16 January 2024 – Rajbinder Sangha
Hide video Show video
(10.00 am)
Mr Blake: Good morning, sir, can you see and hear me?
Sir Wyn Williams: Yes, I can, thank you very much.
Mr Beer: Sir, we may have a small technical issue with the transcription service – no, it’s been fixed, wonderful.
Sir Wyn Williams: Yes, that’s fine.
Mr Blake: Thank you, sir. This morning we’re going to hear from Ms Sangha.
Rajbinder Sangha
RAJBINDER SANGHA (sworn).
Questioned by Mr Blake
Mr Blake: Thank you very much. Can you give your full name, please?
Rajbinder Sangha: Yes, it’s Rajbinder Sangha.
Mr Blake: Thank you very much, Ms Sangha. You should have in front of you a bundle that contains a witness statement behind the first tab, that witness statement has the unique reference of WITN10300100. Is it dated 20 December 2023?
Rajbinder Sangha: Yes, it is.
Mr Blake: Thank you. Can I ask you to turn to the final page, page 11. Can you see a signature there?
Rajbinder Sangha: Yes.
Mr Blake: Is that your signature?
Rajbinder Sangha: It is.
Mr Blake: Can you confirm that that statement is true to the best of your knowledge and belief?
Rajbinder Sangha: Yes, I can.
Mr Blake: Thank you very much. That statement will be published on the Inquiry’s website and anything I’m asking you today will be supplementary to that. I think we see you in some documents as Rajbinder Bains; is that right?
Rajbinder Sangha: That’s correct, yes.
Mr Blake: Was that your maiden name?
Rajbinder Sangha: It was, yes.
Mr Blake: Thank you. You currently work at Fujitsu?
Rajbinder Sangha: Yes.
Mr Blake: Before joining Fujitsu, you graduate from university with a BSc in information systems in 2005; is that right?
Rajbinder Sangha: Yes, that’s correct.
Mr Blake: Thank you. You held a number of customer services and administrative roles at various companies before joining Fujitsu –
Rajbinder Sangha: Yes.
Mr Blake: – and that included Customer Service Liaison Officer at Dell –
Rajbinder Sangha: Yes.
Mr Blake: – and a senior administrator at a Hewlett Packard company?
Rajbinder Sangha: Correct.
Mr Blake: Thank you. You joined Fujitsu in 2010, July 2010?
Rajbinder Sangha: Yes, that’s correct.
Mr Blake: You were part of the Fraud and Litigation Support Office?
Rajbinder Sangha: Yes.
Mr Blake: You held the role of Fraud and Litigation Support Officer?
Rajbinder Sangha: Yes.
Mr Blake: You took some leave in 2016 and rejoined in 2017 as part of the Release Management Department; is that correct?
Rajbinder Sangha: That’s correct.
Mr Blake: Thank you. You stayed in that role until the present day?
Rajbinder Sangha: Yes.
Mr Blake: Today we’re going to be focusing on the period 2010 to 2016, so the period where you were part of the Fraud and Litigation Support Office. You, as part of that job, processed audit data and extracted audit data –
Rajbinder Sangha: Correct.
Mr Blake: – for what we know as ARQ requests –
Rajbinder Sangha: Yes.
Mr Blake: – or sometimes referred to as ARQ requests?
Rajbinder Sangha: Yes.
Mr Blake: Can you assist us with very briefly explaining what an ARQ or ARQ request was?
Rajbinder Sangha: An ARQ request was a record of transactions that had taken place over the counter at different branches.
Mr Blake: Is it held at Fujitsu as part of Fujitsu’s records?
Rajbinder Sangha: Yes.
Mr Blake: What was your job in relation to that?
Rajbinder Sangha: My job was to process the requests that we received from Post Office, so just extract the data and then send it back to Post Office.
Mr Blake: Can we please look at FUJ00232107. This is an email from 2013. If we turn to the third page, we can see the background to this. It seems as though there’s been a question about the amount of data that is archived. I think this particular email relates to an investigation that is, at that time, going on by Detica.
If we turn to the first page, you set out there the contractual limits to ARQ requests; is that right?
Rajbinder Sangha: Yes, that’s correct.
Mr Blake: So can you please explain to us what the difference between those three figures are in that email?
Rajbinder Sangha: So this would have been information that I would have got from senior members of the team, because I was actually – I was not involved in how many ARQs were set per se, so obviously I’ve gone back – I’ve gone back on this email to say that in terms of ARQs we would have 720, for litigation it would be 250.
Mr Blake: So am I right to distinguish those two: ARQs would be, the first line, for some other purpose than litigation –
Rajbinder Sangha: Yes.
Mr Blake: – perhaps some sort of disciplinary investigation or something along those lines, or what might that typically be?
Rajbinder Sangha: I can’t remember. I can’t remember.
Mr Blake: Okay. Over time I think you became involved in other tasks.
Rajbinder Sangha: Correct.
Mr Blake: You described in your witness statement maintaining a database of user permissions and processing security check applications as well.
Rajbinder Sangha: Correct.
Mr Blake: You have also described working on reconciliation?
Rajbinder Sangha: Yes.
Mr Blake: Can you assist us with what you meant by that?
Rajbinder Sangha: So reconciliation was a report that we would run every morning for transactions that not – had not completed over the counter, so our role was to investigate why these – raise calls and investigate why these transactions hadn’t completed.
Mr Blake: There were other tasks that members of your team were involved in. You’ve described those as providing Fujitsu Helpdesk logs and also providing witness statements in court cases; is that right?
Rajbinder Sangha: Correct.
Mr Blake: Can you assist us, who were the members of your team?
Rajbinder Sangha: So the two people I worked with were Penelope Thomas and Andrew Dunks.
Mr Blake: Can you assist us with what their roles or titles were?
Rajbinder Sangha: I think their titles were the same. The roles were – so Andy provided ARQ requests – he processed ARQ requests, he provided Helpdesk calls and also witness statements; and Penelope was the same minus the Helpdesk calls, so she processed ARQ requests and provided witness statements.
Mr Blake: Thank you. I’m just going to stop you there briefly.
Sir, I don’t know if you’re hearing any feedback at all, we have a little internal noise in this room, I’m just wondering if it’s affecting –
Sir Wyn Williams: It’s not coming through to me and, even allowing for my growing age, Mr Blake, it’s – if there is a noise, I can’t hear it.
Mr Blake: Thank you. I’m told it’s the monitor next to the transcriber.
Ms Bains, does it affect you at all?
Rajbinder Sangha: No.
Mr Blake: You describe there are two people in your team?
Rajbinder Sangha: Correct.
Mr Blake: Are you in the same room, is it a small room, a large room?
Rajbinder Sangha: It’s a small secure room, so we would only have access to that room.
Mr Blake: So there are three of you, all had access to a secure single room –
Rajbinder Sangha: Yes.
Mr Blake: – working together in that room. Where was that office?
Rajbinder Sangha: It was based in Bracknell.
Mr Blake: In Bracknell. I think we will see some emails about people on different floors, where were you in relation to, say, the engineers or those who are working on –
Rajbinder Sangha: So I would be based on the fourth floor and then we had another team called the SSC and they were based on sixth floor.
Mr Blake: Thank you. We have seen in some documents somebody called Gareth Jenkins, who we’ve heard a great deal about –
Rajbinder Sangha: Yes.
Mr Blake: – being described as “Expert Litigation Support”. Is that a description that you’re familiar with at all?
Sorry, because it’s being transcribed …
Rajbinder Sangha: No.
Mr Blake: Did somebody like Gareth Jenkins have a different role in relation to the team to yourself?
Rajbinder Sangha: Not that I’m aware of, no, I don’t remember.
Mr Blake: Can we please look at FUJ00226099, please. Thank you. This is an email from Penny Thomas to yourself, and you can see there it’s headed “Expert Litigation Support”.
Rajbinder Sangha: Yes.
Mr Blake: She seems to be having a conversation or sets out a conversation between herself and somebody called Matthew Church; do you know who Matthew Church was?
Rajbinder Sangha: I don’t remember, sorry.
Mr Blake: She says there:
“Hi Matthew, I’m in the process of putting together the charges for Gareth Jenkins for the Expert Litigation Support he provides. Could you please advise what you charge as an hourly rate?
“This is to cover February activities.”
His response is:
“His daily rate is £1,199 divided by 7.4 hours to get the hourly rate of £162.02.”
So that term “expert litigation support” was not something you were familiar with?
Rajbinder Sangha: No.
Mr Blake: Can we look at one more document, it’s FUJ00226107. This is again an email from yourself, the subject here is “Expert Litigation Support”. There are various figures given. So we’ve heard about a case, the case of Mr Ishaq. It sets out there the expenses for Gareth Jenkins for the Ishaq case. Then it says:
“Gareth booked 31 hours for Expert Litigation Support, his time was expended on the following:
“Predominantly this is for the Ishaq case, but also related to the Second Sight investigation and other cases, eg Sefton and Nield, Dixon and Brown.”
It sets out there a cost for March of £5,291. What was your role in relation to the fees of Mr Jenkins?
Rajbinder Sangha: I had no role in the fees. Basically this information would have been sent to me and then my role was just to send it on to Post Office.
Mr Blake: Who would it have been sent to you by?
Rajbinder Sangha: It would have been – so this information – the information on here would have been sent from Gareth himself to myself.
Mr Blake: Are you able to assist us with typical fees per month that might have been incurred by Mr Jenkins?
Rajbinder Sangha: No, I can’t remember.
Mr Blake: I’m going to move on to the topic of your knowledge of issues with the Horizon system. I’m going to begin with when you joined. Can we start by looking at FUJ00122588, please. Thank you very much.
This is a document that pre-dates your joining the team, so it’s a 2008 document. It’s unlikely that you saw it but, if you did see it, please do let us know.
Rajbinder Sangha: No.
Mr Blake: I’m just going to take you through to see what kind of information was passed on to you when you joined the team. If we look at the first page, it’s called “HNG-X CP – Strengthen the HNG-X Audit Solution, and enable analysis of Counter Event messages”. Are you able to assist us with what “HNG-X CP” means? I’m able to say that HNG-X is what we know as Horizon Online. Does that assist you at all? It seems to be a proposal.
Rajbinder Sangha: Yeah, it’s – yeah, um, I can’t remember, sorry.
Mr Blake: If we start with a description of the audit system and ARQ service, it says:
“We are contractually applied to support the Prosecution Support Service via CS, and provide historical extracts of data from the audit archive (7 years of data), used in legal proceedings often to prove accusations of fraud against postmasters.”
Pausing there, was that your understanding of the contractual position?
Rajbinder Sangha: Yes.
Mr Blake: The completeness of the data extracts provided is assumed, and witness statements state as much (see last page).”
We’ll get onto that last page where we see a form of words that’s used in witness statements:
“The service has worked by providing extracts of Riposte of Message Store data only.”
Then it says:
“This service (worth the best part of the annual £850,000 security revenue …)”
That says, “PS”, it looks as though that’s inserted by Pete Sewell:
“… will remain to 2015 and beyond.”
Were you aware of the financial significance of the ARQ service –
Rajbinder Sangha: No.
Mr Blake: – to Fujitsu?
Rajbinder Sangha: No, I wasn’t.
Mr Blake: Were you made aware when you joined the business of its significance?
Rajbinder Sangha: Yes, I was.
Mr Blake: Yes. There’s then identified in this document a problem and it quotes from the number of a PEAK. Are you aware what a PEAK is?
Rajbinder Sangha: Yes.
Mr Blake: We’ve looked at those?
Rajbinder Sangha: Yes, I –
Mr Blake: Thank you. So it has:
“PC0152376 highlighted that in certain error conditions in the EOD process Riposte cannot be relied upon to write a consistent set of messages to the local store.”
Just pausing there, I wonder whether we can get up PC0152376, it’s FUJ00154684. We’ll just have a quick look at the underlying PEAK that’s referred to here. It can either be alongside or separate.
This is the first page. Could we look at about halfway down the page, if we could have a look at the words, if we scroll down slightly. It’s refers to a member of the NBSC – scroll down a little, thank you:
“Ibrahim from the NBSC [that’s the National Business Support Centre] has asked that an issue be investigated by our software team regarding discrepancies still showing when the MIS stock unit is rolled to clear the local suspense account.”
I’m not going to ask you in any detail about this particular problem but it relates to discrepancies showing. If we turn over to page 3, please, there’s an entry on page 3 from Mr Barnes, who we are going to hear from tomorrow, 2 January 2008. So this is two years before you joined the team. He says, as follows:
“The fact that EPOSS code is not resilient to errors is endemic. There seems little point fixing it in this one particular case because there will be many others to catch you out. For example when I tried to balance with CABSProcess running I found that declaring cash failed with the same sort of error message!
“It may be worth passing on the general message to the HNG-X team that in many cases code should always try and exit gracefully after an error and not just blunder on regardless.”
Just pausing there, were you made aware when you joined the team, of issues with, for example, the EPOSS code?
Rajbinder Sangha: No, I wasn’t.
Mr Blake: Could we go back to the document we were just looking at, that’s FUJ00122588, please. Were you aware with general issues regarding data integrity and concerns about the integrity of, for example, ARQ data?
Rajbinder Sangha: No, I wouldn’t have been because it wasn’t part of my role and, if there was issues within the system, it would be dealt with by higher members of the team, so members that had more knowledge of the system.
Mr Blake: So sticking where we are, let’s move to the second bullet point under “Problem”. It says:
“This particular issue has been fixed … but it is very probable that similar problems exist in the Horizon system.
“Therefore the process of providing data now needs to include the extraction and cross-checking of event data to help identify where data integrity might be compromised.
“The statements currently asserted in witness statement cannot be guaranteed in all cases (even after this CP) … but this CP seeks to strengthen the process and allow us to reliably identify where the assertion can or cannot be made.”
Then it says:
“Current Process
“Many manual steps, requiring great care and skill from individual resources, obvious potential for human error.
“Data distributed or transferred over too many platforms/media: inherently insecure.
“Tactical solution [which has] introduced further manual steps.”
Were you aware, as somebody who was involved in that overall ARQ process, that there had been a number of manual steps to work around certain problems?
Rajbinder Sangha: No, I wasn’t.
Mr Blake: If we go over the page, please. Under “Benefit/Risk”, if we scroll down slightly, it says as follows:
“If we cannot better identify where data integrity can or cannot be guaranteed, then we are in breach of contract and may:
“Be fined heavily [or]
“Not be able to offer the ARQ service, or will undermine confidence in the service.”
Then scrolling down, we have that witness statement extract. We will see in due course this becomes part of or is included in a pro forma type witness statement.
Rajbinder Sangha: Yes.
Mr Blake: It’s a form of words that says, as follows:
“An audit of all information handled by the TMS [that’s Transaction Management Service] is taken daily by copying all new messages to archive media”, et cetera, and it details the process that’s involved.
Just pausing there on this particular document, do you think, at the time that you joined the ARQ team, it would have been helpful to have known about concerns about the integrity of the ARQ system?
Rajbinder Sangha: Yes.
Mr Blake: Yes. You joined then in July 2010. Could we please move to FUJ00122925, please. This is 14 July 2010 and you are sent by Penny Thomas a standard witness statement for you to read through. Now, was this very early on in your time, was this reading in to the job or –
Rajbinder Sangha: Yes, it was. It was reading into the job and understanding what roles we were undertaking.
Mr Blake: We can have a look at that standard witness statement. That’s at FUJ00122926. Thank you very much. This is the attachment to that email.
Rajbinder Sangha: Correct.
Mr Blake: It doesn’t currently have anybody’s name on it. It’s a pro forma standard witness statement that says, as follows, it says:
“I am authorised by Fujitsu Services to undertake extractions of audit archive data and to obtain information regarding system transactions recorded on the Horizon system.”
Scrolling down, it says:
“Horizon’s documented procedures stipulate how the Horizon system operates, and while I am not involved with any of the technical aspects of the Horizon system, these documented processes allow me to provide a general overview.”
Was it your understanding that this was to be used by those who weren’t, for example, the engineers, like Gareth Jenkins, so didn’t have quite that technical understanding, but who nevertheless provided –
Rajbinder Sangha: Correct.
Mr Blake: – some evidence in criminal proceedings?
Rajbinder Sangha: Yes, correct.
Mr Blake: Could we scroll over to page 3., and that second paragraph on page 3. That’s the form of words I just took you to in that earlier 2008 document. Then if we go over the page again, there is, if we scroll down a little bit, it sets out the various controls that apply to the audit extraction process. For example, it says at 1:
“Extractions can only be made through the [audit workstations] which exist at Fujitsu Services”, and it gives the address.
Stopping there, were you part of that extraction process?
Rajbinder Sangha: I was.
Mr Blake: Did you extract?
Rajbinder Sangha: Yes.
Mr Blake: Number 2:
“Logical access to the [audit workstation] and its functionality is managed in accordance with certain principles. This includes dedicated logins, password control and use of Microsoft Windows NT security features.”
So, in order to extract the data, there were certain security controls in place to ensure that there were sufficient controls on that data.
Rajbinder Sangha: Correct, yes.
Mr Blake: “3. All extractions are logged on the [audit workstation] and supported by documented Audit Record Queries, authorised by nominated persons within the Post Office. This log can be scrutinised on the [audit workstation].
“4. Extractions are only made by authorised individuals”, et cetera.
So this is essentially reassuring a court as to the integrity of the data because it can be accessed by those who don’t have a login, for example?
Rajbinder Sangha: Yes.
Mr Blake: If we go over the page, please, to page 5. If we scroll down three-quarters of the way, we can see where a particular ARQ number is inserted. So this is the pro forma part that would be filled out by the person completing that statement:
“ARQ (NUMBER) …”
I think that means into insert the number there.
Rajbinder Sangha: Yes.
Mr Blake: “… was received on [insert the date] and asked for information in connection with the Post Office at [and then you give the Post Office branch details].”
That’s effectively producing the ARQ data for the court; is that your understanding?
Rajbinder Sangha: Yes, that’s correct, yeah.
Mr Blake: Can we scroll down to the final substantive page, it’s page 7. Thank you very much. It’s that paragraph that’s currently on the screen, the bottom paragraph there, this is another part of that standard witness statement. It’s a form of words that we’ll see in a number of other witness statements. It says, as follows:
“There is no reason to believe that the information in this statement is inaccurate because of the improper use of the system. To the best of my knowledge and belief, at all material times the system was operating properly, or if not, any respect in which it was not operating properly, or was out of operation was not such as to effect the information held within it.”
Did that paragraph at the time cause you any concern?
Rajbinder Sangha: No.
Mr Blake: No. Does it now cause you any concern?
Rajbinder Sangha: No, it doesn’t, no.
Mr Blake: It doesn’t cause you any concern? We’re now going to look at an issue in 2010, so the year you joined, with the ARQ data. Could we look at FUJ00172183. The document we’re going to look at is another PEAK and we’re going to look at an entry that begins on 21 June 2010, so very soon before you joined – it’s FUJ00172183 – and it covers the period when you did join subsequently.
Rajbinder Sangha: Okay.
Mr Blake: But we’ll start with the entry on 21 June 2010. Thank you. So we can see there it’s PC0200468. There’s a summary at the top, perhaps I will read that out, but we will get to that. That’s further down in the log. It says:
“From Penny – In a nutshell HNG-X [that’s Horizon Online] application is not removing duplicate transactions (which may have been recorded twice on the audit server) and they are appearing in the ARQ returns. For the old Horizon application Riposte automatically removed duplicate entries. An initial analysis shows that one third of all ARQ returns (since the new application has been in play) have duplicated transactions.”
If we scroll down, please, to the second entry in this log. We have an entry from Penny Thomas. She says:
“While performing an audit retrieval for branch [and it gives the branch details] duplicate transactions have been found [and it gives the date]. Initial analysis shows that duplicate records are held in 2 different audited TMS files.”
Scrolling down the page, the final entry on the page from Mr Barnes, he says as follows, on 22 June:
“The way it works is that it processes all the results in a given file building up an internal table of transactions consequence for that file. Then at the very end of the processing the file it dumps the internal table to [and gives it details of the table]. It does not cross-check the transactions in one file against another file.”
He outlines two solutions that are possible. There’s an “easy solution”:
“As each transaction is processed a check is made … and if it’s already there the transaction is ignored writing a warning to the query logged. The problem with this solution is that a query needs to be made to the database for every transaction.”
Then he describes the “more difficult solution”, if we go over the page, please:
“The internal table which at the moment is built up on a per file basis is changed to being built up on a per query basis. The check for duplicate transactions is then done within the internal table. This is a much more thorough approach but will take much more work.”
Then there’s an entry on 22 June 2010 and it’s outlining the details of a fix to this problem. If we scroll down, it describes the impact on the user of the issue. It’s about halfway down this page. “Impact on user”, it’s almost at the bottom of this page. It says:
“Occasionally duplicate transactions are listed in the spreadsheets produced and presented to court for prosecution cases. These can give the defence teams ground to question the evidence.”
Then it says:
“Have relevant KELs been created or updated?”
So a KEL – are you aware of a Known Error Log.
Rajbinder Sangha: Yes, I am.
Mr Blake: It says:
“No [Known Error Logs] have been created for this since we intend to fully resolve the issue shortly.”
If we scroll down, there are risks that are outlined of the fix. It says:
“If we do not fix this problem our spreadsheets present in court are liable to be brought into doubt if duplicate transactions are spotted.”
If we go over the page, please. It says:
“QueryDLL.dll is a recent component introduced for [Horizon Online] and has not yet fully bedded down and so it is likely to change as problems such as this one are spotted.”
Just pausing there, on joining the team – and you’ve said it was a small team, three people – were you not made aware of issues such as this that could affect the reliability of information presented to court?
Rajbinder Sangha: No, I wasn’t made aware.
Mr Blake: If we scroll down, let’s look at the entry of 23 June from Penny Thomas, it’s just there. It says:
“Initial analysis of all ARQ returns since the [Horizon Online] application has been implemented identifies approximately one third (of all returns) have duplicate entries. This is now extremely urgent.”
Scroll down, please, towards the bottom of the page, an entry of 7 July 2010. It’s the bottom entry that’s currently on screen, 7 July 2010, and it says as follows – it says that it’s been fixed, essentially, 7 July.
Then if we keep on going down, scrolling down, over the page, 30 July. It seems as though there’s, in fact, a testing of the fix, 30 July. It says there, the entry from Sheila Bamber, it says:
“This PEAK has been tested in LST and fix will be read with Release 2 data centre.”
Down the bottom of the page, an entry from Penny Thomas, 1 September 2010:
“Fix successfully deployed …”
So it seems as though that PEAK, that incident log, was open from before you joined and wasn’t closed until 1 September and wasn’t successfully fixed until 1 September –
Rajbinder Sangha: Correct.
Mr Blake: – so after you had joined the team?
Rajbinder Sangha: Yes.
Mr Blake: Does it cause you any concern that you weren’t aware of that kind of thing on joining and –
Rajbinder Sangha: Yeah, it does because, obviously, I didn’t realise that the data – that there was these issues with the –
Mr Blake: Yes. Can we please have a look at FUJ00172047. This is an email chain from Penny Thomas. You are copied in at this stage and this is 21 July 2010. Its title is PC0200468. So it does certainly as though you were copied into conversations about this particular issue. Do you agree with that?
Rajbinder Sangha: Yes, I do.
Mr Blake: Could we, please, look at page 10. It’s a discussion of the issue we’ve just looked at, from Penny Thomas to Pat Lywood are you able to assist us? I think it says at the top there “Service Implementation Manager”; was Pat Lywood someone who you knew?
Rajbinder Sangha: No.
Mr Blake: If we look at that final entry, Penny Thomas says as follows, at the bottom entry:
“We have a very significant problem which has been recorded [and it gives the PEAK details]. In a nutshell the [Horizon Online] application is not removing duplicate transactions (which may have been recorded twice on the Audit Server) and they are appearing in the ARQ returns. For the old Horizon application Riposte automatically removed duplicate entries. An initial analysis shows that one third of all ARQ returns (since the new application has been in play) have duplicated transactions.”
That’s the entry, effectively, taken from the PEAK –
Rajbinder Sangha: Yes.
Mr Blake: – or later effectively inserted into the PEAK. Could we look at page 9, please. You’re not, at this stage, copied in. This was a chain that was ultimately copied to you but you’re not part of this discussion. If we look at the bottom of page 9 there’s an email from Graham Welsh to a number of people at Fujitsu. Looking at their names, they all seem to be involved in the technical aspects of Horizon.
Sorry, it’s the bottom of page 8. The email itself is on the top of page 9 as we can see. He says:
“Please see below attached [for your information]. In essence we have a problem with the ARQ extraction tool. Under Horizon this would inhibit the duplicate transactions held for the audit server and thus supply evidence for court etc without duplicated records.
“However the [Horizon Online] tool does not and thus duplicate records that cannot be differentiated are supplied as evidence. Thus could allow for legal challenge to the integrity of the system.”
Now, you don’t have recollection of discussing this issue, you’re at least on the chain of this issue. Why do you think it is that the significance of this was not more prominent in your three-person team?
Rajbinder Sangha: I think this was dealt with senior members of the team that obviously had more knowledge of the system –
Mr Blake: But you are one of those people who is extracting –
Rajbinder Sangha: Correct.
Mr Blake: – the ARQ data. It says there that there is a problem with the ARQ extraction tool and could allow for a legal challenge to integrity of the system. Looking back at it, does it seem to you to be quite a significant issue?
Rajbinder Sangha: Yes, looking back at it now, yes, it is.
Mr Blake: But it wasn’t at the time or wasn’t seen in your team as a significant issue –
Rajbinder Sangha: Yes.
Mr Blake: – or not sufficiently prominent to draw to your attention?
Rajbinder Sangha: Yeah.
Mr Blake: Could we look at the bottom of page 6 and into page 7, please. There’s an email from Andrew Mansfield. He says as follows:
“Gerald has produced a fix for Release 1 and it is ready to go.
“He has added an impact statement to the PEAK that includes a brief statement on testing: an audit request should be performed that retrieves and processes TMS files containing duplicate transactions. It should be confirmed that the duplicate transactions have been removed from the final spreadsheet generated by the audit application.
“The PEAK is currently with RMF for targeting.”
He says:
“There has been discussion of a possible workaround. This involves modifying the audit queries so that the message numbers are included in the output to the spreadsheets (currently they are not). This would allow the duplicate messages to be identified and removed by running a macro on the final spreadsheet generated by the application.”
He then says, if we could scroll down slightly:
“Penny Thomas is in discussion with the Post Office over whether this workaround is acceptable in the short-term.”
Could we scroll up, please, to page 5 – at the bottom of page 5, thank you, if we could stop there. So we have a message from Penny Thomas saying that:
“[The Post Office] has gone to [Post Office Legal] for guidance and further returns have been identified this morning as bound for court.”
So it seems as though ARQ data is heading to court whilst this issue has already been identified. Then we have a response from Graham Welsh, and he says:
“[Please see below] from Penny. I understand that there are more court cases pending and whilst the briefing to the investigation has taken place they are coming back requesting help due to the level of activity and nervousness regarding the current workaround.”
In terms of workaround, as somebody who was extracting the data, were you performing this workaround? Were you, for example, manually checking certain things due to this kind of issue?
Rajbinder Sangha: I can’t remember.
Mr Blake: I mean, presumably you did more than just press a button and the ARQ data comes up on your screen.
Rajbinder Sangha: So my role was to extract the data. So, basically, enter in the branch code, enter a date range and that would extract the information. That was my responsibility.
Mr Blake: So you weren’t involved in the actual checking of the underlying data?
Rajbinder Sangha: No, I was not there – I wasn’t – my role was not to analyse the data.
Mr Blake: Who was it that would perform things like these workarounds that we see here? So something like checking for duplication within the –
Rajbinder Sangha: So part of the process would have been that, when we extracted the data, another member of the team would ensure that whatever had been requested was included in the – is what was produced.
Mr Blake: You say another member of the team. There were three of you?
Rajbinder Sangha: So yeah, it would either have been Andy or Penny.
Mr Blake: Can we, please, look at page 4, if we scroll up. At the bottom of page 4, from somebody called Tom Lillywhite, who is the Principal Security Consultant. He says:
“Have just read this … suggest keeping Penny in the loop … she knows just how nervous the customer is/will become and may have something to add …”
At this stage, did you have any interaction with the Post Office?
Rajbinder Sangha: Apart from processing the requests, that’s the only interaction I had with them.
Mr Blake: So you would send something to the Post Office?
Rajbinder Sangha: Yes, yes.
Mr Blake: If we have a look at the email above, Mr Welsh says, as follows, he says:
“The sooner this is resolved the easier the task will be in managing [Post Office] expectations in this area while minimising the risk of duplicated effort by having to reproduce reports already provided.”
So it seems as though there is an attempt to avoid having to reproduce those reports that have already been provided in, for example, court proceedings?
Rajbinder Sangha: Correct.
Mr Blake: Given that you were part of a very small team and you were part of that ARQ process, does it not strike you as odd that there wasn’t a conversation with, for example, Penny Thomas about this issue?
Rajbinder Sangha: Penny was a senior member of the team so she dealt with this – these kind of issues. But I don’t recall having a conversation, no, being told.
Mr Blake: Was she somebody who kept her cards very close to her chest? Was she not somebody who discussed issues that were actually affecting the very job that you were carrying out?
Rajbinder Sangha: Sorry, can you repeat that question again? Sorry.
Mr Blake: I think what we’ve identified is that there is a serious issue that’s identified in various emails. If we look at the first email, it seems to have at least been copied to you but I think your evidence is that you don’t remember any discussion about it and, really, my question is why didn’t you have a conversation with Penny Thomas? Was there something about Penny Thomas that made it unlikely that you would that have a conversation about it?
Rajbinder Sangha: I don’t recall having a conversation.
Mr Blake: In terms of Penny Thomas, what was your relationship?
Rajbinder Sangha: My relationship with Penny was, if there was something that I was unsure of in terms of the ARQ requests, I would approach her.
Mr Blake: Yes. There were three of you in the room. Did you sit in silence all day? Did you discuss matters that affected your work?
Rajbinder Sangha: We discussed things, yes.
Mr Blake: Why do you think it is or may be that Penny Thomas didn’t discuss an issue such as this with you?
Rajbinder Sangha: I don’t know. I don’t know.
Mr Blake: Now, we saw that you received that draft statement early on in your time at the Post Office, the draft pro forma statement, that we looked at. That was dated or sent to you on 14 July. We’re now on 21 July when this chain is being sent to you, so not that far after having received that statement. Did it not cause you any concerns about the reliability of the statement, the pro forma statement?
Rajbinder Sangha: At the time no because I was not involved in producing a witness statement for going to court proceedings.
Mr Blake: Yes. Does it cause you any concern now?
Rajbinder Sangha: Yes, it does.
Mr Blake: Why does it cause you concern now?
Rajbinder Sangha: Because, obviously, we had bugs in the system.
Mr Blake: Can we please look at FUJ00225940. This is another PEAK. Thank you. This is PEAK PC0204310 and it relates to “duplicate JSN detected”, and the summary there, if we scroll down:
“Support overhead currently 30 incidents in 5 days. Resolution will mean that any future occurrences will have a different root cause and require investigation. Such incidents are currently getting masked by the volume associated with this PEAK with consequential risk to the integrity of the audit trail used for litigation support.”
If we scroll down, please, we have an entry of 13 September. So a couple of months into the job. It says there:
“Exception raised while processing message event. Serious system error.”
Perhaps we could look at page 8, please, and a couple of entries on page 8. Thank you. The second entry there, 19 October 2010, an entry that says, as follows:
“A new Business Impact has been added:
“This PEAK will reduce the support overhead because it will reduce the number of alerts associated with duplicate JSNs. It will also mean that any future occurrences will have occurred for a different reason and would require investigation. Such incidents are currently getting masked by the occurrences associated with this PEAK.
“There is no immediate benefit to the customer other than support engineers no longer need to worry about these types of alert and can focus on another alerts.
“Currently this type of error generates a high frequency of alerts on a daily basis, masking other types of error.”
If we scroll down near the bottom of the page to 2 November, an entry from Steve Parker. He says as follows:
“The Business Impact has been updated:
“Support overhead currently 30 incidents in 5 days. Resolution will mean that any future occurrences will have a different root cause and will require investigation. Such incidents are currently getting masked by the volume associated with this PEAK with consequential risk to the integrity of the audit trail used for litigation support.”
Do you recall this issue?
Rajbinder Sangha: I do.
Mr Blake: You do?
Rajbinder Sangha: Yeah.
Mr Blake: Why do you recall this issue? What was it about this issue that stood out?
Rajbinder Sangha: I just remember being made aware that we had duplicate JSNs. In terms of the severity of it, technical background, I didn’t have that much knowledge of the system but I remember being told about this, yeah.
Mr Blake: Thank you. I’m going to take you to an email chain, it’s at FUJ00228770. It’s an email chain from November 2010. Can we start on page 5, please. We’ll see your name appear within this email chain, on – if we look at page 5, you’re not currently copied in on this particular email but this describes the workaround to this particular problem.
The subject is “PC0204310 duplicate JSN detected”, so that’s the PEAK that we’ve just been looking at?
Rajbinder Sangha: Yes.
Mr Blake: Mr Parker says:
“It will be difficult to get the right technical people together on this one for a face-to-face discussion. My take on it is:
“1) Risk to support is large. It is impossible for support to check all the duplicate JSN events to ensure they are the same issues described on”, and it gives details.
“2) The 100 [approximately] incidents that support have checked all fall into the scenarios described in [and that’s a particular reference]. These are all safe to ignore.
“3) We risk other parts of the program by trying to force through a fix …”
I think we’ve heard during earlier parts of this Inquiry things like code regression, or other issues with code caused by particular fixes.
Rajbinder Sangha: Yes.
Mr Blake: Is that something you’re aware of?
Rajbinder Sangha: I’m not aware of it.
Mr Blake: As a general principle?
Rajbinder Sangha: Yeah, as a general principle, yeah.
Mr Blake: “4) Risk to audit is very small. Should a true duplicate JSN slip through then it will be noticed by a failure as described [and it gives a reference]. Such incidents will still need to be investigated urgently.
“Pragmatic approach, given the approach, is to ignore all duplicate JSN messages in BAL logs until [the issue] is resolved ([which is going to be] early next year). There is a small risk that by ignoring this event we will be missing an issue that needs investigation.”
If we look at page 3 at the bottom of page 2 into page 3 there is discussion about the workaround that is going to take place until the actual resolution of the issue. Bottom of page 2 into page 3. It’s an email from somebody called Sarah Selwyn to Steve Parker and others. Is Sarah Selwyn someone familiar to you?
Rajbinder Sangha: I know of her, yeah.
Mr Blake: Do you know what her role was at all?
Rajbinder Sangha: I can’t remember.
Mr Blake: So, thank you. She says as follows:
“Steve,
“I agree with your approach as long as any JSN duplicates that match the criteria described … are still investigated urgently, as you described below in item 4.”
The second paragraph describes the workaround. She says:
“What has not been described below is the additional effort that the existence of duplicates places on the Litigation Support Team. Until the two fixes relate to the PEAKs described below are delivered to live, Penny and team will need to run the macro provided as a workaround against very spreadsheet generated by the fast ARQ method to determine if there are any duplicate spreadsheet rows present. These rows do not include JSN.”
Just pausing there, we are going to hear more about it this week, but can you assist us with what a fast ARQ was, as opposed to a slow one?
Rajbinder Sangha: A fast ARQ, what I can recall, is I think we used to put the date range and the FAD code and click a button and it would return the data.
Mr Blake: And a slow one would be what, more manual?
Rajbinder Sangha: Yeah, I think so, yeah.
Mr Blake: Thank you. If we go back to the substantive email, please. It continues as follows, it says:
“If there are any duplicate spreadsheet rows present, these rows do not include JSN. If there are duplicates present, then Penny and team run one of the slow ARQ queries which have been modified to include JSN in order to determine if the duplicate is a true duplicate.”
So it seems as though the workaround involves looking for duplicate entries and then changing the process if duplicate entries are found. Is that something you recall at all?
Rajbinder Sangha: I – yeah, I don’t recall, sorry.
Mr Blake: If we look at that third paragraph on the page, halfway through that paragraph, it says:
“Penny and team will need to continue manually running the workaround macro until at least April next year. The resolution [and it gives it the reference of the PEAK] delivered 3.20 delivered early next year should reduce JSN duplicates in any [Horizon Online] audit analysed but the macro will still need to be run until [a certain release] just in case the audit being analysed is”, et cetera.
So it seems as though Penny Thomas and team – would that team involve – who would that be, “Penny and team”? Is that your team, is that –
Rajbinder Sangha: Correct. Penny, myself and Andy.
Mr Blake: So the three of you would need to continually manually running the workaround macro until April next year. So a considerable period of time; we’re in November until April?
Rajbinder Sangha: Yes.
Mr Blake: Do you recall the first document that I took you to or second document, the 2008 document about Legacy Horizon having these manual workarounds which creates risks workarounds. If we apply those concerns here, can you see potential risks to this workaround process for audit data?
Rajbinder Sangha: Yes, I can see potential risks, yes.
Mr Blake: We’ve seen those witness statements, for example, that talk about the security needed, the logins needed, the special entry to the special room, but if those who are carrying out the process are themselves carrying out a more manual process, do you see risks involved in that?
Rajbinder Sangha: Yes, yes.
Mr Blake: Then if we go to the first page – the very first, thank you very much. You’re not copied in here. So you were copied into the earlier chain but this seems to be the top of the chain that you don’t appear to be copied into. It’s from Penny Thomas and she says:
“All
“The analysis we have conducted (covering …)”
Sorry, in fact, if we could go over the page, if we could start on the second page because there’s an email from Sarah Selwyn to Penny Thomas, yourself and others. She says:
“Penny and Raj,
“Thank you both for your analysis of the business impact of running the workaround fixes for detection of JSNs in HNG-X audit.”
Do you remember carrying out analysis of the business impact?
Rajbinder Sangha: I don’t.
Mr Blake: Do you think it’s likely that you were part of that, given that she has addressed both of you and said, “Thank you both for your analysis”?
Rajbinder Sangha: Yeah, definitely, yeah, I was part of it, yeah.
Mr Blake: She says –
Sir Wyn Williams: Right, I was just going to say it wasn’t on my screen, Mr Blake, but it popped up as I was saying it. Sorry.
Mr Blake: Thank you, sir.
So it’s an email from Sarah Selwyn, who says:
“Penny and Raj,
“Thank you both of your analysis of the business impact of running the workaround fixes for detection of JSNs in HNG-X audit.
“Penny, the permanent fixes to the audit workstation for JSN detection and analysis will be supplied in release 4.37 … which is currently expected to be out of LST on 04/05/2011. There is no live data predicted yet for [the release] but usually this would follow within a few days. You should expect to be running the workaround solution until May 2011.”
That’s similar detail to the detail we’ve just seen.
Then if we go, please, to the first page, which is where Penny Thomas addresses a wider audience, explains the workaround and she says:
“The analysis we have conducted (covering receipts over the last 4 months) …”
Does that assist you with your recollection as to the analysis that was conducted?
Rajbinder Sangha: Yes.
Mr Blake: Yes, and what might that mean, “covering receipts over the last four months”?
Rajbinder Sangha: I think this was the request that we received. So we would have – can you explain that? Can you just explain that question, sorry.
Mr Blake: I think you conducted some analysis and she is telling people about the analysis that’s been conducted. Perhaps you can assist us with that first paragraph.
Rajbinder Sangha: I can’t remember the analysis that I had done.
Mr Blake: Yes.
Rajbinder Sangha: Yeah.
Mr Blake: But are you able to assist, by looking at that, what it might have involved?
Rajbinder Sangha: I can’t, sorry, no.
Mr Blake: It continues:
“We anticipate that by May 11 the bulk of the requests we receive will be for Horizon Online records covering the time frame January to December 2010. Indeed from February 2011 the bulk of the requests may consist of [Horizon Online] records.”
Then there’s a paragraph that describes the workaround, if we scroll down, and she says:
“Therefore, for all retrievals we will need to include additional spreadsheets and [a checking process].”
Then:
“Running additional reports, using a macro and manually checking spreadsheets will increase significantly the time to complete a retrieval; I estimate that an additional 20 minutes will be required to complete each ARQ, and that will require an additional 3 working days per month to be found. Additional work requirements are already being placed on the Prosecution Support Team in the form of supporting Reconciliation and there is a possibility we will be more than stretched to fulfil are required ARQ return timeframes. These changes will alleviate unnecessary pressure on the team and should be implemented at the earliest opportunity.”
Do you recall during this period your small team being particularly busy with these workarounds with creating, for example, manually checking various spreadsheets?
Rajbinder Sangha: Yes.
Mr Blake: Did you personally get involved in that process?
Rajbinder Sangha: Yes, I did.
Mr Blake: We’ve talked about earlier issues that I think your evidence can be accurately summarised as not being seen as such significant issues or not significant enough to have been brought to your attention?
Rajbinder Sangha: Correct, yeah.
Mr Blake: Was this particular issue seen as a significant issue?
Rajbinder Sangha: Yes, it was, yes.
Mr Blake: Yes, and it was seen as a significant issue, why?
Rajbinder Sangha: Because of the data we were providing back to Post Office.
Mr Blake: What were the concerns?
Rajbinder Sangha: That there could be duplicate transactions and obviously, because this information was being provided to Post Office to be taken to court, so it seemed quite –
Mr Blake: Because it could impact on the reliability of the audit –
Rajbinder Sangha: Correct.
Mr Blake: – data that is being provided to the Post Office?
Rajbinder Sangha: Correct, yes.
Mr Blake: I want to ask you about the provision of witness statements. Could we please look at FUJ00225719, please. I’m going to start on the bottom of page 2 into page 3, please. That’s an email of 6 September 2010. So we’re going slightly back in time now. There’s an email from yourself to Andrew Dunks and you say, as follows:
“Just spoken to Maureen and the data for Kirkoswald [that’s a Post Office] needs to be resent as this had duplicated data. I have rerun the reports for you, so please can you check and get these sent to [the Post Office], also I understand that these also require the Witness Statements as well so please can you get this done as we will need to get this out [as soon as possible] as this will be going to court on 20 September.”
So it seems as though there is an ARQ return that contains duplicated data in September 2010, and you are asking Andrew Dunks to check; is that right?
Rajbinder Sangha: Correct, yes.
Mr Blake: If we scroll up, we have an email from Penny Thomas to yourself. It says:
“Please make sure Andy also presents 059.”
Are you able to assist us with what that means?
Rajbinder Sangha: No, I don’t know what 059 was.
Mr Blake: No.
Rajbinder Sangha: I don’t know whether it was an ARQ request or a request for a witness statement.
Mr Blake: If we see there, “Subject” and then it says, “ARQ P048-P058”, and it then refers to 59, does that assist you at all?
Rajbinder Sangha: No, I don’t know what 59 would have been, no, sorry.
Mr Blake: If we scroll up, please, there’s an email from yourself to Penny Thomas. You say as follows:
“Andy has checked the information on the disk, but with regards to the witness statement mentioned that would need to compile this as I have ran the retrievals off, please advise whether I would need to do this.”
Then if we go over to the first page, please, there’s a response from Penny Thomas to Andy Dunks, and she says as follows:
“Andy
“As you know, these returns are reruns of work you have already completed; 3 out of the 12 of your returns contain duplicate transactions, all … need to be re-presented on one disk. They are due in court on 20 September. We have to get them to Salford and they, in turn, have to relay to the Investigator, so we need to get them in the post today.
“In order to get these out as quickly as possible I asked Raj to rerun them for you. If you really are unhappy to take ownership of the work she has completed on your behalf please could you rerun them for yourself; raj can show you the fast ARQ process which really is not onerous at all. You will need to update your witness statement. The disk and statement need to be in the post today, please.”
It seems as though Penny Thomas has effectively stepped in to say that Andy Dunks should be providing the witness statement but not yourself –
Rajbinder Sangha: Correct, yes.
Mr Blake: – is that right?
Rajbinder Sangha: Yeah.
Mr Blake: Is that something that she did regularly?
Rajbinder Sangha: So when I joined the company, I wasn’t comfortable in providing a witness statement. So any requests that did come in for a witness statement it would be either Penny or Andy that would provide that.
Mr Blake: Yes. Can we please look at FUJ00156224. This is a different case, Preston Road post office, and you’re being asked by a Detective Constable so this is a police request for a witness statement, and he says:
“Firstly, many thanks for producing the further disk with ARQ data for the Preston Road office.
“Could I please trouble you to provide a statement exhibiting the CD that you sent through.
“I can provide a draft statement and send it to you unless you have a corporate document that can be used. Penny Thomas provided one such statement previously for this case.”
If we look at the top email there’s an email from yourself to Andy Dunks, and you say:
“I have been requested to produce a statement for Preston Road, please see below. When you’re free are you able to go through the witness statement document with me as I have not produced one?”
Is that a reference to the pro forma statement that we saw earlier today?
Rajbinder Sangha: Yes, it is, yeah.
Mr Blake: It seems at that point that it may be that you’re considering whether you will give a witness statement or not.
Rajbinder Sangha: Yes.
Mr Blake: Can we please look at FUJ00122622. The officer in that email refers to an earlier witness statement from Penny Thomas and we have that earlier witness statement. It’s dated 28 September 2009. This is very much like the pro forma that we saw earlier; do you agree with that?
Rajbinder Sangha: Yes.
Mr Blake: If we look at page 3, for example, about halfway down, it has the words “Preston Road branch” and that’s the insertion of the particular branch into the pro forma, and then that final paragraph, for example, is the one that we saw from the 2008 document into the pro forma –
Rajbinder Sangha: Yeah.
Mr Blake: – and here we see it in an actual statement.
Over the page, please. We have the various controls that apply. We’ve already looked at those in the pro forma.
Then over another page, please, we have below that, please, over the page at the bottom of this page, we have that paragraph that appeared in the pro forma:
“There is no reason to believe that the information in this statement is inaccurate because of the improper use of the system. To the best of my knowledge and belief at all material times the system was operating properly, or if not, any respect in which it was not operating properly, or was out of operation was not such as to effect the information held within it.”
When we began today and you looked at the pro forma I think your evidence was that you didn’t see a problem with that paragraph. Having gone through all of the various issues that have been identified, would you yourself sign up to that form of words?
Rajbinder Sangha: No. No, I wouldn’t.
Mr Blake: Why not?
Rajbinder Sangha: Because of all the issues that have been identified.
Mr Blake: Yes. Could we please look at FUJ00123081. This is the statement that was ultimately provided. It seems as though bomb actually provided the Preston Road statement in the end. We’re there, 19 October 2010. She’s provided a second statement, and she says:
“Further to my statement dated 28 September 2009 …”
So that’s further to the statement we just saw.
If we could scroll down, please, she then exhibits the data that you produced and we see at the bottom of the next page again that form of words used at the very bottom into the next page, same form of words from the previous statement, same form of words from the pro forma; is that right?
Rajbinder Sangha: Yes.
Mr Blake: If we keep scrolling, please. Thank you. Can we look at page 17 of this same document, please. There’s an email chain on page 17 which is the original request from the officer at the bottom of the page, that we’ve already seen. If we have a little look at that, so that’s the request from the officer. Then if we scroll up, Penny Thomas says as follows, she says:
“Please be advised that all requests for Fujitsu support, ie data, statements, court attendance, etc, are to be requested via the Security Team at Salford. We are unable to respond to any requests you may make.”
If we scroll up to page 15, please, there’s the response from the officer. He emails Mark Dinsdale. Is Mark Dinsdale the appropriate person at Fujitsu to contact about the obtaining of a witness statement? Is that something you’re able to assist with?
Rajbinder Sangha: Mark Dinsdale … I recall the name but I can’t remember whether he was Fujitsu or Post Office.
Mr Blake: The officer says as follows:
“Penny Thomas provided a very thorough statement previously for this case. All I require now is a brief statement to exhibit the second disc, which has been produced. Raja Bains dealt with this request from the Post Office … and I have received the disc.”
Then he says out a potential form of words for a statement, essentially just exhibiting the audit records that you had produced.
If we scroll down, the final paragraph there, he says:
“I would appreciate if this could be passed on to Raj (ideally) or another, who could complete this statement TOMORROW … and get it sent to me as a matter of urgency, as this trial commences next week …”
If we scroll up to the very top, we have Mark Dinsdale to Penny Thomas, he says:
“Penny are you able to ask Raj to do the witness statement please.”
So it seems as though very much the officer and those elsewhere potentially at Fujitsu had in mind for you to be the author of that statement but, ultimately, it was Penny Thomas that provided the statement.
Rajbinder Sangha: Yes, correct, yeah.
Mr Blake: If we go back to page 1 of this document, the statement that we’ve just been looking at, there’s just a paragraph on the first page that I would like to take you to. So I’m just going to read that second substantive paragraph on this page. It says:
“The data requested was originally supplied to Post Office Limited on 26 August 2010 by Rajbinder Bains. I have reviewed the archived ARQ data extracted by Rajbinder Bains and confirm that the data provided was extracted from the Horizon system in accordance with the requirements of ARQs 226 to 228/1011 and that the extraction process followed the outlined procedure. I produce a CD containing the results and exhibit them. The CD contains a certified true copy of the original transaction data supplied in August 2010. This data has been held securely on the audit data workstation since its original extraction and contains no additions, deletions or other amendments.”
Would you agree that it’s quite unusual for somebody who didn’t actually download the data to be using this form of words within a witness statement, to be exhibiting data as having been, for example, extracted in accordance with the requirements and following a certain procedure, when, in fact, she wasn’t the one who did the downloading or extraction? Does that strike you as unusual at all?
Rajbinder Sangha: Looking at it, yes.
Mr Blake: Yes, how, for example, could she have said that it was extracted in accordance with procedure? Do you think she was well placed to make that statement?
Rajbinder Sangha: Well, it would have been a process that I would have followed to extract the information, so …
Mr Blake: Because if we scroll over and go to the standard wording that I’ve – that we’ve been through a couple of times, at the bottom of the second page into the third page, that standard wording again that I think your evidence has been that it’s not a form of words you would have signed up to, that there’s no reason to believe that the information is inaccurate because of the improper use of the system and that, to the best of her your knowledge and belief at all material times, the system was operating properly.
Did you have a conversation with Penny Thomas in advance of her signing this statement about the system, about the process that you undertook to obtain the ARQ data?
Rajbinder Sangha: No, no, I wouldn’t have.
Mr Blake: Because I think in your witness statement, paragraph 21 – I can take you do it – I think you say you didn’t have a conversation with Penny Thomas –
Rajbinder Sangha: No.
Mr Blake: – about the contents of witness statements?
Rajbinder Sangha: No, because I never – that was not part of my role, to deal with witness statements.
Mr Blake: Do you see it as in any way odd or unusual that Penny Thomas would have signed a statement that said that the system was operating properly, exhibiting the ARQ data that you had extracted, but not having any conversations with you about that data?
Rajbinder Sangha: No.
Mr Blake: You don’t see that as unusual?
Rajbinder Sangha: No.
Mr Blake: Why not?
Rajbinder Sangha: It was a process that I followed and I was not involved in witness statements, so I just assumed it was part of the process that whatever had been extracted would be included in there, into the witness statement.
Mr Blake: I’d like to look at another witness statement – sir, I’m just looking at the time. Perhaps this is a good moment for our mid-morning break. We will be finished by lunchtime today.
Sir Wyn Williams: Yes, certainly that’s fine, Mr Blake. When do you want to start again?
Mr Blake: Thank you very much, if we start in 15 minutes’ time, so 11.35, please.
Sir Wyn Williams: Yes, certainly.
Mr Blake: Thank you very much.
(11.18 am)
(A short break)
(11.35 am)
Mr Blake: Thank you, sir, can you still see and hear me?
Sir Wyn Williams: Yes, I can, thank you.
Mr Blake: Thank you. I’m going to take you to another witness statement and that’s FUJ00123054. This is the statement that you were the witness for. If we scroll down to the bottom your signature is at the bottom, it’s covered over by our redactions that obscure signatures but this is a document referred to in your witness statement as the one that you were the –
Rajbinder Sangha: Yes.
Mr Blake: – witness. If we scroll up, please, it’s 14 October 2010, so it’s five days before the statement that we were looking at before the break.
Sir Wyn Williams: Sorry, Mr Blake. This says this statement is of Ms Thomas.
Mr Blake: Yes, it is. So Ms Thomas wrote the statement and, if we look down at the bottom there is a signature of both the author of the statement and also the person that witnessed the statement.
Sir Wyn Williams: I’m with you, thank you.
Mr Blake: On the bottom, on the right-hand side, although it’s covered over Ms Sangha’s signature is on the right-hand side at the bottom.
So this was five days earlier. If we scroll down we can see again it’s the same pro forma witness statement. If we look at page 5, please, this relates to a branch at Wattville Road. We can see that on page 5, in the first substantive paragraph there, near the top, it says Wattville Road branch. Then about halfway down that page, it says:
“The requested data for ARQs 197-199/1011 was originally supplied to Post Office on 26 August 2010 by Rajbinder Bains. I have now reviewed the archived ARQ data which was extracted [by you] and confirm that the data provided was extracted from the Horizon system in accordance with the requirements … and that the extraction process followed the outlined procedure.”
If we continue down near the bottom of the statement, so pages 6 into 7, we have that form of words that we saw in those other statements. It’s the bottom of page 6 and 7, thank you.
Can I just take you to your witness statement and that’s WITN10300100. The statement we’ve just been looking at refers to data that you had extracted but Penny Thomas provided the witness statement for –
Rajbinder Sangha: Correct, yeah.
Mr Blake: – and you witnessed her simply signing the statement itself. It’s paragraph 21, page 9 of your witness statement, please. Thank you. You’re referring here to the statement that we’ve just looked at, and you’ve said:
“… the Inquiry has asked (i) to what extent witness statements were ‘generic’ or fact specific, (ii) whether I read or discussed the content of witness statements that I witnessed, (iii) any concerns I had about the contents … (iv) any communications I had with Ms Thomas … (v) any concerns Ms Thomas may have had regarding her role … I understood that witness statements were fact specific based on what [the Post Office] had requested. To witness Ms Thomas signing her statement, I was not expected to read the contents of the witness statement. I cannot now recall having discussions or concerns about the contents of her witness statement that I witnessed. I did not have any discussions with Ms Thomas in relation to the accuracy of witness statements nor any concerns she had about her role as a witness in court proceedings.”
Now, where you refer to “fact specific based upon what the Post Office had requested”, I think we’ve seen today that there was a basic pro forma statement –
Rajbinder Sangha: Yes.
Mr Blake: – that was adapted to the specific case. In terms of a conversation with Ms Thomas, again, in relation to a statement that you had both witnessed the signature of and also that referred to data that you had extracted, do you think it’s unusual in any way that there wasn’t a conversation between yourself and Ms Thomas about the accuracy or the contents of that statement?
Rajbinder Sangha: No, looking back on it, no.
Mr Blake: No?
Rajbinder Sangha: No.
Mr Blake: Can we please look at FUJ00171848, please. This is another PEAK. It’s FUJ00171848. When we get to it, it’s a PEAK that begins on 27 October 2010. So we have the Wattville Road statement dated 14 October 2010. We’ve got the Preston Road statement that we’ve also seen, 19 October 2010, both with the same form of words at the end of both of those statements.
We’re now 27 October, so very shortly after both of those statements were signed and we have an error log PC0205805 and it is summarised as follows, if we scroll down:
“The Fast ARQ interface does not provide the user with any indication of duplicate records/messages.
“This omission means that we are unaware of the presence of duplicate transactions. In the event that duplicates are retrieved and returned to [the Post Office] without our knowledge the integrity of the data provided comes into question. The customer and indeed the defence and the court would assume that the duplicates were bona fide transactions and this would be incorrect. There are a number of high-profile court cases in the pipeline and it is imperative that we should provide sound, accurate records.”
If we scroll down, please, to the second entry, 27 October, we have an entry from Penny Thomas. She says as follows, she says:
“A Fast ARQ interface does not provide the user with any indication of duplicate records/messages.”
I think it’s the same form of words that we’ve just seen.
Rajbinder Sangha: Yes.
Mr Blake: Then there’s an entry from Mr Barnes at the bottom. It says:
“Andy and I have looked at this. We think the method most compatible with existing behaviour is as follows …”
If we scroll on to the next page, please. Thank you. The top entry there, please, thank you:
“Check for duplicates for [Horizon Online] in a similar method to how duplicates are checked for in Horizon.”
I think that’s a reference to Legacy Horizon:
“For Horizon, they are legitimately logged in the audit log and then ignored (because it’s just that identical messages are stored by mistake in more than one transaction file). For [Horizon Online], in the Fast ARQ case, their detection will cause them to be logged in the QueryLog and a count kept of how many there are”, et cetera.
He sets out further detail there.
If we scroll down, please, to the entry of 5 November, thank you, if we could zoom out and look at the final entry on that page, so that final box. Mr Barnes provides a technical summary, and he says as follows:
“HNG-X [Horizon Online] can rarely produce transactions with duplicate journal sequence numbers. At the moment when running a [I think that’s a Fast ARQ] on the audit server, these duplicates are not noticed. This means that the evidence presented by the prosecution team may show duplicate transactions without being noticed. The defence team may spot this and call into question the integrity of our data.”
If we scroll down “Impact on user”:
“Horizon Online transactions with duplicate JSNs may not be noticed. This will call into question the reliability of evidence presented by the prosecution team.
“Impact on operations:
“The prosecution evidence will be more consistent and so prosecution cases will go through more smoothly.”
Over the page. Thank you very much. If we could highlight that top entry, please:
“Have relevant [Known Error Logs] been created or updated?”
It says as follows:
“It was not felt that a [Known Error Log] was required because there are only two people in the prosecution team and they are both fully aware of the problem.”
That reference to two people in the prosecution team, are you one of those two people?
Rajbinder Sangha: I can’t remember. I don’t –
Mr Blake: I think you described there being three people in the team itself?
Rajbinder Sangha: Yes.
Mr Blake: Were two people, at that stage – so we’re 5 November 2010, were there two people who particularly dealt with these kinds of things?
Rajbinder Sangha: Yes, there were.
Mr Blake: Who were those two people?
Rajbinder Sangha: Andy and Penny.
Mr Blake: The reference there to not creating a Known Error Log because only two people were in the team and they were fully aware of it, was this was something that you were fully aware of, being the third person in the team or a third person in the team?
Rajbinder Sangha: I can’t remember.
Mr Blake: If we look at the “Risks” it says:
“If the fix is not released then duplicate [Horizon Online] transactions will continue not to be noticed by the prosecution team which will call into question their evidence.”
Then it says:
“There are no particular problems [with] the fix.”
Does this not seem quite a significant issue for your team?
Rajbinder Sangha: Yes, it is.
Mr Blake: Can you assist us with why a Known Error Log was not created in relation to this issue?
Rajbinder Sangha: I don’t know. It wouldn’t have been part of my job role. It would be dealt with senior levels of management.
Mr Blake: Would it have been helpful to you to have been briefed on this particular issue?
Rajbinder Sangha: Yes, it would have.
Mr Blake: Yes. You will recall –
Sir Wyn Williams: When you say – sorry, when you say “senior levels of management”, I may have been anticipating Mr Blake’s next question but do you mean by that Mr Dunks and Mrs Thomas or do you mean persons higher in the hierarchy than them?
Rajbinder Sangha: Persons higher.
Sir Wyn Williams: So could you give us an indication, if you can remember, who those persons were?
Rajbinder Sangha: The likes of Donna Munro. So, at the time, she was my boss.
Sir Wyn Williams: Right.
Sorry, Mr Blake.
Mr Blake: Not at all, sir. Thank you.
Was Donna Munro involved in issues relating to the integrity of ARQ data, from your recollection?
Rajbinder Sangha: I don’t know whether she was or not. I don’t know.
Mr Blake: Was she somebody who was quite hands-on in terms of the management of the team or hands-off?
Rajbinder Sangha: No, she was hands-on. She knew what was happening within the team and –
Mr Blake: This kind of an issue that Penny Thomas was involved in, would you have expected that to have been discussed between Penny Thomas and Ms Munro?
Rajbinder Sangha: Yes. It would have been, yes.
Mr Blake: The reference there by Mr Barnes to “the defence team might spot this”, can you assist us with what the attitude within your team might have been to defence teams in prosecutions based on Horizon data; is this a typical comment that you might find expressed within the team?
Rajbinder Sangha: I’ve never heard that expression within the team, so I don’t know what the thoughts would have been of the team members.
Mr Blake: You’ll recall paragraph 21 of your statement that I took you to, where you said that you didn’t have discussions with Ms Thomas in relation to the accuracy of her witness statements. Looking at this now, does that cause you some concern about the witness statements that were submitted?
Rajbinder Sangha: Yes, it does.
Mr Blake: Yes. Can you think of a reason why it may have been that that wasn’t subject to discussion between the two of you?
Rajbinder Sangha: The only reason I can think of is maybe because I didn’t actually go to the court proceedings. So it wasn’t discussed with myself.
Mr Blake: Given that there were only three of you in the team, you all shared a room, it was a secure room that couldn’t be accessed by other people: is it surprising to you that this kind of an issue wasn’t more thoroughly discussed?
Rajbinder Sangha: Yes, yes it is.
Mr Blake: Can we please look at FUJ00155517. We’re now moving to 2011, February 2011. Now, there is an email that is forwarded to you. We will go through the email. The email is actually dated 14 September 2010 and it’s about an issue with Horizon Online. Are you able to assist us at all as to why this was sent to you or copied to you in 2011? I can read the top entry is to Andy Dunks, and it says:
“As you’re starting to cover audit you need to be aware of the situation detailed in this mail string.”
Do you know why you’re copied there in 2011?
Rajbinder Sangha: I think it was for information purposes, so that I was aware that there was a situation.
Mr Blake: If we scroll down we can see it’s an email from Tom Lillywhite, who is described on the final page as Principal Security Consultant. This email states as follows:
“Our RMG Account Fraud and Litigation Service are currently acting on an ARQ … This request for transactions records, which covers March 2010, is in respect of an outlet … which is already migrated to Horizon Online.
“Because of a number of technical issues (errors detected) that arose during migration up to June 2010, and which [Post Office] technical specialists are aware of, the information gathered in respect of this particular ARQ may be subject to issues of integrity. Our technical staff have investigated the record in question and, at this stage, although they report that there is no obvious evidence of suspicious behaviour, they can add nothing further with any certainty and they do not have the ability to determine if there really are any financial implications with the messages. In other words, any response from us would have to bear the health warning that there was no guarantee as to the integrity of the data provided by us.
“The issue is of particular relevance in light of the fact that provision of an ARQ could result in a request for a Statement of Witness to support litigation activity. As such, any statement of witness provided would, in real terms, have to reflect this.”
So there seems to be the identification of a particular issue that impacts on the integrity of data. Was it common for these kinds of issues to be shared over email in this way? If we look at the top, it seems as though Penny Thomas is sending it to Andy Dunks, essentially a historic email, to update him as to issues relating to integrity because he’s taking up some sort of role relating to audit.
Rajbinder Sangha: Yeah, it was normal.
Mr Blake: That was normal?
Rajbinder Sangha: Yeah.
Mr Blake: So, in terms of notifying people of issues relating to integrity of the audit data, would you say it was more common to receive emails rather than there had to be some official documentation or notification of some sort?
Rajbinder Sangha: Yeah, emails.
Mr Blake: Thank you. Did this particular issue cause you any concerns?
Rajbinder Sangha: Not at the time, no, because I didn’t have that much knowledge of the system.
Mr Blake: Knowing now all the things that we’ve already been through, all of those various statements that were provided, does this cause you any concerns?
Rajbinder Sangha: It does now, yes.
Mr Blake: Can we please look at FUJ00229039. We’re now in April 2011, two months later. We can see that’s an email from Gareth Jenkins to yourself and perhaps we could go to page 3, please. Bottom of page 2, into page 3. Thank you very much.
There’s an email from Gareth Jenkins to somebody called Saheed and it says as follows – it’s in reference to a particular branch and it says:
“I’ve had a quick look at the info that Mark has already provided to [the Post Office] about the discrepancies issue last year and how it relates to this branch.
“It looks like this branch hit the problem twice …”
It then says:
“These put together result in a net discrepancy of [minus] £335.66.
“The data relating to the first period will now have gone from BRSS as will some of the data from the second period.”
Are you able to assist us with “BRSS”, what that means?
Rajbinder Sangha: No, I don’t remember, no. Sorry.
Mr Blake: At the bottom of the page, it says:
“Given that it is now clear the branch hit the bug twice (as did [another branch]), is that sufficient information for [the Post Office], or do we need to retrieve the audit data for the relevant periods and carry out a more complete analysis?”
Then if we go to the bottom of page 1 into page 2, we have the email from Gareth Jenkins to Penny Thomas and yourself. It says:
“Penny (or Raj if Penny’s still away),
“Saheed has asked me to do some investigations into this Branch and most of the relevant data will have been archived from BRSS by now.
“Therefore please can you retrieve the data as for an ARQ.”
So Mr Jenkins is getting in touch with you and Penny Thomas about a particular bug that’s affected a branch that’s caused a discrepancy and has asked for you to run an ARQ; is that right?
Rajbinder Sangha: Yes, that’s correct, yeah.
Mr Blake: It seems from these emails that there are a number of emails relating to bugs, errors or defects in the Horizon system. Can you assist us with why, at that particular time, you weren’t more concerned about the integrity of the system?
Rajbinder Sangha: Because I didn’t have much knowledge of the system, I just saw my role as to process the ARQ requests. I wasn’t involved in or didn’t have much knowledge of, the issues, obviously, that have been identified.
Mr Blake: I said a few times about the small number of people in the room, the room being secure, et cetera. People might see it as surprising that you didn’t have, you know, team meetings, discussions within the room, about problems generally with Horizon, the bugs that we’ve just been referring to that caused discrepancies, issues with the integrity of the ARQ data. Can you assist us with why that might be the case, that there were no discussions?
Rajbinder Sangha: Obviously, there were discussions but, obviously, I didn’t understand the severity of these issues, compared to other senior members of the team.
Mr Blake: You’ve been very candid today about now having concerns, looking back.
Rajbinder Sangha: Yes.
Mr Blake: In reality, though, did you have concerns during the period that we’ve gone over –
Rajbinder Sangha: No.
Mr Blake: – about the integrity of the system, the reliability of the ARQ data?
Rajbinder Sangha: No, I didn’t.
Mr Blake: Because there’s an email I want to take you to that you’ve seen, it’s FUJ00156518, please. It’s the email about the Swiss Cottage trial. If we have a look at the bottom page, there’s an email from Graham Brander to Penny Thomas and Andy Dunks, and it says:
“I still haven’t had confirmation as to whether either/both of you are required.”
It’s for a particular trial relating to, it seems, Swiss Cottage post office:
“The provisional batting order for witnesses … shows Penny being required on day 2 … and Andy on day 3 …
“Andy: can you confirm that if required, you are able to cover Penny’s evidence in case she is unable to attend.
“Penny: I have asked our legal person that if you are required that if possible can you give evidence on day 1 but I am still awaiting a definitive response.”
If we scroll up, please. The next email says, it’s from Andy Dunks to Graham Brander and Penny Thomas:
“As far as standing in for Penny, I do not think my knowledge is good enough to be able to do this to the level that Penny could answer any questions. I would not be comfortable to do it.”
Then it’s the first email in this chain the top email of the chain that we are concerned with, which is from Penny Thomas to Donna Munro, so who you’ve explained was effectively a line manager of the team.
Rajbinder Sangha: Correct, yes.
Mr Blake: She says:
“Donna
“It would appear you have a gaping hole as far as prosecution support is concerned; Andy being uncomfortable to support a basic statement and Raj not being prepared to submit one.”
It seems, from this email, a reading of this email, is that what Penny Thomas is doing is distinguishing two different positions: the one of Andy, who said that he doesn’t have enough knowledge to be able to support the statement; and, on the other hand, yourself, not being prepared to submit one, suggesting that you had some real concerns about submitting a statement. Did you have real concerns about submitting a statement?
Rajbinder Sangha: My concern was not feeling comfortable to submit a witness statement about a system that I didn’t have knowledge about, proper knowledge about.
Mr Blake: Yes. Did you think that those within your team that sufficient knowledge to submit them?
Rajbinder Sangha: Yes, yes they did.
Mr Blake: Did you, during the period that we’ve looked at – as I’ve said, you’ve been very open about your concerns now – was it in no way to do with those kinds of concerns that you weren’t prepared to submit a witness statement back in 2010, 2011, 2012?
Rajbinder Sangha: No, it wasn’t. I just was not comfortable in producing a witness statement.
Mr Blake: Can we please look at a later bug, and that is FUJ00226106. We’re now in 2013. At the bottom of this page, we have an email from Gerald Barnes to CSPOA Security. Who is CSPOA accurate?
Rajbinder Sangha: It’s another – it’s the same team as the Fraud and Litigation Support.
Mr Blake: Then copied to yourself, Andy Dunks and others. He says as follows:
“A serious flaw has recently been spotted in the audit code. It was introduced in the fix to [and it gives a PEAK reference] quite some time ago (but post [Horizon Online]). There is a small possibility of missing transactions on generated spreadsheets if the query handling was run during the evening Query Manager shut down.”
Then he says, for example, a little further on:
“You may consider it necessary to check the previous ARQs run.”
He describes what’s called a serious flaw –
Rajbinder Sangha: Yes.
Mr Blake: – that goes back some time ago. If we scroll up we have Andy Dunks responding, copied to you. He says:
“Gerald,
“Can you confirm that we’re talking as far as back as September 2009?”
Actually, if you look at the PEAK, it seems to be closed in 2009, so certainly it seems to have been closed in 2009:
“Are you able to pop down and explain and show us what to look for, as we need to put together some time scales to complete this task.”
Then Gerald Barnes says:
“Hi Andy,
“I will come down in a few minutes.”
Is this something you remember at all?
Rajbinder Sangha: I do remember obviously being copied into this email and there were discussions but as to what those discussions, looking back on it now, I can’t remember.
Mr Blake: Because there’s no answer in that email about the 2009 issue?
Rajbinder Sangha: No.
Mr Blake: There’s just the query?
Rajbinder Sangha: Yeah.
Mr Blake: “Can you confirm that we’re talking as far as back as September 2009?” The response is “I’ll come down”; was that coming down to your team?
Rajbinder Sangha: Coming down to the secure room to show us.
Mr Blake: Yes, and you can’t recall significant discussions about this issue at all?
Rajbinder Sangha: No.
Mr Blake: No. We’ve seen throughout the course of today your name on various emails, often copied in –
Rajbinder Sangha: Yes.
Mr Blake: – about certain bugs or certain issues with ARQ data and the integrity of the data. We’ve seen statements from Penny Thomas, exhibiting ARQ data that you had obtained and providing that standard form of words towards the end of the statement. You’ve, in at least one of those cases, witnessed the statement being signed?
Rajbinder Sangha: Yes.
Mr Blake: We have here, for example, serious flaws in the audit code being identified. Why do you think it was that that was not more of a topic of discussion within your small team?
Rajbinder Sangha: I don’t know.
Mr Blake: Do you think it would have been appropriate or not appropriate to have had those kinds of discussions?
Rajbinder Sangha: No, it would have been appropriate, yes.
Mr Blake: The Chair is going to be considering recommendations for this Inquiry in due course. So it would be helpful to know what you think went wrong in terms of the lack of knowledge that you had about the integrity of the data that you were providing for use in court proceedings. Are you able to assist us with that?
Rajbinder Sangha: Sorry, can you repeat that again?
Mr Blake: If the Chair is trying to understand what went wrong, why wasn’t there this discussion about the integrity of Horizon data? What can we learn from this experience? What is it that you think, what is the reason you think for not being sufficiently informed or for there to have been insufficient discussions about what received like quite serious problems with the integrity of the audit data that’s being provided for prosecutions?
Rajbinder Sangha: Sorry, it’s gone over my head.
Mr Blake: Is there anything you can think of that you didn’t do as a team that would have been better, had you done it?
Rajbinder Sangha: Maybe obviously have more conversations about these issues that were identified and it being filtered down to everyone within the team.
Mr Blake: Can we look at FUJ00229905, please. This is a final email chain that I’ll take you to that relates to bugs, errors or defects in the Horizon system. We’re now on 13 May/14 May 2013. If we could go over the page, please, and start on page 2. From Gareth Jenkins to yourself:
“Raj,
“There was a significant bug on [Horizon Online] that affected about 14 branches that was identified earlier this year that is causing some interest with high levels of management in [the Post Office] and Fujitsu.
“Anne Chambers has done some detailed analysis based on information in BRDB/BRSS, but I think it would be helpful to pull back the transaction logs for the relevant periods to confirm the analysis.”
If we go over the page, the first page, there’s a response from you to Gareth Jenkins. You say as follows:
“Hi Gareth,
“No problem – I have set the retrievals off now and should be able to get all the information by Friday if not earlier however I am not in the office until Wednesday and Thursday we have an audit taking place so not sure whether I will have time, however I will try to get this to you [as soon as possible].”
Then you have an email from him to you:
“Raj,
“Please can you retrieve similar info for the following branches …
“This work should be chargeable to the ‘Second Sight’ CP I mentioned yesterday. If there is no more covering on that then please let me know and I’ll take it up with Pete Newsome.”
So we have an email from Gareth Jenkins on 13 May that refers to some interest with high levels of management in the Post Office and Fujitsu, we have a reference there to Second Sight; were these not topics that were discussed within your team?
Rajbinder Sangha: With myself, no, they weren’t.
Mr Blake: Did it not stand out that high levels of – you were being told in an email that there’s something of interest to high levels of management within the company?
Rajbinder Sangha: At the time no, it wouldn’t have.
Mr Blake: Was it known within the company that there were concerns with management and that related to, in some way, the job that you were performing or the task that you were performing?
Rajbinder Sangha: No.
Mr Blake: No. Can we please look at POL00333387. We’re now in 2014, August 2014. I think this is a meeting that you don’t actually recall but I just want to very briefly take you to it. So this is Alan Simpson from the Post Office sending you a draft agenda for a meeting.
Then if we go to that agenda, POL00333391, we have there some attendees from Fujitsu, some attendees, some invitees, from the Post Office. It looks as though you are there to talk to the Post Office about the ARQ process; is that right?
Rajbinder Sangha: Yes.
Mr Blake: Yes.
Rajbinder Sangha: That’s what this is showing, yes.
Mr Blake: I mean, given that you seem to at least by 2014, have been a contact with the Post Office about the ARQ process, do you think that you were sufficiently informed about problems with the ARQ process that stage?
Rajbinder Sangha: Yes, I had.
Mr Blake: Did you have discussions with the Post Office about concerns relating to the integrity of –
Rajbinder Sangha: Oh, no.
Mr Blake: – of data?
Rajbinder Sangha: No, I didn’t.
Mr Blake: Do you think you had enough information – being a person who was meeting with the Post Office, do you think you had been provided with enough information internally relating to concerns about the integrity to be able to have passed those on?
Rajbinder Sangha: No, no.
Mr Blake: Why not?
Rajbinder Sangha: Because of – obviously, all these issues that have been identified, and not having the technical knowledge of the system, but obviously in what context this meeting – this agenda is about, I can’t remember what was discussed that my name is against the ARQ process, I don’t – I can’t recall.
Mr Blake: I’d just like to take you to one final document today and that’s FUJ00180028. We’re now in 2018 so quite a while later. There seems to be a general communication that’s sent out. It states as follows:
“Post Office Limited are currently involved in a class action brought by a group of current and former subpostmasters which is due in court in November 2018 and March 2019. During this period, it is possible that Freeths Solicitors, who are representing the subpostmaster group, may attempt to contact Fujitsu staff. If you receive such contact please do not respond, but instead pass the details of the contact to Pete Newsome or Chris Jay.
“Thank you for your cooperation in this matter.”
It’s from Garry Stewart, the Delivery Executive Post Office Account. So he was somebody who worked at Fujitsu but was responsible for the Post Office Account; is that right?
Rajbinder Sangha: Correct, yes.
Mr Blake: At the top, you seem to be forwarding it to somebody called Victoria Griffin; do you recall who she was at all?
Rajbinder Sangha: Yes, she was a former colleague.
Mr Blake: Was she senior to you, junior to you, same level?
Rajbinder Sangha: She didn’t work for part of the Fraud and Litigation Support, no.
Mr Blake: Can you recall why it may have been that you forwarded to her at all?
Rajbinder Sangha: Just for information.
Mr Blake: Yes. Looking at the contents of that, relating to potential contact, “If you receive such contact … do not respond”, do you think that within Fujitsu people were able to speak freely about problems with the Horizon system?
Rajbinder Sangha: Yes, they were.
Mr Blake: Yes. Do you think that the members of your team were able to speak sufficiently freely about the integrity issues relating to the ARQ data?
Rajbinder Sangha: Yes, they were, yeah.
Mr Blake: In those circumstances, why do you think it is that, within your very small team, you didn’t build up sufficient knowledge to be able to raise any concerns yourself?
Rajbinder Sangha: At the time, I didn’t see a concern. So I didn’t see a concern at the time.
Mr Blake: Why did all those emails, email after email after email, not cause you some concerns?
Rajbinder Sangha: Because it would have been dealt with by more senior technical people that had more of an understanding. I just saw my role as just processing ARQs.
Mr Blake: Thank you.
Sir, I don’t have any further questions. There are questions from Mr Stein and also from – just from Mr Stein, I think.
Sir Wyn Williams: Right, over to you, Mr Stein.
Questioned by Mr Stein
Mr Stein: Thank you, sir.
Ms Sangha, I represent a number of subpostmasters and mistresses and I’m instructed by a firm of solicitors called Howe+Co. Now, I’ve got a few questions for you that arise out of your statement –
Rajbinder Sangha: Okay.
Mr Stein: – and one document that I want to put to you from the background material. Can I take you to your statement, which is WITN10300100 and paragraph 7 of that at page 3, please. Thank you.
Now, at this part of your statement, Mrs Sangha, you identify the period of time that you were with Fujitsu, as you still are, as a Fraud and Litigation Support Officer in the Fraud and Litigation Support Office, okay? We can see that, if we go down to paragraph 8, if you would, please, same page, that you went on leave in April 2016 and, upon returning from leave in January 2017, you commenced in a new role.
So between those two paragraphs we’ve identified within your statement where you are saying you joined in July 2010 the Fraud and Litigation Support Team, and then you go on leave in April 2016 and you come back from leave in January ‘17 and you’re in another role?
Rajbinder Sangha: Correct, yes.
Mr Stein: All right.
So when you join the Fraud and Litigation Support Team and part of your tasking is to provide ARQ data for litigation conducted by the Post Office, did you have an understanding that the provision of such data was a serious job that could have serious consequences for subpostmasters?
Rajbinder Sangha: No, I wasn’t.
Mr Stein: Now, as time has evolved and, in particular, as you may have become aware of the issues involved in this Inquiry, you probably have seen that this was a very serious job with real life serious consequences for subpostmasters; is that fair?
Rajbinder Sangha: Yes, that is fair. Yes.
Mr Stein: So when you joined the Fraud and Litigation Support Team, were you given any training as to the need for you to be as complete and as accurate as possible, or to be as diligent as possible, to make sure that the data provided was as good as possible?
Rajbinder Sangha: So the only training that I received was from fellow team members.
Mr Stein: Was there any training as to how ARQ data might be used within the court process?
Rajbinder Sangha: No.
Mr Stein: Now, of course, you do understand from all of the publicity created by the Inquiry and indeed by the decades of campaigning that subpostmasters have conducted over the years, that, because there were bugs in the system, innocent people have gone to prison –
Rajbinder Sangha: Correct.
Mr Stein: – or have lost their livelihoods –
Rajbinder Sangha: Yes.
Mr Stein: – lost their houses, homes, lost their ability to look after themselves and their families, whether that’s through prosecutions or court actions taken in the civil courts. You understand that?
Rajbinder Sangha: Yes, I do, yeah.
Mr Stein: Okay. Now, you’ve discussed matters with the gentleman that’s been asking you questions this morning, Mr Blake, about how the system worked. So can we piece that together. The ARQ data was something that you supplied on request by others; is that right?
Rajbinder Sangha: That’s right, yes.
Mr Stein: You’ve also discussed with Mr Blake the fact that you and others in your team were aware that there were bugs or issues within the Horizon system, yes?
Rajbinder Sangha: Yes.
Mr Stein: Okay. Within your statement, if we can now go, please, to paragraph 24 – and, to assist, that’s at page 10, so page 10, paragraph 24 – you’ll see there that you’re saying:
“The Inquiry has asked me to set out any other knowledge I had in respect of bugs, errors and defects in Horizon (including Legacy and Horizon Online) [and, secondly] how I consider they impacted the work I carried out, and … any concerns I may have in respect of their impact on [Post Office’s] litigation or criminal proceedings in which [you] played a role.”
You go on to say this:
“I do not have any specific recollections of bugs, errors or defects in Horizon, however I would have been aware of them during my time in the CSPOA Security Team.”
Okay?
Rajbinder Sangha: Yes.
Mr Stein: All right. So you provide the ARQ data on request?
Rajbinder Sangha: Correct.
Mr Stein: Others, that’s Mr Jenkins and Ms Thomas, provide witness statements in relation to the operation of the Horizon system, yes?
Rajbinder Sangha: Correct.
Mr Stein: Where there are bugs, errors and defects within the system, was it your understanding that Ms Thomas and Mr Jenkins would provide any information that was required for court cases about how they may have had an impact on the operation of the system?
Rajbinder Sangha: I don’t know.
Mr Stein: You’ve seen statements, draft statements and statements – you’ve been shown those by Mr Blake this morning.
Rajbinder Sangha: Yes.
Mr Stein: You’ve seen statements, therefore, that have been given by Ms Thomas and Mr Jenkins, yes?
Rajbinder Sangha: Yes.
Mr Stein: You are aware that their roles were to provide, if you like, the court-facing evidence, yes?
Rajbinder Sangha: Yes.
Mr Stein: Okay. So it may be I asked a bad question and, if so, I’ll have another go at putting it, all right? Would you have expected those witnesses to provide any other information about problems in the system that may be useful for the court process?
Rajbinder Sangha: Yes.
Mr Stein: Okay. So we’ve got, really, two things: ARQ data that you’re providing, yes?
Rajbinder Sangha: Yes.
Mr Stein: And then an overlay of any other information that may be necessary for the court process?
Rajbinder Sangha: Yes, definitely.
Mr Stein: So that would mean that, let me see, Ms Thomas and Mr Jenkins, if they believed that there were – how can I put it – a bunch of problems with the system that the defence need to know about, you’d be expecting them to provide that information?
Rajbinder Sangha: Yes.
Mr Stein: All right. In the same way, you’d be expecting them to provide information that said “Well, look the system does this, but there are these problems”; is that fair?
Rajbinder Sangha: Yes.
Mr Stein: Okay. Now, during the court process, when individual subpostmasters and mistresses went to the Criminal Court of Appeal, Simon Clarke advices were revealed at that stage and one of them mentioned the question of the integrity or the credibility of Mr Jenkins, and that was in around 2013. Were you ever asked any questions about whether Mr Jenkins was a person that was providing honest evidence?
Rajbinder Sangha: No.
Mr Stein: Was it ever brought to your attention that there was a question about his credibility?
Rajbinder Sangha: No.
Mr Stein: Was it ever brought to your attention that there was any issue with his accuracy, in other words getting things right?
Rajbinder Sangha: No.
Mr Stein: Was there any investigation that you’re aware of within Fujitsu of Mr Jenkins’ accuracy, credibility and integrity?
Rajbinder Sangha: No.
Mr Stein: Was that ever brought to your attention?
Rajbinder Sangha: No.
Mr Stein: Rather than going through all of that regarding Ms Thomas, the same questions for Ms Thomas?
Rajbinder Sangha: Yeah. The same. It would be the same: no.
Mr Stein: Now, I’m going to take you to a document which is going back to 2010. As we’ve just discussed, that’s when you joined this particular team, okay? The document is POL00028838. Now, you’ll have that on your screen. I’m afraid the document is a little grey, if you like. It’s sort of a little hard to read because it’s perhaps been copied too many times, or something similar. All right, but you can see there that it’s headed “Receipts/Payments Mismatch issue notes”, okay?
Then we see attendees at this particular meeting and a variety of names there. Can we go to the Fujitsu names? You’ll see there at the bottom Mike Stewart, was he someone that you were familiar with?
Rajbinder Sangha: I know the name.
Mr Stein: In what capacity?
Rajbinder Sangha: That he was a Service Delivery Manager, sorry, but I didn’t have no interactions with him.
Mr Stein: John Simpkins?
Rajbinder Sangha: Yes.
Mr Stein: Dealings with him?
Rajbinder Sangha: No, not really. I know that he worked in the SSC Department.
Mr Stein: Right. Now, regarding your own line management, were these within your own line management? Were they people that were above Ms Munro or sideways, working with her? Can you recall where they existed, if you like, on an organogram?
Rajbinder Sangha: I wouldn’t say they were above her. They were to the side.
Mr Stein: To the side?
Rajbinder Sangha: Yes.
Mr Stein: Now, Gareth Jenkins we can jump over him because we’ve discussed that. He is within your statement as someone that you were aware of in the way we’ve discussed. Then underneath that was Mark Wright, Fujitsu Technical Specialist; someone you came across?
Rajbinder Sangha: No.
Mr Stein: All right. Then it goes on to say, “What is the issue?” So the period of time we’re talking about in relation to this, we’ve looked at variously during the course of this Inquiry, this is 2010, we think around August 2010, all right:
“What is the issue?
“Discrepancies showing at the Horizon counter disappear when the branch follows certain process steps, but will still show within the back end branch account. This is currently impacting circa 40 branches since migration onto Horizon Online with an overall cash value of circa £20k loss. This issue will only occur if a branch cancels the completion of the trading period, but within the same session continues to roll into a new balance period.”
Then you’ll see that underneath that:
“At this time we have not communicated with the branches affected and we do not believe they are exploiting this bug intentionally.”
Now, top of the next page, please, so second page on Relativity, where it says – right, if we look at the top there, it says:
“Note at this point nothing into [that’s the word it says there] feeds POLSAP and Credence, so in effect the POLSAP and Credence shows a discrepancy, whereas the Horizon system in the branch doesn’t. So the branch will then believe they have balanced.”
Then just lastly, in relation to this, look at “Impact”, please, at the bottom of that particular page, so the second page within Relativity, “Impact”. If you can highlight that, please, I’d be very grateful. Thank you.
Under “Impact”, so what does this mean, is where we’ve now got to within this document:
“The branch has appeared to have balanced, whereas in fact they could have a loss or a gain.
“Our accounting systems will be out of sync with what is recorded at the branch.
“If widely known, could cause a loss of confidence in the Horizon system by branches.
“Potential impact upon ongoing legal cases where branches after disputing the integrity of Horizon Data.
“It could provide branches ammunition to blame Horizon for future discrepancies.”
Okay? It may be a bit of an understatement but the impact looks pretty significant?
Rajbinder Sangha: Yes, it does.
Mr Stein: Okay. Now, as we’ve looked at within your statement, you came into this particular part of work within Fujitsu in July 2010?
Rajbinder Sangha: Yes.
Mr Stein: You’ve said within your statement that you were made aware – paragraph 24, we’ve discussed this before – you would have been made aware of bugs, errors or defects in Horizon, is what you’ve said then. So, help us with this particular one, this submarine bug, which can’t be detected by branches, a significant one, you agree; a serious one, you agree?
Rajbinder Sangha: Yes.
Mr Stein: Was that something you were alerted to in 2010, not long after you came into this particular post?
Rajbinder Sangha: I can’t remember.
Mr Stein: If you had been, would this have caused you some concerns?
Rajbinder Sangha: Oh, definitely, yes.
Mr Stein: Yes. You’ve mentioned, in relation to Mr Blake’s questions today – a number of times he asked you that sort of question, which is: “Would this have caused you concern?” Who should, within the system that you were working within – within Fujitsu, who should have kept you up-to-date with these things? Who should have been talking to you about these things? In other words, who was responsible for not dealing with these matters and discussing them with you?
Rajbinder Sangha: I would say management.
Mr Stein: Management. From your perspective, would that have been Ms Munro or would it have been others above her?
Rajbinder Sangha: I would say others above.
Mr Stein: Why not Ms Munro, just out of interest? Why not her as well, if you like?
Rajbinder Sangha: Her as well, yeah.
Mr Stein: Fine. All right. I know this is difficult because, of course, you’re still working for Fujitsu, so we all understand that.
Now, when you left that particular team – we’ve looked at this before, paragraph 8, “I went on leave in 2016”, and you then came back into a different role, okay?
Rajbinder Sangha: Yes.
Mr Stein: All right. By the time you’ve gone on leave in April 2016, who’s taking over from your role? In other words, the role of providing data, ARQ data, for potential prosecutions or civil actions by POL?
Rajbinder Sangha: So at that time, our team had expanded because, obviously, we had taken on more work. So there were other people within the team, as well.
Mr Stein: Right. So when you left, there was other capacity –
Rajbinder Sangha: Yes.
Mr Stein: – in the people capacity –
Rajbinder Sangha: Yes.
Mr Stein: – if I can put it that way? Right.
Now, we know that the Post Office stopped prosecuting cases itself, in other words taking people itself to courts, and the matters were left to the Crown Prosecution Service within England and Wales and other prosecution authorities in parts of the devolved jurisdiction, okay?
Cases are still being prosecuted of subpostmasters, this time by the CPS, using data from Fujitsu. Who is in charge of that? Who is dealing with that now?
Rajbinder Sangha: I don’t know.
Mr Stein: Now, we’ve looked at and examined the question – as has Mr Blake to a large extent, this morning, as well – about the level of bugs, errors, defects within the system that were brought to your attention, okay, and you’ve said in your statement that they would have been brought to your attention, and so on.
Were you aware that the staff members at Fujitsu that the capacity and ability to alter branch data –
Rajbinder Sangha: No.
Mr Stein: – remotely?
Rajbinder Sangha: No.
Mr Stein: Was that ever explained to you?
Rajbinder Sangha: No.
Mr Stein: You’re aware within all of these court proceedings that that’s a matter that has been discussed, the question of remote access into branches and altering branch data? You’re aware that that’s been discussed?
Rajbinder Sangha: It has – yes, I’m aware of it.
Mr Stein: That’s not something you’ve ever asked about?
Rajbinder Sangha: No.
Mr Stein: About your period of time, when you’re looking at these matters, providing data, you’ve never asked whether the data had been altered in any way by someone at Fujitsu?
Rajbinder Sangha: No.
Mr Stein: Are there people still in post within Fujitsu who were responsible for not making sure that you’re aware of these issues?
Rajbinder Sangha: No, they’re not in post.
Mr Stein: To your knowledge, is the Horizon system still being used to provide data that is used in court proceedings?
Rajbinder Sangha: Yes, I think it is.
Mr Stein: Thank you, Mrs Sangha.
Sir Wyn Williams: Thank you, Mr Stein.
Are there questions from anyone else?
Mr Blake: No, sir. That’s all for today.
Sir Wyn Williams: Well, thank you very much, Mrs Sangha, for providing a witness statement, and coming to the Inquiry to answer a good number of questions. I’m very grateful to you.
So we’ll adjourn now until 10.00 tomorrow morning.
Mr Blake: Thank you very much, sir.
(12.34 pm)
(The hearing adjourned until 10.00 am the following day)