1 Wednesday, 11 January 2023 2 (11.30 am) 3 MR STEVENS: Sir, can you see and hear me? 4 SIR WYN WILLIAMS: Yes, I can. 5 MR STEVENS: Than you, sir. If I may call Mr Bansal. 6 SIR WYN WILLIAMS: Certainly. 7 STEVE BANSAL (affirmed) 8 Questions by MR STEVENS 9 MR STEVENS: Good morning, Mr Bansal. As you know, my name 10 is Sam Stevens and I ask questions on behalf of the 11 Inquiry. Please could I ask you to state your full 12 name? 13 A. Steven Bansal. 14 Q. Thank you for giving evidence today to the Inquiry, both 15 in the written form which we're about to turn to and 16 orally today as well. In front of you I see you have 17 a bundle of documents, at the front of which should be 18 you've witness statement. 19 A. It is. 20 Q. That is dated 9 August 2022 and runs to seven pages; is 21 that right? 22 A. That is correct. 23 Q. Have you had a chance to read that statement recently? 24 A. I have. 25 Q. Could I ask you please to turn to page 7 of that 1 1 statement. Is that your signature? 2 A. It is. 3 Q. Are the contents of that statement true to the best of 4 your knowledge and belief? 5 A. They are. 6 Q. Thank you. That statement now stands as evidence in the 7 Inquiry. I will now turn to ask you some further 8 questions on it. 9 Let's start with when you joined Peritas Limited. 10 That was in October '97 as a training instructor; is 11 that right? 12 A. That's correct. 13 Q. Now, just for background, Peritas Limited was a company 14 subcontracted by ICL Pathway to deliver training to end 15 users of the Horizon System; is that right? 16 A. I believe so, yes, yes. 17 Q. You say in your statement that you weren't solely 18 limited to the Horizon project but delivered training to 19 a number of different projects; is that right? 20 A. That's correct. 21 Q. And, because of the passage of time, your recollection 22 has faded and you cannot necessarily say which of your 23 recollections refer to the Horizon project itself and 24 which refer to others. 25 A. That is correct. 2 1 Q. I understand that you delivered on either basis 2 an agreed training plan to end users but you weren't 3 involved in designing the training programme itself; is 4 that right? 5 A. Not technically true. I did put together some of the 6 training that trained the trainers in preparation for 7 some of the Pathway -- 8 Q. I see. 9 A. -- training. 10 Q. So the people who would eventually go out to train the 11 end users, you helped put together documentation for 12 training those trainers? 13 A. Correct -- at the time. 14 Q. At the time. So that's '97? 15 A. '97. 16 Q. You say that in the delays to the rollout of the Horizon 17 IT System, because of that, you transferred to ICL 18 Pathway to work as a trainee tester; is that correct? 19 A. That is correct. 20 Q. So when the further offices started rolling out in '99, 21 with the national rollout in 2000, did you return as 22 a trainer to train people on rollout or did you remain 23 in testing? 24 A. No, I remained in testing. 25 Q. Did you have any knowledge of the training then after 3 1 you left Peritas Limited? 2 A. None at all. 3 Q. I'd like to refer to your witness statement, please. 4 The reference is WITN04770100. Please could we turn to 5 page 6, paragraph 19. Thank you. 6 In the second sentence it says that you received: 7 "... feedback reviews from my Peritas manager at 8 the time, which collated comments from subpostmasters in 9 respect of training. I also read the feedback forms and 10 requested feedback directly back from the attendees, as 11 it was important to me that the training had been 12 received and understood." 13 Just to clarify, the feedback you're referring to 14 there, is that personally how you delivered the training 15 or on the course as it was as a whole? 16 A. A combination of the two. So, if I recall -- and it is 17 vague memory -- there were effectively two forms for the 18 attendees to complete. One was on the training itself, 19 the content, duration, you know, was it technically 20 sufficient, and then the second was on the trainer, how 21 they delivered, what their technique was like, were you, 22 as an individual, comfortable with the information that 23 had been passed to you. 24 Q. But the forms you read they were the ones completed in 25 relation to training you delivered, rather than -- you 4 1 didn't look at other trainers' feedback forms? 2 A. No, it was purely a case of collecting the data attend 3 then perusing it before returning it to the office and, 4 indeed, at the time talking to the attendees. 5 Q. Please could we go to page 3 of the same statement and 6 paragraph 12. This describes the cash account or you 7 have put subheading "Cash Account". Then over the page, 8 the final sentence says that: 9 "... I do not recall having any experience of 10 working with the "Cash Account" software." 11 Does this mean that you didn't train end users on 12 how to use the EPOSS application or how to balance? 13 A. I don't believe so. I think at the time -- it's part of 14 I don't have an awful lot of recollection at all of what 15 the actual specific training was. I don't believe it 16 would have covered cash account at that time but I can't 17 say categorically no. 18 Q. I want to move on now then to testing. Your evidence is 19 that you transferred from Peritas to ICL Pathway, as we 20 say, as this trainee tester. Could I ask, at the time, 21 what qualifications in IT did you have? 22 A. At the time I did not have specific IT qualifications. 23 I think the position was that the rollout or the 24 training of the trainers was paused because the project 25 itself was at a pause. At the time, I was informed that 5 1 we were unsure whether that would be a three-month 2 pause, a six-month pause and, because of the information 3 and the training that I'd gathered, Pathway/Peritas made 4 the decision it would be useful if I were to support the 5 testing community because of some of the knowledge I'd 6 picked up. So, initially, I was there purely to support 7 and give a different perspective to the testing. 8 Q. Just to clarify, had you worked in IT as a tester prior 9 to that point? 10 A. No. 11 Q. Did you receive training from Pathway on your role as 12 a training tester? 13 A. I received on-the-job training. As I say, initially 14 I was there to support but then I ended up shadowing the 15 testers and gradually built my level of experience and 16 knowledge. 17 Q. I'd like to turn to a document. The reference is 18 FUJ00058375. This document is titled "Direct Interface 19 Testing Specification Pathway to HAPS". We will come to 20 the acronyms in a moment. 21 If we could just move down slightly on the screen, 22 please -- thank you -- at the bottom you'll see you are 23 the author. Do you recall writing this document first? 24 A. I vaguely recall writing it, yes. It was quite some 25 time ago but yes. 6 1 Q. On that "quite some time ago", apologies, if I could now 2 ask us to go a bit further up the document to the top, 3 we'll see the date is 3 February 1998. Now, in your 4 statement you say that you joined as a trainee tester in 5 April 1998 so you must have presumably joined the 6 testing team before then. 7 A. Formally, I think I joined -- effectively my contract 8 with Peritas ended. My new contract with Pathway 9 effectively started in April. Prior to that, I was 10 effectively on loan to the testing community. So I'd 11 been there for some time. 12 Q. Can you give any indication, just to place how long 13 you'd been in the testing team at this point. At this 14 point, roughly how long had you been working on testing? 15 A. I'm afraid I couldn't say. 16 Q. Please could we turn to page 5 of this document. The 17 introduction says that: 18 "This document details the direct interface test 19 specification between Pathway AP system ..." 20 Stopping there, that's the Pathway Automated 21 Payment System, isn't it? 22 A. That's correct. 23 Q. It goes on: 24 "... and POCL HAPS System." 25 That being Post Office Counters Limited Host 7 1 Automated Payment System? 2 A. Correct. 3 Q. When we are talking about the interface here, in simple 4 terms, are we saying what you're testing is how data is 5 transmitted from the Pathway Automated Payment System to 6 POCL's back end system? 7 A. From APS to HAPS. 8 Q. The document goes on to say that: 9 "It identifies the requirements that will be used 10 to accomplish direct interface testing between POCL and 11 Pathway, as such this document must be owned and 12 approved by POCL, Pathway and the PDA." 13 Indeed, if we can turn to page 2 of the document, 14 please, and go down to "Approval Authorities", you see 15 there that there are three approvals, Simon Palladino, 16 Pathway; John Robson, POCL; and John Bruce, PDA. Could 17 I ask what the role of the approval authorities was in 18 relation to this document? 19 A. To review and approve the document. 20 Q. Did they have any input into its content from your 21 recollection? 22 A. Not from my recollection. 23 Q. Would it have been possible to conduct this testing, the 24 direct interface testing, without input from Post Office 25 Counters Limited? 8 1 A. I don't think so. 2 Q. Please can we turn to page 10 of the same document and 3 go down to heading 4: 4 "Each party will use its all fault reporting 5 system. Pathway will log any incidents using the fault 6 reporting system PinICL the incident number will be 7 passed back for future progression and clearance." 8 So, in essence, is that any problem that arose 9 during testing will be logged on PinICL on Pathway's 10 side? 11 A. Yes. 12 Q. If we could go back to your witness statement, please -- 13 that's WITN04770100, page 5, paragraph 16 -- you say: 14 "In my role as trainee tester, I was given scripts 15 to run in order to test the equipment and/or counter. 16 I would then record the result of the test and feed the 17 results back to the Fujitsu test manager. It is my 18 understanding that the Fujitsu test manager would 19 communicate the results of the tests with the relevant 20 Post Office test manager." 21 So just to take it in stages, were you involved in 22 passing on any information about testing to the Post 23 Office itself? 24 A. I suspect I was, yes, at some stage. 25 Q. In what forum would that be? How would you pass on the 9 1 information? 2 A. Potentially there may have been triparty calls, there 3 would have been emails and potentially through reporting 4 of the testing that was carried out. 5 Q. Do you recall the type of information that you would 6 have provided to Post Office Counters? 7 A. At the time, and I can't say this because I don't 8 actually recall it, but my assumption is that I would 9 have been passing on details of the PinICL reference 10 number and the faults that were found. 11 Q. Could you just give an overview of the types of areas 12 that you were -- we see here the interface. What else 13 did you test in your role as trainee tester? 14 A. I don't have a good recollection of that at all, I'm 15 afraid. 16 Q. In respect of where you say your understanding was, that 17 the Fujitsu test manager would communicate the results 18 of the tests with the relevant Post Office test manager, 19 what is the basis of that understanding? 20 A. Again, from my recollection when I did the witness 21 statement back in August, is that I wasn't leading any 22 of the discussions. There was always a senior either 23 tester or manager in the meetings initially with myself 24 and any triparty meetings. 25 Q. Are you aware of any formal procedures or protocols that 10 1 were in place regarding the communication of test 2 results? 3 A. I can't say that I am. I think that it generally was 4 agreed -- again, my recollection is vague -- but I think 5 the principle was that, if there was a meeting, then 6 they were documented as part of that meeting. If it was 7 a PinICL and, as I say, or if the Post Office or PDA had 8 any issues, they would be reported via a mail into us. 9 Q. So, overall, your understanding is that things were 10 passed across at these meetings, possibly emails as 11 well, but is it fair to say your recollection is -- 12 A. It is very vague, I'm afraid. 13 Q. My understanding is that you remained in a testing role 14 until 2002 when you left ICL Pathway; is that right? 15 A. That is correct. 16 Q. You then returned to, then, Fujitsu in 2007. 17 A. Correct. 18 Q. At this stage, what's been known as Legacy Horizon was 19 still in use but it was looking for gearing towards 20 changing to Horizon Online and developing Horizon 21 Online. I understand you were involved in the 22 development of Horizon Online? 23 A. In, again, the testing of Horizon Online. 24 Q. Now, the Inquiry will be considering the design, 25 development and testing of Horizon Online in greater 11 1 detail in due course. I want to limit what we discuss 2 to a few small points, starting with testing, if I may. 3 Please could we bring up POL00029327. So this 4 document, and I'll ask you for your held with the title, 5 is "HNG-X: ITU V&I Business Continuity High Level Test 6 Plan". It says you are the author at the bottom. Could 7 you please provide a summary of what this document is 8 describing? 9 A. It is validation and integration and it is business 10 continuity. So it's effectively providing assurance 11 around resilience, business continuity, that the 12 infrastructure will cope with a level of impact. So if, 13 let's say, a server was to go down, that we have 14 sufficient resilience that a single server going down 15 won't impact service and that the service itself will 16 fail over to another component providing the resilience 17 and potentially also the business continuity. So if we 18 were to fall into a disaster recovery scenario, that 19 potentially we could move from one site, one data 20 centre, effectively, to another data centre and maintain 21 service, albeit there would be a period in which we 22 would have to complete that move. 23 Q. This specific area of testing, was this the sole area 24 you were dealing with or did you deal with others as 25 well? 12 1 A. Potentially, I would have dealt with others but I think 2 this was the -- one of the main areas at the time. 3 Q. Could I ask just to move down the document to the 4 "Approval Authorities". Again, here we have three 5 approval authorities. There's the HNG-X test manager 6 and then Andrew Thompson, Post Office Limited test 7 manager, and Tony Wicks, business continuity manager. 8 If it's different to what we went to before, can 9 I just ask you to explain what the role of the approval 10 authority was for this document. 11 A. Again, to review and sanity check the proposal and to 12 provide their approvals from their respective positions. 13 Q. Do you recall what input Mr Thompson from the Post 14 Office had on this document? 15 A. I can't, I'm afraid. 16 Q. Once this document was in its complete form, so 17 approved, would a copy be sent to all the relevant 18 approval authorities as well? 19 A. That is how the process should work, yes. 20 Q. Can I move to a different topic, please, and if I can 21 bring up document FUJ00084350. Actually, let's see, 22 we'll stay there for the moment but we may want to go to 23 the first page, if you need it. 24 This is a spreadsheet that was provided to the 25 Inquiry by Fujitsu and the file title is 20100526_CS 13 1 prayers. It appears to be dated 26 May 2010. Please 2 could you clarify what "CS prayers" are? 3 A. I think it's customer services prayers, and prayers 4 would be a meeting that's held in the morning to discuss 5 issues. 6 Q. Did you attend those prayers meetings? 7 A. I believe I would have attended on occasion, yes. 8 Q. We're looking here at the Closed tab you see at the 9 bottom it says "Closed" and in row 124, column C refers 10 to a problem, saying: 11 "More than 2,000 critical events per day." 12 In column F there are a series of what I presume 13 to be dates listing various entries and at 9/2 in F it 14 says: 15 "Steve Bansal running analysis on all events to 16 see what can be done." 17 Do you have any recollection of these events or 18 what this means? 19 A. Bear with me, I'll just ... 20 Q. Of course. 21 (Pause) 22 A. No, I can't say with any certainty. 23 Q. Are you able to help with what a critical event would be 24 generally? 25 A. A critical event could be a counter going offline, it 14 1 could be many things. 2 Q. You can't assist, yes. No, thank you. We can take that 3 document down now, thank you. 4 Moving on from Horizon Online, your witness 5 statement states that you became a problem manager in 6 around 2010 and that at this point was as a full-time 7 employee? 8 A. That is correct. 9 Q. Again, the Inquiry will be investigating the 10 identification and rectification of bugs, errors and 11 defects in the Horizon IT System in due course but I'd 12 like to explore some general points on the problem 13 management system with you first. 14 Please can I bring up the following document 15 FUJ00080043. This is titled the "RMGA Customer Service 16 Problem Management Process" and it's the second version. 17 Does "RMGA" stand for "Royal Mail Group Account"? 18 A. It does. 19 Q. It states that this is a process definition to describe 20 and document the customer service problem management 21 process. The document was drafted on 22 April 2008, so 22 before your time as problem manager. 23 A. Yes. 24 Q. But would it have described the process of problem 25 management when you became a problem manager in 2010? 15 1 A. The likelihood is yes. 2 Q. Do you know whether this document -- or, to your 3 knowledge, was this document an internal one? 4 Actually, if we can scroll down slightly, please, 5 before I put this question you see the distribution 6 list. To your knowledge, was this document purely 7 an internal document or would the Post Office have 8 received it? 9 A. Based on the information on that page, it would appear 10 to be an internal document. 11 Q. Please could we turn to page 6 of the document. So in 12 this introduction, it sets out the process, objective 13 and scope of problem management and a problem is defined 14 as "the unknown underlying root cause of one or more 15 Incidents". 16 We see in the documentation a distinction drawn 17 between problems and incidents or major incidents, with 18 different processes. Please could you help us with what 19 the difference between a problem and the problem 20 management process and an incident and a major incident 21 process is? 22 A. Okay. A problem could be raised off the back of 23 an incident or an issue in a single branch or multiple 24 branches. We would use the problem itself, a problem 25 ticket, to continue the investigation, the analysis, 16 1 until such time the incident is resolved. 2 For a major incident, the distinction there is the 3 severity and the priority and potentially the impact to 4 the wider estate. So a major incident would mean that 5 potentially a greater number of branches are down, 6 they're offline, there is not a service being offered. 7 So the priority there is resolution to get those 8 branches' services available as soon as possible. We 9 would then subsequently raise a problem ticket for any 10 outstanding issues where we've not developed/understood 11 the root cause to continue the investigation. 12 I think there was an element I haven't covered. 13 Q. Let's just break it down with that first, so we can 14 understand the difference. 15 So, for example, if there was an unexplained 16 discrepancy of a low amount, say, a £5 discrepancy at 17 a single Post Office reported, would that be classed as 18 an incident in itself? 19 A. That would be classed as an incident, yes. 20 Q. The underlying cause of that discrepancy, that would be 21 the problem? 22 A. Yes. 23 Q. A major incident would be, say, if there was a complete 24 outage of service for a period of time, which had a very 25 severe effect on the network but, again, the problem is 17 1 trying to find the underlying cause of that major 2 outage. Is that the distinction? 3 A. Correct. It's getting the root cause. 4 Q. So when we talk about problem management here. We're 5 talking about finding the root causes of bugs, errors 6 and defects, basically, or trying to find whether there 7 is a bug, error or a defect? 8 A. Correct. 9 Q. It refers to reactive and proactive problem management. 10 We're going to, I think, look at that in due course as 11 we go through this document here. 12 Can I start, though, by looking at some of the 13 responsibilities for problem management and, if we turn 14 to page 6 of this document, if we're on page 6, if we 15 could go to the bottom of it, please. Thank you. 16 So the first point here is a "Process Owner" and 17 it says: 18 "The owner of the process this POA Service 19 Delivery Manager responsible for the Service most 20 affected by the Problem. The Process Owner, otherwise 21 known as the Problem Manager, is appointed by the 22 Service Delivery Team Manager." 23 So if a problem arose, who would have day-to-day 24 responsibility for the problem management process and 25 seeing that the problem is investigated? 18 1 A. So unless there is a defined problem manager, it would 2 fall to the SDM, whose service that problem falls under. 3 Q. So the "SDM" being the service delivery manager? 4 A. Service delivery manager, correct. 5 Q. Is it the case that a service delivery manager can 6 appoint a problem manager and delegate responsibility 7 for that particular problem? 8 A. That can happen. 9 Q. In 2010, when you were described as a problem manager -- 10 A. Yes. 11 Q. -- were you a person to whom problems would be delegated 12 or were you a service delivery manager? 13 A. I was a problem to whom -- a person where the problems 14 would be appointed to. 15 Q. I understand that you became a service delivery manager 16 later in your career; is that right? 17 A. That is correct. 18 Q. When did that happen? 19 A. 2010. 20 Q. Right. Sorry, so you were -- you weren't a problem 21 manager in 2010, you were a service delivery manager in 22 2010? 23 A. My apologies. I started out in the service team as 24 a problem manager and then moved into becoming an SDM. 25 Q. In the same year? 19 1 A. Later that year, 18 months afterwards. 2 Q. Roughly, yes. 3 A. But it was a progression. 4 Q. If we turn over the page, there is a role described as 5 a "Problem Resolver", who's responsible for finding 6 a resolution to the problem. Would that be, for 7 example, someone in the SSC who's actually 8 investigating, running diagnostics? 9 A. Possibly someone in the SSC but it would be someone who 10 has the technical knowledge. So SSC, being the third 11 line support team, would have knowledge, articles and 12 information for them to investigate but it may be that 13 the resolution would come from the fourth line support. 14 So there isn't a specific problem resolver and it is 15 allocated case by case. 16 Q. So your role as problem manager would be to, what, 17 oversee them and -- well -- 18 A. To ensure the process is followed and that we have the 19 correct support, et cetera, and that we're doing the 20 communication both internally and externally. 21 Q. Looking then at how this process works, could we start 22 with problem identification and turn to page 10 of this 23 document, with the flowchart at section 4.1.1, please. 24 So we see on the top left there's two ways into 25 the problem management process: incident management and 20 1 alerting of a pattern likely to cause a problem, at the 2 far left. Is that what you would describe as proactive 3 problem management where an incident is detected by 4 Fujitsu itself? 5 A. Yes. 6 Q. Then we also have the major incident management in the 7 second from the left. 8 Then it says to open a problem record at 1.1.1 in 9 the middle. The third box on the top, hard to see but 10 it says "Incident & Problem Alerting Process", was there 11 a written procedure for the incident and problem 12 alerting process that you're aware of? 13 A. The incident and problem alerting process, to my 14 recollection, would be the daily monitoring that is 15 performed by the SMC. So they would effectively see 16 alerts, because they're monitoring the system, and they 17 would then the raise an incident. The incident would 18 then be trended and that would be how we would then 19 raise a problem record. 20 Q. So that may be a way through opening a problem record 21 but, looking at this flowchart, if we look at the box 22 1.1.1, we're at the stage where a problem record has 23 been opened and then the flowchart goes off to three 24 boxes. Now, the middle one is "Start Total Time Clock" 25 and the second one is "Start [I assume Service Level 21 1 Agreement] SLA Clock". 2 Is that referring to, sort of, deadlines for when 3 a problem should be resolved by? 4 A. The SLA clock is if there is a Service Level Agreement 5 in place. So, at that point, effectively, we're 6 starting the clock. 7 Q. Yes. So we're in the position where we've got the 8 problem open? 9 A. Yes. 10 Q. So it may have come from the SMC or not, but the box we 11 didn't look at on the left "Incident & Problem Alerting 12 Process", do you know to what that refers? 13 A. I don't, I'm afraid. 14 Q. When a problem record was opened, who would be told of 15 the problem or provided with the problem record? 16 A. The problem manager would obviously be either made aware 17 or would have raised the ticket themselves. That would 18 then be put onto effectively a spreadsheet, a database, 19 and then that would be informed to the wider account via 20 an update on the actual incident ticket. So the 21 incident ticket would then have a reference back to the 22 problem record. The problem record should then have 23 a reference back to the incident itself. 24 Q. So you said the wider account. That's the wider group 25 of people within Fujitsu working on this account? 22 1 A. Correct. 2 Q. If we follow this flowchart through at the top right we 3 see it says go to "A", after we've taken these various 4 steps. If we could go to page 12, please, of the 5 document -- thank you -- this section concerns 6 classification and, in paragraph 4.1.2.1, which is just 7 below the flowchart, it asks the problem manager and 8 resolver to capture the sense and respond codes. Could 9 you assist with what those are? 10 A. I can't -- no. What I would say is that I'm not sure 11 whether the -- how long the sense and response codes 12 were actually in play and what I would say is that 13 I think we have a matrix which would give us the 14 priority and severity, which I think is further down in 15 the document. 16 Q. Yes, I want to turn to that now, actually. We see 17 "Priority" is in a different section, so if you follow 18 it across, it's to 1.2.3. So after the problem's been 19 classified, a priority's set and, in that regard, we 20 look at the appendix to this document. 21 I'd like to look at the final page first, which is 22 page 23, please. This is a table which says "Priority". 23 Is this the table to which you were referring to get 24 a priority score for the problem? 25 A. Correct. 23 1 Q. On the column on the left, there is an impact score or 2 an impact value of 1 to 5 and then the columns on the 3 top, from the second column to the final column, these 4 are urgency scores, again of 1 to 5, which we'll come to 5 in a moment. But for present purposes, it looks like 6 this ends up with a score -- if you combine these two, 7 the impact and urgency to get a priority score of 8 somewhere between 1 and 5? 9 A. That is correct. 10 Q. Were there any deadlines or -- how were the different 11 scores for priority treated? How was a "1" priority 12 different from a "2"? 13 A. So a 1 priority is the most immediate; so effectively 14 resolve this with the highest priority. A 5 would be 15 the lowest priority. 16 Q. Were there targets or deadlines for a priority 1 and 17 then a priority 2? 18 A. Relating to -- 19 Q. How long they needed to be -- within what time they 20 needed to be resolved? 21 A. Within problem management, I don't believe that there 22 was. 23 Q. In practice, what effect did the priority level have on 24 the speed to which problems were resolved? 25 A. If you had a P1 then, effectively, we would trump any 24 1 other activity that's going on to be able to call SMEs, 2 the support units to come and prioritise this work to 3 look at the resolution of the incident or the issue and, 4 bearing in mind it was a problem that would already have 5 had a high priority incident allocated to it, that 6 activity would have been ongoing. 7 So yes, a higher priority would have meant that 8 people would have paid attention and actually 9 appropriately prioritised the activity. 10 Q. Is it possible to say, if you gave something 11 a priority 1, within what period of time you would have 12 expected the problem to be resolved? 13 A. As I say, with urgency but a problem is different to 14 an incident. An incident does have a four-hour SLA or 15 an eight-hour SLA or a three-day SLA. A problem does 16 not have the same SLAs because those incidents, high 17 priority, are being worked on as part of this problem 18 ticket. 19 Q. If you look at some of the factors that go into giving 20 us the priority, please can we turn to page 21. Thank 21 you. This first refers to, we see in the bottom the 22 impact value. There's a table there but, starting at 23 the top, the first step is to give it a criticality 24 value, which there are, again, five scores from critical 25 to cosmetic. 25 1 This would be assigned by the problem manager; is 2 that right? 3 A. That is correct. Problem manager and SMEs. 4 Q. Was there any guidance on what would be determined as 5 a critical, high, medium or minor? 6 A. The critical would be defined as something which has 7 effectively a show stopper on a wider scale. So, again, 8 if we go back to a P1 scenario, almost a disaster, 9 service has stopped. That's regarded as critical. 10 And then we go down in severity down to the things 11 which are cosmetic or minor. 12 Q. So when you say the show stopper point there, critical, 13 you suggested that's something that's stopping service 14 but also would you take into account how many people 15 were -- 16 A. Absolutely, and I think that's covered in the impact 17 section below. So if it's one user or if it's the 18 entire estate, they will have a different -- 19 Q. So that's taken into account under the impact but the 20 criticality part, is it fair to say it's a judgement 21 call at there's no particular written guidance on what 22 is critical and what is medium? 23 A. So, again, hence why it's the problem manager and the 24 resolver group looking at this. So it's a collective 25 view on how critical things are and it's not one 26 1 individual's judgement. 2 Q. Was there ever an incentive to lower the criticality 3 score or to put in a lower score than you otherwise 4 would have thought? 5 A. No. No, there was never any pressure to do anything 6 like that. 7 Q. When we go down -- if we could move down -- thank you -- 8 to the impact table, the number of users affected, 9 obviously at the top we see it ranges from, on the left, 10 over 70 per cent to, on the right, to a single user and 11 that affects the overall impact score. 12 Was this -- "Number of users affected", was this 13 the number of users that had been affected or would it 14 be an assessment of how many users may be affected by 15 a problem? 16 A. I think for a problem we would be -- we would take both 17 into account. If the problem was well understood and 18 defined, then potentially you'd be looking at just the 19 affected users because, again, we'd be in a position to 20 understand that. 21 If the issue/problem was relatively new and that 22 was still being defined and understood, then we would 23 also look at the potential wider impact and take that 24 into account. 25 Q. So from a criticality point of view, if, say, less than 27 1 100 -- say less than 50 subpostmasters were reporting 2 unexplained discrepancies in their branch accounts, 3 where would that have fallen on the criticality score? 4 A. To my mind, that would have been a critical. 5 Q. That would have been -- 6 A. That would be a critical. If you're getting that many 7 postmasters reporting something of that nature, that's 8 something that needs to be looked at with urgency. 9 Q. So that's if they were reporting all at once. If it's 10 just a single discrepancy that's being reported, how 11 would that change things? 12 A. That would change things because, until we've done some 13 trending along that, we don't know where those 14 discrepancies are. They could be related, they may not 15 be related; so we would, as part of that problem review, 16 pull together any child incidents to see if they 17 actually are related. 18 Q. Could we turn the page, please, to the urgency score on 19 page 22. Before I ask you about the detail of it, in 20 broad terms can you explain how the impact score 21 differed from the urgency score? 22 A. Sorry, my mind's gone blank. Can you repeat your 23 question? 24 Q. Of course. In what way -- what considerations or what 25 different considerations would you take into account 28 1 when arriving at an urgency score, in comparison to the 2 impact score? 3 A. I guess we would look at what potentially may unfold 4 over the next period. Depending on where the scenario 5 of the issue is, it could be that with the batch 6 processing that happens overnight that may then add to 7 the severity or the impact of the issue, and it could be 8 that during a working day there is the opportunity to 9 support the postmasters, support the post office with 10 a resolution so that would make that resolution within 11 that time span far more urgent than if there was a roll 12 on impact of an overnight batch. 13 Q. Let's look at what the urgency table says and go through 14 it there. For the first level, which is the most urgent 15 it says it: 16 "Has a significant adverse impact on the delivery 17 of service to a large number of end users. 18 "Causes significant financial loss and/or 19 disruption. 20 "Results in any material loss or corruption of 21 customer data." 22 It says: 23 "For example, incidents with this urgency may 24 affect the COMPANY." 25 What company is being referred to there when it 29 1 says "the company"? 2 A. I'm afraid I don't know. 3 Q. Would it be the Post Office as a whole rather than 4 individual subpostmasters? 5 A. I don't know, I'm afraid. 6 Q. The urgency value 1, as we say, refers to significant 7 financial loss or disruption. The second score, it says 8 it causes -- sorry, urgency score 2 -- it says: 9 "Causes a financial loss and/or disruption to the 10 customer which is more than trivial but less severe than 11 the significant financial loss described in the 12 definition of an Urgency level of 1." 13 Are you aware of any guidance on how a problem 14 manager was to distinguish between trivial or 15 significant financial loss or somewhere in between? 16 A. No specific guidance. 17 Q. At the bottom of urgency score 2, it says: 18 "For example, incidents with this urgency may 19 affect a VIP SITE." 20 Do you know to what that refers? 21 A. I think, historically, Post Office did have a number of 22 sites that they determined as VIP and -- yes, I'll say 23 no more. 24 Q. Could we look at an actual problem report. It's 25 POL00029568. We see this is a problem report. It says 30 1 it affects 14 branches. It was reported by Steve Parker 2 and you're listed as the problem manager. 3 Now, this concerns a bug in the system described 4 by Mr Justice Fraser as bug number 3, the suspense 5 account bug and, in essence, what this document shows or 6 suggests is, in some branches, there was data entered 7 into the local suspense account that was relevant to 8 balancing in trading periods 9 and 10 in 2010 and '11 9 and this data in the suspense account was retained in 10 the database. Therefore, when the branches came to 11 balance in the corresponding trading periods 9 and 10 in 12 later years, that 2010 data was reused incorrectly. 13 Is that a fair summary of the problem? 14 A. I believe so, yes. 15 Q. Whilst branches had experienced the error in 2012, it 16 was only reported to Fujitsu in 2013; is that right? 17 A. As I understand it, yes. 18 Q. If you could move down, please, the page, we see at this 19 stage 14 branches are listed as having discrepancies. 20 Some of them are small amounts. For example, the third 21 one down is 1 penny but, as you see the fourth one down 22 is £9,799.88. Can we go to the top of the table, 23 please, again. Thank you. 24 Now, this was given an urgency score by you of 2. 25 Could you explain why this had an urgency score of 2? 31 1 A. I think at the time because there were 14 branches and 2 because at the time we were looking to get the 3 investigation underway. So I think, if memory serves, 4 this had come through to us from Post Office. So we 5 raised an immediate problem record to do effectively 6 a historical investigation into those 14 branches. 7 That's why I think it was a 2 rather than a 1. 8 Q. When you were -- in a case like this when you're given 9 an urgency score, would you consult the appendix to 10 which we just referred or was it more of a sense of 11 experience and feel to what score would you ascribe an 12 urgency score? 13 A. So I think I can't hand on heart say that I looked that 14 appendix for this one. I think I may have done; I may 15 not have done. So I can't comment. But, normally, 16 I think the advice to the problem management team is to 17 look at the appendix. 18 Q. Please could we go back to the appendix -- it's 19 FUJ00080043 -- and turn to page 23. On what you just 20 said, the paragraph below the table does say: 21 "For example, if the agent decides that the 22 Urgency score is 3, and the Impact has been calculated 23 as 2, then from the Priority table, the final Priority 24 will be automatically generated as 2. The assigned 25 priority can be overridden if the problem is serious and 32 1 discussed with the Service Delivery Team Leader, but the 2 Problem Management process must be followed." 3 Now, in the problem record there that we just 4 looked at, the priority score given was 4. If we look 5 on the urgency score for an urgency score of 2, the only 6 priority scores you can give are 1, 2, 3, 3 and 5. 7 Could you assist with why you considered that or gave 8 the priority score of 4 for that problem? 9 A. I'm afraid, I can't. 10 Q. In practice, giving it a score of 4 rather than, say, 3, 11 what difference do you think that would have made in 12 practice to how the problem was resolved? 13 A. On this occasion, I don't think an awful lot. Having 14 read through the rest of the pack, I know that that 15 particular issue was dealt with by a number of people 16 and I think there were a number of high priority PEAKs 17 that were raised and the investigation was quite 18 intensive. 19 Q. You have raised it. Let's look at that. It's 20 POL00029671. Can we turn to page 6, please. 21 There's an entry, 6 March 2013. I should say, for 22 the record, that this PEAK is the PEAK referred to in 23 the problem report we've just seen but the entry on 24 6 March 2013 at 4.05 says: 25 "There was a conference call with POL (Laura 33 1 Darby, Mark Wardle and others) on 28th Feb about this 2 call, and the spreadsheet showing the impact of the 3 problem on the 14 branches was sent to them by Steve 4 Bansal. We are waiting to hear from Mark whether this 5 is sufficient information for them to resolve the 6 consequences on the branches and POLSAP." 7 So do you recall how this problem was resolved 8 thereafter following this call? 9 A. I don't. 10 Q. You, mentioned that you remember this in particular at 11 there were several people on it. Was this problem given 12 more resources than, say, another priority 4 problem 13 would be given? 14 A. I think that, in this particular scenario, I think Anne 15 Chambers, it was her priority. She effectively dropped 16 all other work, to my approximate knowledge, as it were, 17 and this was her main focus. I believe there was 18 another PEAK open and I think that in the background 19 other teams were also looking at different aspects in 20 support of this. So Anne wasn't looking at this on her 21 own; there were wider teams looking at the scenario and 22 the issues. 23 MR STEVENS: Sir, I don't know if you want to have a break 24 this morning but this would be a good point to break for 25 the hour mark? 34 1 SIR WYN WILLIAMS: Yes. Well, certainly that's okay. All 2 I don't want to do is to have a break and then have 3 another long break if you see what I mean. How are we 4 going with the witness, generally? 5 MR STEVENS: Quicker than -- yes, there's probably about the 6 same again, maybe less. 7 SIR WYN WILLIAMS: Okay. So should we -- let's ask 8 Mr Bansal. If we have, say, a ten-minute break now, 9 should we then complete his evidence without having 10 a formal lunch break, so that that would take us to 11 maybe 1.30 or would he prefer to have a formal lunch 12 break at 1.00? 13 A. I'm easy to go through. 14 SIR WYN WILLIAMS: You would prefer to go through? 15 A. I would prefer to go through rather than stop for lunch. 16 SIR WYN WILLIAMS: Is that all right with you, Mr Stevens? 17 MR STEVENS: It is, sir. 18 SIR WYN WILLIAMS: So we will have a ten-minute break now 19 and complete Mr Bansal's evidence and that will be it 20 for the day. 21 MR STEVENS: Thank you, sir, fine. 22 (12.26 pm) 23 (A short break) 24 (12.38 pm) 25 MR STEVENS: Sir, can you see and hear me? 35 1 SIR WYN WILLIAMS: Yes, I can. Thank you. 2 MR STEVENS: Thank you. Mr Bansal, we will continue. Can 3 you please bring up on the screen POL00029671. 4 That's my error in the reference. It's 5 FUJ00080043. 6 Thank you. If you could turn to page 13. Sorry, 7 page 14. Thank you. 8 So once we've got the priority, it goes into this 9 flowchart of managing root cause process and 10 establishing corrective action and it's, at this point, 11 I assume, where the problem resolver takes over the 12 mainstay of the technical work? 13 A. Yes. 14 Q. Now, as a problem manager at this stage, how do you 15 review or, in 2010, how would you review or keep track 16 of how problems were being resolved or diagnosed? 17 A. So we would have regular meetings with the resolver and 18 possibly the support teams to understand exactly where 19 we are with getting to resolution. 20 Q. Would the regularity of those meetings be connected to 21 the priority of the problem or ... 22 A. Yes. If it was a high priority incident, then we would 23 be having almost daily conversations to track progress 24 but, again, we would also be reliant on the SME, the 25 support teams, providing sufficient feedback to 36 1 determine the regularity of those conversations. 2 Again -- yes. 3 Q. Could we turn to page 16, please. Now, this is, again, 4 part of this error control process and the step in the 5 flowchart here at 2.3.1 says "Assess if permanent 6 solution is required", and it gives eight options for 7 this assessment, ranging from "Impact minimal: not 8 cost-justifiable" with other ones requiring -- it says 9 "Resolution requires POL funding" or "Resolution 10 requires action by POL". 11 I want to look at the first two. Would anyone 12 from the Post Office be involved in this assessment of 13 whether a permanent solution was required? 14 A. Yes, they should be. So we would hold a regular review 15 of problem records with the Post Office and we would 16 take to them our findings and, if we were in a scenario 17 where we had to look at the justification in this 18 manner, if it wasn't apparent, ie that we had to fix it, 19 then we would have a conversation with Post Office. 20 Q. These meetings in 2010 that started around that point, 21 how often would you have those meetings with Post 22 Office? 23 A. I can't say with absolute certainty. I would suggest 24 minimally at a month but I can't say with any greater 25 recall. 37 1 Q. At these meetings, would you discuss all active problems 2 or a certain priority of problems? 3 A. I think the approach would be that we would discuss all 4 active problems but with priority given to those that 5 are of the highest urgency priority, hot topics, 6 et cetera, and then you would work your way down the 7 list. 8 Q. So where it says in this chart "2.3.2 Impact minimal: 9 not cost-justifiable", if the problem resolver had found 10 a bug in the Horizon System that Fujitsu had provided to 11 the Post Office, to whom was the cost not justifiable to 12 enter a permanent solution: Fujitsu or the Post Office? 13 A. That would be determined by what the root cause of the 14 problem was. 15 Q. You've identified the root cause of the problem and now 16 the question is what action to take with it and one of 17 the options is not to do anything or not to implement 18 a permanent fix because it's not cost justifiable. 19 A. Yes. 20 Q. Whose costs are we looking at here? 21 A. Again, that is dependent on the resolution. If what is 22 found to be which case is that it's missing requirements 23 or incorrectly stated requirements originally, then that 24 may be something that we would look to Post Office. 25 Because it's a change of requirements, they would need 38 1 to confirm that what they would like is that issue 2 addressed with a new set of requirements. 3 Q. So that would be covered, would it, by, if we look in 4 this diagram, "2.3.8 Resolution requires POL funding" or 5 "2.3.9 Resolution requires action by POL"? 6 A. Correct. 7 Q. So in 2.3.2, "not cost-justifiable", does that refer to 8 cost to Fujitsu? 9 A. No, that would also refer to cost to Post Office. So 10 I see what you're saying but it falls under that 11 category as well. So, yes, there may be an occasion 12 where Fujitsu, depending on what the impact is, may say 13 it's not justified, as Post Office might have done 14 historically as well. 15 Q. So in what circumstances would Fujitsu say the cost to 16 them meant it was unjustifiable to implement a permanent 17 fix? 18 A. So it could be that the impact is only to our support 19 teams. So if it means that we see something within our 20 monitoring, let's say, our error handling, and it -- 21 effectively we could then potentially ignore that 22 particular scenario. So what we then do is write 23 a knowledge base article to that effect so we don't then 24 have to put in a software or hardware update to achieve 25 that. So it's cheaper, therefore, to put together the 39 1 knowledge article in that scenario, where we see that 2 again we know that, under those circumstances, we can 3 ignore that event. 4 Q. You said that the Post Office were involved in this 5 assessment. What happened if there was a divergence of 6 views on whether or not to implement a fix? 7 A. Then we would follow the customer's recommendation. 8 Q. Please can we bring up another document. It's 9 FUJ00085191. This is another spread sheet. We're on 10 the first page there. It was provided to the Inquiry by 11 Fujitsu, the title was "POLS [so POL's] Weekly Problem 12 Review 241013". 13 Do you recognise this type of document? 14 A. I do recognise this type of document. 15 Q. Who created this or -- not specifically which person, 16 but which corporate entity would create this document? 17 A. I'm going to say I'm little bit hazy whether it would 18 have been something that Fujitsu produced or whether 19 it's something Post Office produced but it was something 20 that we reviewed collectively and updated collectively. 21 Q. Can you recall when spreadsheets like this were used for 22 collective discussions first? 23 A. I can't say that I can recall when it started or whether 24 it was practised when I joined. 25 Q. In terms of the meetings you discussed earlier about 40 1 going through various problems that you had, you said at 2 the least regular interval's monthly but you couldn't 3 remember how often precisely -- 4 A. Yes. 5 Q. -- would this be the document that was used -- 6 A. This would be the type of thing, yes. 7 Q. The type of thing? 8 A. The type of thing. So, again, just to your point, 9 I think when I initially started as a problem manager it 10 was monthly but, to my recollection at that time, 11 I found that to be insufficient, so I brought that 12 forward to fortnightly and then that also was quite 13 slow, so we went to a weekly meeting. 14 Q. At these meetings, who from POL -- or at least who in 15 terms of what job roles from POL -- would attend? 16 I should say Post Office. Which job roles from Post 17 Office would attend? 18 A. So I think we would have some representation from Post 19 Office from a service perspective. We would have I'm 20 going to suggest some SDMs and senior service person. 21 From memory, I'm not going to say names because I can't 22 remember all of them but, certainly, there was at least, 23 you know, two to three people at all of these meetings 24 early on. 25 Q. Could we now just turn to the "Closed" tab, please, on 41 1 the spreadsheet and if we could go to row 20. Thank 2 you. 3 I think what we'll need to do, if we could drag 4 the row 20 down so that it's -- I think if you go to the 5 left there and -- yes, thank you -- drag it down we'll 6 get that detail in. Thank you very much. 7 So this we see from column D refers to the 8 14 branches and the local suspense account issue and we 9 have in column F, which is titled "Supplier Updates" 10 a series of entries with dates. If we could go across 11 to column G which is titled "POL Updates "there are also 12 entries on dates as well, not necessarily the same. 13 How were these columns updated; can you recall? 14 A. Yes. They were potentially drawing the meetings. As we 15 mentioned earlier, we would go down the sort of priority 16 list and we would look for updates from either side on 17 the particular issues. 18 Q. Who maintained -- because if this is an updating 19 document that was used for meetings, was there a master 20 copy or was someone responsible for maintaining a master 21 copy? 22 A. I'm going to say that I think the master copy was held 23 with Fujitsu and was shared post every meeting with the 24 Post Office representatives. 25 Q. Thank you. We can take that document down now. 42 1 Could we please bring up FUJ00085175. We were 2 previously looking at version 2 of the "Customer Service 3 Problem Management Procedure" before the break. This 4 is, if we go to the bottom, version 2.3, which 5 I understand means it's in draft form; is that right? 6 It's a draft, not approved version? 7 A. I can't see approved on here. 8 Q. If we go up, sorry, slightly -- 9 A. Ah, yes. 10 Q. Draft version. 11 A. Yes. 12 Q. At page 4 -- go to the bottom, please -- you're listed 13 as a mandatory reviewer. So presumably you would have 14 seen this document at the time. 15 A. I would have done. 16 Q. Please turn to page 9 of the same document. Under 17 heading -- can we go to heading 1.5.1. This says: 18 "The Problem Records for [Post Office Account] is 19 held on the ..." 20 Is it TRIOLE? 21 A. TRIOLE. 22 Q. Thank you. 23 "... Service Desk [system]." 24 Who would have access to the TRIOLE service desk 25 system? 43 1 A. Fujitsu staff. 2 Q. Post Office didn't have access to that? 3 A. Correct -- I don't believe so, no. 4 Q. Over the page, if we may, it refers to, at the top: 5 "Problem Managers can access the Problem Action 6 Plans by ..." 7 Then it gives a reason -- sorry, the way to do it, 8 and it says: 9 "These reports are held within a spreadsheet which 10 contains three tabs: Horizon, POLSAP and Closed." 11 Is that referring to the spreadsheet or the types 12 of spreadsheet we were seeing that we just took you to? 13 A. I believe so. 14 Q. Please could we move to a different document. It's 15 FUJ00085985. We see from it there's a note on the 16 second paragraph of the substance: 17 "Note Jan 2018: Document updated to reflect the 18 changes on the POA Account." 19 So we've jumped forward quite a bit. 20 Have you seen this document before? 21 A. I think I have seen this document before, yes. 22 Q. Do you know -- it says "IP Handover". Do you know what 23 it was drawn up for? 24 A. Sorry, could you? 25 Q. Sorry, do you know for what purpose this document was 44 1 drawn? 2 A. Yes, it was effectively a task list for IPs. 3 Q. "IPs" being? 4 A. Industrial placement. 5 Q. These people would help with the problem management and 6 incident management processes? 7 A. They would help with various tasks across the service 8 team to give them some scope and bandwidth of some 9 training and some understanding of how business works. 10 Q. Could you turn to page 4, please. This refers to the 11 "ATOS Problem Spreadsheet". I think at this stage it 12 would be helpful to introduce ATOS. Could you state how 13 ATOS fit into the problem management process? 14 A. So I think at the time Post Office effectively brought 15 in a managing agent to work on their behalf and, as part 16 of that, they procured a problem management service 17 through ATOS. 18 Q. So ATOS were when it says "ATOS Problem Spreadsheet" and 19 we talk about ATOS, that is subcontract -- or people 20 contracted by the Post Office? 21 A. Yes. 22 Q. It says that: 23 "... the Problem Spreadsheet to ATOS Problem 24 Management ... with Fujitsu updates which are discussed 25 on the weekly Problem Management call every Friday." 45 1 I think earlier in your evidence you referred to 2 these calls going from a monthly to fortnightly and then 3 an even shorter period of time. 4 Do you have any recollection as to when they went 5 to weekly calls? 6 A. I don't, I'm afraid. 7 Q. The spreadsheet that's referred to, is it basically -- 8 was it in a similar form to the spreadsheet we looked at 9 earlier? 10 A. I believe so. 11 Q. So, fundamentally, the process you're describing hadn't 12 changed, just the frequency of the -- 13 A. The -- yes. 14 Q. -- meetings? 15 A. Actually, if you -- I think if you go back to the other 16 spreadsheet that might give an indication of the 17 regularity of that particular spreadsheet, whether that 18 was weekly. 19 Q. In due course, the Inquiry can look at the documents to 20 see that but thank you. 21 Can we please move to page 12 of this document. 22 Now this talks about PEAK reporting. The Inquiry's 23 heard a lot about PinICLs because of the time-frame but 24 PEAKs were effectively the same as PinICLs in order that 25 they -- well, the PEAK system was a system in which 46 1 problems were recorded and it was a flow of the 2 information done to rectify those problems. Is that 3 fair? It was a log, basically, of actions taken. 4 A. It was or it could be used in that form, yes. 5 Q. Was there any material difference between the PEAK and 6 the PinICL systems? 7 A. At a high level, I'm going to say I don't believe so. 8 Q. Page 12 says that these instructions are: 9 "... to generate a PEAK Report for Steve Bansal in 10 preparation for the Leadership Team Meeting on Friday." 11 Who attended the leadership team meeting? 12 A. That would be an internal Fujitsu meeting, from 13 recollection. 14 Q. What was its purpose? 15 A. To provide an update to the leadership team on the 16 status of service. 17 Q. With the reporting of PEAKs, the second paragraph says", 18 As PEAK Reporting is used to keep track of the trend of 19 PEAKs", and goes on to say "when there is a sudden 20 increase or decrease", can you explain what Fujitsu did 21 in respect of trend analysis. How did it analyse trends 22 in PEAKs? 23 A. We would have members of the service team and the MAC 24 team looking at trends, effectively, if we were seeing 25 an increase in them and, if we were seeing an increase 47 1 or a decrease, in which areas, and was that associated 2 to any new releases, was that associated to any updates 3 that had gone out, positive or negative we needed to 4 understand what was going on and then potentially, 5 proactively be able to get ahead of any issues as well. 6 Q. So who within Fujitsu was responsible for that trend 7 analysis? 8 A. So responsibility for it would ultimately come to myself 9 in that particular phase, but it was a number of teams 10 that were producing that activity. 11 Q. Which teams would they be? 12 A. So I think at the time they were the -- I think they are 13 currently called the MAC team and I can't for the life 14 of me remember what they were called then. 15 Q. At the start of your evidence, or near the start, we 16 discussed proactive problem management. Presumably this 17 is an example of proactive problem management analysing 18 the trends of PEAKs? 19 A. Yes. So the PEAKs will be done via that team -- 20 apologies to talk across you -- and the problem manager 21 would be doing the problem effectively trending. But 22 the two should meet. 23 Q. Was there anything else other than this PEAK -- those 24 two points you said, that Fujitsu did in respect of 25 proactive problem management? 48 1 A. I could say probably, yes, but nothing is coming to 2 mind. Apologies. 3 Q. Before moving on, can we please move to page 16. This 4 refers to major incident reports and when we discussed 5 this earlier you referred to a major incident being 6 a particular incident that had a significant effect on 7 the network or it was particularly severe. 8 A. Yes. 9 Q. In this, it says, starting with the second line: 10 "In the event of a Major Incident, you alongside 11 the rest of the team will be expected to drop whatever 12 you are doing to manage the issue in the most effective 13 way." 14 I don't need to read the second paragraph. The 15 next one is: 16 "When producing the Major Incident Report, you 17 will be assisted by the Duty Manager who was running ... 18 the incident, who will provide you with a detailed 19 timeline of events, including calls that were made and 20 resolution steps taken by the individual teams. With 21 this information, you will do the typing of the first 22 draft using the account template ... 23 "Once you have completed the report, you will 24 review with the relevant parties, eg Duty Manager 25 involved and Steve Bansal, before sending the report to 49 1 Steve Bansal. From this point Steve will make the final 2 edits and send to the customer. Your main job is to 3 type up the report and make sure all detail is recorded, 4 Steve will make the decision to remove any unnecessary 5 detail." 6 So in respect of major incidents, you were the 7 sort of final point of call or the final interface of 8 information between Fujitsu and the Post Office? 9 A. Correct. 10 Q. What type -- when it says you would make the decision to 11 remove any unnecessary detail, what types of thing would 12 you remove? Would they be substantive or -- 13 A. They would be what I call the "he said/she said". So 14 effectively some of the chit-chat. So, again, the 15 purpose of having an IP recording what was going on, 16 effectively as a transcript almost, they would document 17 everything that was kind of said and when that then came 18 to me to review, I would remove some of that because it 19 wouldn't be pertinent to the actual final report. 20 Q. Were you ever under any pressure to downplay 21 an incident? 22 A. No, no. 23 Q. Please -- 24 A. Apologies. I was going to say that, while the report is 25 being produced and while the major incident is ongoing, 50 1 I would have open dialogue with my Post Office 2 counterpart and I would be providing them with updates. 3 That's before and after a major incident. So the 4 technical written would support everything I'd been 5 saying to him. 6 Q. Can we please go back to FUJ00085175 and can we please 7 turn to page 9. This was a document we were looking at 8 a moment ago, version 2.3 of the problem management 9 process. 1.4 refers to metrics to be reported monthly, 10 which will be used to measure effectiveness of the 11 process and drive performance of the process and overall 12 service in general. That included things such as number 13 and end impact of incidents occurring before root 14 problem is identified and resolved. 15 Do you know who was responsible for including this 16 in this document? 17 A. I think, at the time, it was my predecessor or my 18 manager at the time. 19 Q. Please can we now turn, on this issue, to the Horizon 20 Issues judgment, which can be found at POL00022840 and 21 page 97. In this section, Mr Justice Fraser is making 22 findings on Mr Godeseth's evidence. Presumably you know 23 Mr Godeseth as a member of Fujitsu as well? 24 A. Correct. 25 Q. At paragraph 322, Mr Justice Fraser refers to a later 51 1 version of the problem management document we've been 2 discussing. So as you will see in the second line, it 3 refers to being copyrighted in 2017. 4 At paragraph 324, he refers to paragraph 1.4 of 5 that document. He says: 6 "The following metrics, to be reported monthly, 7 will be used to measure effectiveness of the process and 8 drive performance of the process and overall service in 9 general ..." 10 Over the page, we see the list which we saw in the 11 document previously. 12 If we could go down please to 325 -- thank you -- 13 it says that: 14 "... the Claimants ... sought to obtain reports 15 that would [have been] expected to exist [as a result of 16 this policy]." 17 It says that: 18 "... Fujitsu stated (through the Post Office's 19 solicitors) that 'Fujitsu believes that it does not 20 record problems in such a way that would allow this to 21 be determined without retrospectively carrying out 22 detailed analyses' and that it would require 23 'a disproportionate effort and cost' to provide these." 24 Mr Justice Fraser then quotes from Mr Godeseth's 25 evidence and he says: 52 1 "I have spoken to my colleague Steve Bansal, 2 Fujitsu's senior service delivery manager, who has 3 informed me that the Post Office account customer 4 service problem management procedure document was 5 introduced by Saheed Salawu, Fujitsu's former Horizon 6 lead service delivery manager and that Saheed Salawu 7 left the Fujitsu Post Office account in around 8 February 2013, before the new procedure had been 9 implemented. I understand from Steve that Saheed 10 Salawu's replacement did not wish to implement the 11 changes and therefore the records referred to by 12 Mr Coyne in paragraphs 5.157 to 5.159 of his report do 13 not exist, as we continued to follow the previous 14 existing reporting methodology." 15 Do you recall having a conversation like with 16 Mr Godeseth? 17 A. I do. 18 Q. Is Mr Godeseth's evidence correct in that regard? 19 A. It is. 20 Q. So the 1.4 documentation and procedures were never 21 implemented? 22 A. They were not implemented. 23 Q. Why was that? 24 A. I think when I then subsequently took over, in my view 25 most of the data was being captured in alternate 53 1 locations, not necessarily as a specific problem KPI 2 dashboard, shall we say, and the majority of that 3 information was being discussed with Post Office. So if 4 I took those points and reviewed them in the context of 5 that meeting that we were having weekly, those points 6 were being picked up. 7 What I hadn't -- what I didn't do was put them 8 into a dashboard. 9 Q. Fujitsu through the Post Office solicitors is recorded 10 to have said that to retrospectively carry out detailed 11 analyses, it would require a disproportionate effort and 12 cost to provide these. If we could go up the page, 13 would it have been difficult to ascertain these issues 14 or ascertain this data retrospectively? 15 A. I think some we of the data would have been available 16 but I don't think it would have been easy to have then 17 subsequently collated all of it. 18 As I say, some of those points, I think, are 19 discussed but is it there for anyone to root out? 20 I don't think so. Depending on how far back anyone 21 would like anyone to go to retrieve that historical 22 data, it would take some effort. 23 Q. Do you think it would have been helpful to have this 24 information available in a dashboard form, as you 25 suggested? 54 1 A. With hindsight, yes. 2 Q. Please can we move to another document. It's 3 FUJ00085953. This is a 2015 "[Post Office Account] 4 Problem Management -- Problem Review" and the abstract 5 says: 6 "This report contains the trend analysis of the 34 7 problem records raised in the [Post Office Account] 8 Problem Management TfS database during 2015." 9 The TfS database, is that the TRIOLE -- 10 A. Yes, TRIOLE for Service. 11 Q. Can you recall when this type of annual review was first 12 conducted? 13 A. Possibly '13 onwards. 14 Q. Do you know why it was implemented? 15 A. Because I wasn't -- I didn't have that information, so 16 I requested it to commence. 17 Q. What did you want that information for? 18 A. So I could do a review of the problems and then I would 19 have effectively a year-on-year view of what was going 20 on so I could trend at a much bigger scale. 21 Q. Was this an internal document or was it shared with Post 22 Office? 23 A. I think it's internal. 24 Q. That document can be taken down, please, and if we can 25 go to POL00029084 and if we could go to the email at the 55 1 bottom, please, this is an email chain in September 2010 2 from Gareth Jenkins. Did you know Gareth Jenkins? 3 A. Yes, I did. 4 Q. Did you work with him? 5 A. I worked with him on occasion, yes. 6 Q. His role at this stage was distinguishing engineer. 7 What does that mean or how did you see his role in 8 Fujitsu at this point? 9 A. I saw him as a 4LS, so fourth line support. He was 10 an architect and an SME. 11 Q. Here he's referring to a receipts and payments mismatch 12 issue and he's attached a document. This has now become 13 known as the receipts and payments mismatch book. Are 14 you aware of the nature of that book? 15 A. I am at a high level, yes. 16 Q. Just for ease, could you give your high level 17 description of the book and how it operated? 18 A. So I think effectively ... I'm not sure how to put it 19 into words but -- I'm going to say no then. 20 Q. It was a case, was it, where postmasters would try to 21 put -- to do a trial balance, so they wouldn't do 22 a complete balance but would try to do a trial balance, 23 and there would be a discrepancy that they were asked to 24 put into -- whether they wanted to put it into the local 25 suspense account. Does that sound right, so far? 56 1 A. That sounds about right. I was going to say it had 2 suspense account. 3 Q. Then if you cancelled at that stage and you were taken 4 back to another screen where you're given various 5 options, but when they cancelled, in the local cache, 6 the counter's own system, the discrepancy was zeroised. 7 Does this sound right? 8 A. Yes. 9 Q. So far. The problem was if they rolled over again from 10 that point there would be -- the fact the discrepancy 11 had zeroised would be essentially recognised and there 12 would be a discrepancy between what the counter showed 13 and what was in the actual back end systems. At that 14 a high level, does that sound -- 15 A. At a high level I think that sounds ... 16 Q. So going back to one of the problems we said earlier, in 17 terms of problems, that's really quite a significant 18 problem. 19 A. That's a very significant problem. 20 Q. In this email, the third paragraph down, Mr Jenkins 21 says: 22 "We probably need to formally raise this as 23 a problem with POL. I'm not sure how this is done, but 24 presumably you can initiate that. We should then plan 25 to do the initial analysis and provide POL with a view 57 1 as to the scope and then agree how to progress it." 2 Why do you think it is that a senior member of 3 Fujitsu's front line staff was not aware how to formally 4 raise a problem at this point? 5 A. I can't comment on that but yes. No, I can't comment on 6 that. 7 Q. Are you aware of any steps that Fujitsu took to make its 8 own staff aware of the problem management process? 9 A. I think that it is done not necessarily by a broadcast 10 or a communication but the wider account are aware that 11 the service team have a problem management function 12 because they are involved, shall we say. Third line, 13 fourth line are all aware of the problem management 14 function and, as I say, they do support it. So I'm not 15 sure why Mr Jenkins didn't know how the process worked. 16 Q. If we go to the top of this email, we see that Mark 17 Wright subsequently sends you this? 18 A. Yes. 19 Q. Were you the problem manager? 20 A. I then pick this up, yes. 21 Q. In broad terms, do you remember how the problem was 22 handled? 23 A. So I think from that point on we raised the problem 24 record. We got -- Gareth was more heavily involved to 25 do his analysis. When we discussed it, my feeling was 58 1 that we would need what I called a White Paper, just 2 a paper that would go through in detail because -- as 3 I failed to articulate the summary of the scenario back 4 to you is because in -- I think it was quite complex and 5 so what I asked Gareth to do was to actually write it up 6 in a White Paper so effectively when we communicated 7 with Post Office we were clear in what the issue was, 8 what the scenarios were and where we'd got to with our 9 investigation, and it was that paper that I subsequently 10 shared with Post Office. 11 I think thereafter there were multiple 12 conversations which Gareth and occasionally I was also 13 party to. 14 Q. Can we leave that there then thank you and move to one 15 final -- sorry, penultimate point, and it concerns 16 a briefing to the NFSP. Please can we bring up 17 POL00002091. This is a Fujitsu NFSP briefing on 4 July. 18 If we turn to page 2, you see in the right-hand column 19 that you're there to talk about capacity management, 20 transaction monitoring and event management as well as 21 major incident history. 22 Do you recall why you were asked to give this talk 23 to the NFSP? 24 A. I think it's possibly because there had been in that 25 particular period or leading up to that period a number 59 1 of major incidents and I think it was effectively to 2 confirm that and acknowledge that we'd had them, what 3 had been done about them and, effectively, how many 4 actions came out of them and to give them a level of 5 assurance and also to allow them to pass on any feedback 6 and comments around each of those scenarios -- each of 7 those incidents, rather. But, again, a long time ago, 8 so I think. 9 Q. Do you recall, when you were drafting the content of the 10 briefing, was that reviewed by anyone or was it your own 11 work? 12 A. I think it was my own work. 13 Q. Did the Post Office have any input into the briefing? 14 A. I don't believe they did directly. 15 Q. You say "directly". 16 A. I may have discussed it with whoever at the time to give 17 them an overview of what we were doing and whether they 18 were comfortable with that. 19 Q. Do you recall from anyone having any pressure put on you 20 as to what should go in the briefing? 21 A. No. 22 Q. Can I ask the screen to go to page 48, please. This is 23 part of the problem management section on which you gave 24 a presentation and you provide a problem report and you 25 give the example of counter transaction processing. The 60 1 summary states: 2 "Last week we analysed the milliseconds each 3 transaction takes and found an issue in the recent 4 version of the IBM Tivoli software that has affected 5 counter transaction performance. 6 "Most of the counters have this version of the 7 software. We are still well inside SLA but it is not as 8 optimum as we would usually prefer." 9 So is it a fair summary that this problem that's 10 been picked here is that transaction times were 11 intermittently taking longer than expected? 12 A. Yes. 13 Q. How much longer are we discussing in this case? 14 A. Again, I'm going to suggest it's in the milliseconds 15 and, as is alluded there, we are within SLA but it's 16 something we picked up and we are informing both Post 17 Office and the National Federation of SubPostmasters. 18 Q. Can you recall why you chose this as the example 19 problem? 20 A. Possibly because it's a proactive update. 21 Q. Did you discuss problems such as the receipts and 22 payments mismatch book? 23 A. I don't think so unless it's in the list in the -- 24 sorry, in the presentation. 25 Q. In the presentation. 61 1 A. Apologies. Then not. I don't know and I can't recall 2 what the scope was, whether I just went back six months, 3 whether I went back whatever. But, no, it wasn't 4 mentioned if it wasn't on that slide. 5 Q. Can you explain why you use this problem rather than 6 a more significant problem like the receipts and 7 payments mismatch book? 8 A. No, I can't. 9 Q. Do you think choosing this as a problem was a fair 10 reflection of the problems in the Horizon IT System? 11 A. I think at the time possibly I was trying to get 12 a balance because, again, as I say, we'd gone through 13 a period where there had been some major incidents and 14 I was trying to give some closure and some level of 15 confidence that we were kind of over that period and 16 this particular one would have been -- I don't recall 17 exactly but possibly would have been fairly current. 18 I can't say at this moment where either of the two other 19 investigations you're referring to where they were, 20 whether I had something substantive to be able to 21 provide. 22 Q. This was in 2012, this presentation. The receipts and 23 payments mismatch book was in 2010? 24 A. Yes, so I don't think the presentation would have gone 25 back two years. 62 1 Q. That's what I want to deal with on problem management. 2 Just before I finish, I wanted to go back to a document 3 I took you to right at the start. It's FUJ00058375. 4 Forgive me for the jumping around in the chronology. 5 It's the "Direct Interface Testing Specification". 6 Please could I ask that we turn to page 11. So I took 7 you earlier to what it said about PinICLs, PinICLs being 8 recorded. This says "IT SERVICES", at the top. This is 9 the next paragraph: 10 "IT SERVICES will fax details of incidents raised 11 to Pathway ..." 12 Now, where it says IT services there, is that 13 referring to Post Office? 14 A. I believe it is. 15 Q. It says: 16 "IT SERVICES will fax details of incidents raised 17 to Pathway, if any incidents are found to be software or 18 hardware faults these will be entered into PinICL. 19 A copy of the PinICL report will be faxed to HAPS." 20 Is HAPS Post Office? 21 A. I think HAPS is external in Farnborough. 22 Q. If we look at the -- it may help to look at page 3, just 23 for your assistance and the abbreviations. 24 A. Yes. 25 Q. "HAPS (POCL) Host Automated Payment Service". So is 63 1 that a branch in Farnborough? 2 A. Memory may be vague but I think HAPS themselves were 3 based in Farnborough. So I think the initial -- no, 4 I could just be completely wrong here. It could be that 5 IT Services provide us the fax, we then raise the 6 PinICL, and then send it back to them but then I don't 7 know why it's worded that way. 8 Q. So is the wording being raised a PinICL -- sorry, they 9 inform Fujitsu of the problem, a PinICL's raised and the 10 PinICL's faxed back to Post Office? 11 A. Yes, why it says "HAPS" and it doesn't say "IT Services" 12 I'm unsure because HAPS, to my mind, is effectively 13 third party and is Farnborough-based, rather than IT 14 Services which would have been back to Post Office. 15 Q. Can we turn to page 12 of this document and down to 16 section 7. It says "Responsibilities test sign off": 17 "This will be via a handover meeting at which the 18 interfacing systems will give their approvals." 19 We've got POCL, PDA, Pathway. Could you just 20 explain what that -- was it the case that all three had 21 to approve the test scripts? 22 A. Effectively, yes. 23 Q. Thank you. I have no further questions but there are 24 questions from -- may I just turn my back for a moment? 25 A. Yes. 64 1 MR STEVENS: Sir, there are questions from Core 2 Participants, I believe. 3 SIR WYN WILLIAMS: Certainly. At the moment, I've just got 4 that last document on the screen so I can't see you 5 sadly. 6 MR STEVENS: I'm sorry. 7 SIR WYN WILLIAMS: Who is going first then? 8 MR STEIN: None from us, sir. 9 Questions by MR HENRY 10 MR HENRY: Thank you, Mr Bansal. Could I just ask you to 11 reflect on your role as a problem manager and then 12 a service SDM. When were you a problem manager? 2010? 13 A. Yes. 14 Q. Right. What I'm going to be suggesting to you is that 15 the White Paper was, in fact, to do with the receipts 16 and mismatch bug. You can't recall that specifically 17 but if a document were to arise that in fact establishes 18 that fact, you would not dispute it, would you? 19 A. I wouldn't dispute it. 20 Q. No, and so therefore that can be dealt with another 21 witness. But the fact is there were a number of 22 problems, both when you were a PM and an SDM, to do with 23 Horizon Online. You must have been extremely busy. Do 24 you agree? 25 A. It was the role, it was the job. 65 1 Q. Now, in 2010 -- there's a document that I'd like to put 2 up which is FUJ00084531. Now, can we just have a look 3 at this. Page 1 we can see -- and thank you. Could you 4 scroll so that we can see the -- yes, thank you, just 5 there. 6 Can you just make the screen a little bit smaller, 7 please? That's fine. Maybe a little bit bigger to your 8 previous position -- forgive me. I just wanted to be 9 sure that nothing had been cut out yes. Thank you. 10 So we have at that point a number of PEAKs. There 11 are 25 on hold, correct? 12 A. Yes. 13 Q. Three are impacting. Does that mean that they have some 14 adverse effect on the system because they are impacting? 15 Is that the description given? 16 A. I'm not sure if they're out for impacting; so to be 17 assessed. 18 Q. The fact is that "impacting" often has a deleterious 19 connotation, doesn't it, at times? 20 A. I believe they were out for assessment. 21 Q. You've got investigation, which is -- again, it's not 22 resolved. So "hold" is not resolved, "investigation" is 23 not resolved. What does -- "monitoring" obviously 24 you're just waiting to see whether it's going to get 25 worse; is that right? 66 1 A. They are monitoring to see what is happening, yes. 2 Q. Right, exactly. "Release to live": what does that mean, 3 please? 4 A. "Release to live" means that they are being packaged and 5 distributed. 6 Q. And then "to be closed tomorrow", that means that they 7 are supposed to be sorted? 8 A. Effectively they could be the ones that are currently on 9 monitoring -- 10 Q. Right. 11 A. -- a period has been agreed and that potentially is 12 tomorrow. 13 Q. And then "waiting fix" means that there's 14 that are 14 awaiting resolution? 15 A. Correct. 16 Q. So really we've got roughly, haven't we, 60 issues that 17 you're not on top of because you've got 25 on hold, 19 18 which are being investigated, and 14 which are awaiting 19 a fix -- roughly 60? 20 A. That's what I think it says, yes. 21 Q. Thank you. Could we just scroll up, please. 22 Why is it called "prayers" as a matter of 23 interest? 24 A. I think it was just a given name. I've no idea. This 25 is dated a couple of days after I joined, but yes. 67 1 Q. Sometimes prayers means "heaven help us". I mean, you 2 can't think why there was that sort of -- was it an 3 acronym? Did it stand for anything or was it "Oh my 4 God", you know, "look what we've got to deal with"? 5 A. I couldn't say. 6 Q. You couldn't say. You never asked? 7 A. Never asked. 8 Q. Right, okay. Scroll up, please. Thank you. 9 Pat Lywood: who was Pat Lywood? 10 A. Pat Lywood was a member of the Post Office account. I 11 can't remember her exact role. 12 Q. Don't worry. Let's just scroll up a little bit further, 13 please. You see we've got Mr Godeseth there as well. 14 Do you know notice his name? 15 A. It's not jumping out at me but, yes, it will be there. 16 Q. Don't worry. Torstein Godeseth. Carry on, please. If 17 you scroll up. And a little bit further up, please, so 18 I can get to the text of the message, if you please. 19 Isn't a suggestion that here you are being asked if 20 there's anything you can do to speed things up? I think 21 there's some further text and delay. Carry on, please. 22 A. Yes, I think you can see it's a communication to -- 23 Q. Sarah P, ENT: 24 "Sheila, please could you take a look at the ones 25 on you and try to resolve some of the hold investigation 68 1 ones." 2 So this was basically a constant battle, wasn't 3 it? 4 A. It was an update communication to everyone to see if 5 they can address any PEAKs to speed things up, yes. 6 Q. I mean, a constant battle with instability and errors, 7 wasn't it? 8 A. It was a call to address PEAKs, yes. 9 Q. 60 of which were, you know, as you've already said, 10 remain to be resolved. 11 Could we now move on please to POL00029493, 12 please. 13 I'm so sorry, I thought that that was notified. 14 So it's 00029493. That's interesting because I actually 15 have it on my system and it's one of the ones that I was 16 allowed to ask. Don't worry, we don't need to put it 17 up. 18 I just want to ask you, please, just very -- were 19 you aware that this was a retro-engineered system? The 20 way in which it had been eventually allegedly made 21 acceptable was to reverse engineer it. Rather than 22 prospectively design a logical system, it was 23 essentially dealing with a number of problems and trying 24 to reverse engineer it but at the same time, when that 25 was happening, further problems were being introduced. 69 1 Were you aware of that? 2 A. I wasn't necessarily aware of that. 3 Q. You weren't told that, fine. 4 Could we move on, please, to POL00029460. This is 5 a major incident report and you're the owner. And could 6 I ask you, please, to help me. Was that data -- 7 familiarise yourself, please, with the document. Has it 8 been shown to you in advance? 9 A. Possibly. 10 Q. Do make yourself familiar with it. Tell me you're -- it 11 would help if you could let the officer know who is very 12 kindly assisting in scrolling up. So you just let her 13 know when you need to read more of the document. 14 (Pause) 15 A. Could you scroll up. 16 Q. I would like you to concentrate, please, on the first 17 three paragraphs within the box "analysis of problem". 18 Obviously, you must read everything but I want you to -- 19 of particular interest to "Post Office Limited" down to 20 "1 February". 21 A. Are we also able to go down to the root cause? 22 Q. Yes, of course. Please do. I'm not trying 23 to ... (Pause) 24 A. Okay. Then if we could go back to that section. 25 Q. Of course. Of course, Mr Bansal. 70 1 Right, now my question is that this is described 2 as an update but if you see the second paragraph in 3 analysis of the problem: 4 "This was a second iteration of the data as 5 a problem had been identified with the initial dataset 6 that had been supplied during validation of the token 7 data. The CTO update was in effect a primary package 8 with an incremental update." 9 Was the data update in fact actually a fix because 10 of the problem identified within the initial dataset? 11 A. I couldn't tell you. 12 Q. You couldn't say. 13 A. I couldn't say at this moment in time. 14 Q. Fair enough. Don't worry. I now have permission, sir, 15 to refer to POL00028830. We can see the date of this. 16 This is 28 September 2010 and it relates to PEAKs 17 PC024765 and PC0204263 and then 64 and 63 is again 18 mentioned. 19 Can I just ask you, please, to look at this 20 document which I have permission to put to you. So, 21 again, if you could inform the officer who is presenting 22 it. (Pause) 23 A. Okay, if we could go up. 24 Q. You can see there, can't you, receipts and payments 25 mismatch. If you go back, PC0204263 describes a problem 71 1 with SU balancing that will result in a receipts 2 payments mismatch. So given the fact that is authored 3 by Gareth Jenkins, we can fix his knowledge as to that 4 problem in September and isn't that the White Paper that 5 you were -- 6 A. I believe this is the White Paper I was referring to. 7 MR HENRY: Well, I'm very grateful, sir. Thank you very 8 much for your time and your patience. 9 SIR WYN WILLIAMS: Are there any other questions? 10 MR STEVENS: No, sir. That's everything, thank you. 11 SIR WYN WILLIAMS: Well, I'm very grateful to you, 12 Mr Bansal, for coming to give evidence to the Inquiry 13 and also being very flexible about the time when you 14 started giving evidence and the progress of your 15 evidence and that's helped us to move along efficiently. 16 So thank you very much again. 17 A. Thank you, sir. 18 MR STEVENS: Thank you, sir. That concludes today but we 19 have Steve Muchow tomorrow. 20 SIR WYN WILLIAMS: So that will be at 10.00? 21 MR STEVENS: Yes, sir. 22 SIR WYN WILLIAMS: It is just Mr Muchow tomorrow? 23 MR STEVENS: Yes, it is. 24 SIR WYN WILLIAMS: Fine, thank you. See everyone tomorrow. 25 Goodbye. 72 1 (1.42 pm) 2 (Adjourned until 10.00 am the following day) 3 I N D E X 4 STEVE BANSAL (affirmed) ..........................1 5 Questions by MR STEVENS ..........................1 6 Questions by MR HENRY ...........................65 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 73