Official Report 413KB pdf
I welcome everyone to the Justice Sub-Committee on Policing’s second meeting of 2016. I ask everyone to switch off mobile phones and other electronic devices.
I have received apologies from Alison McInnes and Margaret Mitchell, but—by no mean feat—I welcome Graeme Pearson to the committee.
Item 1 is on information and communication technology systems. I welcome Martin Leven, who is the director of ICT at Police Scotland; Chief Superintendent Hamish Macpherson, who is the i6 programme director; and Chief Superintendent Alan Speirs, from the contact, command and control division. You have all been here before, so you know how the whole thing works, which is excellent. If a member has not asked you a question directly and you want to say something, just indicate that and I will call you.
Thank you for your written update on Police Scotland’s various ICT projects. We will go straight to questions. John Finnie has a look on his face. Does that mean that a question is brewing?
Good afternoon, gentlemen.
A number of weeks ago, I was told that this was all off the rails—that it was not going well. Can you please give us a general overview of where things currently stand?
Are you talking about i6?
Yes.
I was about to clarify that, because everything else is doing well.
Were there other things on the rails? [Laughter.] Perhaps I should not have intervened.
According to one of this morning’s national papers, we are about to cancel i6. That is not a fact; i6 is currently in commercial renegotiation with Accenture.
When I appeared in front of the committee in September last year, there had been a delay caused by a problem with hardware—I gave you a very long description of that hardware. That issue was rectified very quickly, within the timescales that we expected. However, around that time it became clear that the standard of some of the coding that we were receiving from the third-party supplier for the i6 build was not anywhere near a standard that we were willing to tolerate. Since then, we have been in discussions with Accenture. Last week, a meeting was held with Accenture. Unfortunately, I was unable to attend that meeting due to personal illness, so I ask Hamish Macpherson to pick up on where we are with that.
First, I will give a bit of background. In June, we received for user-acceptance testing a product on which product testing had been completed. Members will remember that I gave the committee a long and very drawn-out description of how the supplier’s test cycle should work before we get the product for user-acceptance testing. According to the agreed exit criteria, by the time it left the product-testing stage it should have had a maximum of 80 defects, none of which should have been major or critical errors. However, very quickly after we went live into user-acceptance testing, we saw that the number of defects was significantly higher than we wanted.
What was the number?
At one point, we ended up with 12 critical errors in the system—errors that stop one progressing through the system—when there should have been none. There were a number of other errors, too. It is almost pointless to get into numbers, because some of the errors are cosmetic.
The point is that some of the errors were critical.
Yes. There were a number of critical and major errors that were stopping us progressing through the application. At that point, the supplier was struggling to get rid of the errors: as they fixed one, another error was produced.
As a result, in September—just after we met here—it became obvious to us that we were not going to make the December date. In September, the supplier confirmed that. We asked it to carry out a root cause analysis of why that had happened and the reason for the level of deficiency. As I said in the submission, the supplier re-engaged with Police Scotland, and said that it had carried out the root cause analysis but wished to move into a without prejudice period. We are still in the without prejudice period.
What does that mean?
Basically, it means that the information that the supplier gave us was subject to commercial embargo: there is a limit to what I can say to the committee about it in a public meeting because we are still within that period. The supplier shared with us its assessment of what had caused the defects and errors in the system and we worked with it on a plan for moving forward. The original plan that the supplier provided to us was too long and would have required additional finance from Police Scotland. That plan was rejected.
Wait a wee minute. We were told that
“a contract variation agreement was signed in April 2014 protecting the delivery of the original requirements at no additional cost to Police Scotland.”
We clung on to that.
Indeed—but we are still there. The original plan that the supplier provided to us asked for additional finance. Also, the timescales did not work out for us. There is little point in my going into the timescales here—you will appreciate that 300 person days to fix something is 300 people for one day or one person for 300 days. so the timescale was not suitable for Police Scotland. Moreover, an injection of additional finance was certainly not suitable for Police Scotland.
As a result, a summit meeting last week was set up by the chair of the Scottish Police Authority. Accenture, the SPA, the Scottish Government and Police Scotland attended that meeting. As a result of that meeting, we have a potential positive way forward that still involves no additional injection of cash from Police Scotland or the SPA. However, we still have to work through the details and timings of that.
The original solution that was proposed by the supplier was to fix everything and deliver it at a later date. Our proposal, which we are working through with the supplier, is to sort it on a modular basis. That way, we can get modular releases of different parts of the application—crime, criminal justice and custody—over the next few months.
Do we have the right people sitting in front of us? Clearly, we want the practitioners, who will advise what is to be designed, and we want senior management. Should we have a lawyer here? Do contractual issues run alongside the present challenges?
We are having contractual discussions with the supplier. We are not at the lawyer stage yet. What is on the table is a potential way forward. We have just got to work out the detail of that way forward, which we think will take eight to 12 weeks.
At what point will we move beyond things being deemed to be commercially confidential? Our obligation as a committee is to understand the financial arrangements. Do you consider every commercial matter in confidence?
We will end up walking out of the without prejudice period. A without prejudice stage is only a stage that allows you to have open negotiation without putting yourself in a situation in which you have a commercial problem. It means that the supplier can share with us stuff that it would not normally want to share with us. That period will end—the intention is that this is time bound, so there are four-weekly checkpoint meetings for the next 12 weeks with the people who attended the summit—the SPA, the Government, Police Scotland and Accenture senior management. At the end of that, we will exit the without prejudice period with a positive way forward.
Is that a new initiative or was it built into the contract? We all understand that these are incredibly complicated areas. Was some latitude allowed around timeframes?
Every contract has within it a remedial process. It is an Office of Government Commerce model contract, so it is just following guidance, if you like. If we miss a critical milestone, which is what happened—the go-live milestone in December was missed—we move into a period of mediation in which we look at how to work that through with the supplier. That is where we are now.
What could go wrong? If the situation cannot be resolved, what will that mean? Will the whole thing be kaput?
I like to think that we now have a positive way forward. If we cannot reach a positive way forward, we will have to consider options. There is always the option to part company, but at this point we are not currently considering that.
Can I have a final question please, convener? It is in relation to other decisions that have been made, including on control room closures—
Before you move on, I want to talk more about i6.
The question is entirely related to i6.
Okay.
Thank you, convener. If there is the potential for future problems with i6, this is clearly not the time for important decisions to be taken about command and control and other issues, is it?
I will pick that up. I am keen that the committee understands the matter clearly. The i6 project does not have any major impact on the control room project, which is a separate project using completely separate technology and—
There must be a link.
There is a link, but that is within our gift and under our control. As part of the i6 project, we created a module—it is mentioned under the general IT update in the committee’s pack—called the integration data archive project. We are using that as a switchboard for information between a variety of Police Scotland systems.
We have more than 700 different systems. We are hoping to reduce that number significantly as finance allows us. In our team, and within our gift, we have created an innovative approach to how we are going to exchange data, because otherwise we would always be stuck in the situation where we are beholden to suppliers to integrate products with any new systems that we bring online. That would be a dangerous place for us to be, so we have created technology that will allow i6, or whatever we bring in for the deliverables, to store information. In C3—Alan Speirs’s area—there will be completely separate information storage. We will have a switchboard that will exchange information at the right place and at the right time between the two systems. We are not producing one massive IT system for all Police Scotland; our approach is modular. I hope that explains the situation a little bit more clearly to you.
Thank you.
I want to go back to the costs. There may be something about those within the contract. As you have said, if there are any additional costs in the contract, they will not pass to Police Scotland. What happens about consequential costs to Police Scotland because of delay? I presume that the project is being introduced to deliver not only more work-related efficiency but more financial efficiency. There has been extensive delay, so is there anything in the contract through which Police Scotland will be able to recover additional costs to it from Accenture?
The contract as built has penalty clauses in it, in which financial penalties—
That is different.
That was the way that the contract was built. The matter was discussed at the summit. Again, I will pass to Hamish Macpherson to provide the detail.
Again, I do not wish to appear to be evasive in any way, but there is a limit to what I can say because we are in the middle of commercial and contractual negotiation, so some of the information is obviously commercially confidential and we are in a without prejudice period, as I have explained. I can assure you that although there are indirect costs for Police Scotland as a result of the delay, those are forming part of our negotiation with Accenture.
So, have you boxed in any direct additional costs within the contract frame, and any costs because of delay are in addition to that? That is what I am getting at.
The direct and indirect costs of running the project are part of the financial envelope that forms the i6 contract. It is not just, if you like, to supply the software but to supply the services with the software. Our discussions with Accenture include our direct losses as a result of the delay.
But those losses are a negotiable part, whereas the costs within the contract are not negotiable because they are written into it. What I am getting at is whether there are additional costs to Police Scotland for
“protecting the delivery of the original requirements”.
I want to know if that is fixed, which apparently it is.
You are talking about the contract variation agreement, which was done earlier on.
Yes.
That position has not changed, and there is no additional cost—
Am I correct in saying that part of your negotiation in principle includes any indirect costs, which you may or may not have to moderate or write-off and so on?
I am confident that the indirect costs will fall within the settlement that we agree with Accenture, so there be no additional costs.
Can I have an answer to that? It was a simple question. Because of all the additional costs that you have, part of the negotiation will be you asking the deliverer what they will do for you. The costs are part of a negotiation; they are a flexible bit of those negotiations. You may write-off the costs, you may have to moderate them, you may get all the costs. Am I right?
Yes, they costs would be part of the negotiation.
That is what I wanted to know.
I am sorry—that is what I was trying to explain. The costs will be part of the negotiation.
So, that is not commercially confidential. It lets us know that there could be a cost to Police Scotland, if you have to write off those costs.
13:30
Yes—but I am by no means saying that that is what will happen. That is certainly part of our negotiation and our supplier accepts that there are direct costs for Police Scotland, which will have to form part of that negotiation.
We cannot know what that figure might be.
This has probably been explained before, but all the figures in the—
No. I am talking about indirect costs—additional costs to Police Scotland outwith the contract.
To have a discussion regarding that, I would have to give you the commercial costs of the contract, and the contract forbids me from doing so—certainly in public. Obviously, that would allow other suppliers to know exactly what the costs are.
I might come back to that in due course.
We are at a stage of without prejudice negotiation, where Police Scotland and the contractor are still speaking. The contractor has not walked away, as has been reported. Is that correct?
That is correct.
How long do you expect the without prejudice period to last? It cannot be indefinite.
I imagine that we will get it concluded in between eight and 12 weeks. I am not saying that we will get all the commercial issues sorted in that time, but we will know the way forward and be able to form some heads of agreement between the two organisations in order to move forward with the programme.
So, in three months we should have a clear indication of the path forward for the i6 project.
Indeed.
I want to ask about some of the things that you have said. I am by no means an IT expert—I do not think that any committee member is—but I have dealt with IT contracts, and there are certain phrases that bother me. You have talked about code deficiencies. The other day, I read a BBC article about what is deemed to be one of the worst-ever computer games, which was made by Atari. It was called “ET the Extra-Terrestrial”, after the film.
I am wondering where this is going.
I will tell you exactly where it is going in two seconds. One of the issues with that game was that the software designer had very little time to deal with it. There were a number of code deficiencies, which meant that ET, rather than going through the game as he should, kept falling into holes, which of course annoyed folks playing the game. You talk about code deficiencies in the project. Does that mean that we would have a situation in which an operator would fall into a hole at a certain point and be unable to move forward with the software? What implications would that have for the ability to process the necessary things, say in the crime section? What would be the repercussions of that kind of thing? Explain code deficiencies to me in simplistic language, like the BBC did for “ET”.
I will give that to Hamish Macpherson to explain to you.
I first want to reassure you that that would not happen. The whole purpose of the complex test cycle that we have described to the committee is to ensure that the system is fully tested before it ever reaches an officer on the street or a member of police staff in a back-room function. The thing was built and a completion certificate for the build was issued on 10 December 2014. Since then, we have been in testing. The system was always going to be thoroughly tested; it is the testing that has identified the deficiencies.
The biggest issue that we have is with ISCJIS—integrated Scottish criminal justice information system—compliance, which demands that the data be put together in a certain way. For example, a charge can have one or more accused, an accused can have one or more compensation orders and compensation orders can have one or more bodies seeking compensation. Because it is such an integrated system, the crime automatically follows through to criminal justice, via reporting through ISCJIS. The underlying problem is how data has been put together, which makes it a fairly major problem to resolve, for the supplier. That is the level of code deficiency that I am speaking about.
Are you saying that the major flaws are only in that area of the system, or is that just an example?
There is more than one area. In the paper that is before the committee, I identify four areas where there are flaws, but for me criminal justice is the biggest area where there are flaws in the system.
For the record, can you describe in layman’s terms the four major areas where there are difficulties and what those difficulties are?
Yes. I will try to do that. I was hoping that I was speaking in layman’s terms.
Criminal justice obviously also relates to the crime area, because the crime information gets created in the first place and it has to be ISCJIS compliant. That then passes through criminal justice.
There are also some deficiencies within the audit module. We must ensure that we know what officers have done in the system—it is critical that that gets audited. That area is less problematic than criminal justice—there are multiple small defects, but there are still defects.
Bits of administration do not work. That affects our ability to administer the system—for example, to add to constrained lists and so on within the system. Some of the functionality just did not work, and some of the search functionality did not work as specced. Those areas are smaller than the criminal justice area. The criminal justice area is the biggest area that requires to be fixed, if you like, and, because that sits at the database level, it is quite complex to fix.
When you appeared before the committee previously, I talked about system failure that I had come across in the past when too many bells and whistles were added on to a system, often at the tail end of its design. Have there been changes from your side during the project that have maybe caused some of those difficulties, or is this entirely a situation in which the developer—the contractor—knew all the requirements at the very beginning?
As I have explained, the biggest issue is the criminal justice issue. I think that version 1 of the ISCJIS data standards came out in 2007, and that is the biggest issue, so—
That does not answer my question. Was the contractor, Accenture, aware of the requirements for all of this at the very beginning of the contract?
Yes. It is a contractual obligation to comply with ISCJIS, which is named. It is written in cold English that it is necessary to comply with ISCJIS data standards as part of the contract. I am not going to say that there has not been change in the contract—just as there will have been change in the contract for this very building, for example. There have been changes to legislation and so on, so there have been some change requests in relation to the contract, but fundamentally those are not what has caused the issue.
What I am trying to get at is that you have made no major change to what the contractor needed to develop since the very beginning. Has there been any major change?
There has been no major change that would have caused the defects.
So Accenture was well aware of the requirements of the system at the very beginning when the contract was signed.
ISCJIS data standards are named and must be complied with within the contract.
What does that mean?
The contract actually says in cold English, “You must comply with the integrated Scottish criminal justice information system’s data standards”. Every copy of the manual has been supplied to—
We understand about the complexities of criminal justice, as there can be multiple accused and various bits of information about previous convictions and so on. Was all that in the specification at the very start?
Yes, it was in the specification from the start.
It pains me to go here, because I think that you will probably turn round to me and say, “I cannot answer that because of commercial confidentiality or because we are in a without prejudice period.” With hindsight, do you think that the contractor is capable of carrying out all the instructions that it was given to develop the system?
At this stage—again this is a without prejudice thing, but you are asking for the personal opinion of a couple of individuals who are sitting up here—I have doubts about the capability of the contractor to deliver this going forward. The contractor will need to prove to us as part of the negotiations that it is completely capable of delivering this, because it has very clearly let us down and put us in the situation that we are in.
So you have doubts. Do you think that Accenture will be able to retrieve your confidence in its ability to deliver the system?
It is a possibility that it could retrieve our confidence, depending on what it comes back with. I certainly imagine that, as part of any negotiation going forward, there will be a change in certain of the personnel who have been involved in the project at the supplier end. If a robust plan can be produced, we will look at it, but I again emphasise that a robust plan was produced the first time round.
If we reach a situation in which the contractor is found to be incapable of delivering the system that is required and you have to seek another contractor to carry out the work, will the work that has been carried out thus far go to the new contractor so that it does not have to start from scratch, or will we be back at the very beginning of the process? That is important to know. In addition, I would like to know—if you can share this with us—what the penalty for the contractor will be if it is found to be unable to deliver the contract.
I will probably not be able to share with you what the penalty is—
I thought that that would be the case.
It is still to be negotiated.
That said, I can confirm that a lot of the work that has been done as part of the journey that we have been on has been done within Police Scotland by the team that reports through Hamish Macpherson, so we have a lot of internal intellectual property, if you like, relating to how any future system would work. I will pass over to him on this, but I would be surprised if any of the code work that Accenture has done would be passed on to another party. I do not think that we would want that to happen if we reach the stage—I repeat that it is a big “if”—of looking at an alternative supplier.
Before Chief Superintendent Macpherson expands on that, when it comes to the intellectual property, in percentage terms, or perhaps in terms of importance, is it the case that the important intellectual property rests in your hands rather than in the hands of Accenture?
I will let Hamish Macpherson deal with that, because he has been more involved in the legal side and the contracts, but the workflow intellectual property, which is the important part of the process, is absolutely with Police Scotland. We designed that. It is my understanding that the coding is the intellectual property of Accenture, but I will pass over to Hamish to correct me if I am wrong.
There are probably two points on which I can reassure members. Since September, when it was first declared that there was a potential issue with going live in December, I have not rested on my laurels. We continue to look at contingencies—including interim contingencies—in relation to what we can do to make sure that we get live ICT systems across Scotland. The integrated data archive project, which Martin Leven mentioned, is a good example. That will give us a view of crime, missing persons and so on across Scotland, but it was not predicated on the delivery of i6 because we deliberately separated things for the reasons that Martin has set out.
As far as intellectual property is concerned, I came along after the system was built. At that point in time, I looked back at what information was provided to the supplier. I must give credit to my team and my predecessor’s team, because the information that was supplied at that time was absolutely superb. The business process work that they did in advance of going out to tender for the system was superb.
However, time has moved on. New legislation has come in and other new legislation is coming. The Criminal Justice (Scotland) Act 2016 and the Children and Young People (Scotland) Act 2014 are on their way, so my team has been reviewing and updating all the business process maps, which hold the intellectual property of the journey through policing. My team has been doing that since September and those maps are now in a good state. I am hoping that we do not have to use them as a contingency, but it would have been remiss of me if I had not done some work as a contingency.
Kevin Stewart mentioned code deficiencies, which would indicate that the software engineers have made mistakes in the coding and the programming and, as a result, have not been able to produce what your system requires.
As far as you are aware, has Accenture brought in any additional expertise? Has it brought in other software engineers to try to sort out the problems?
Throughout the entire journey, Accenture has made several attempts to address issues.
To return to Mr Stewart’s point about the Atari game, a key difference was that the game was launched on an unsuspecting public without the proper quality assurance and then the problems came and people got frustrated with ET falling down holes. I cannot believe that I am sitting in a parliamentary committee talking about ET falling down holes, but I will move past that.
13:45
We have heard worse.
It is the layman in me. It was the easiest way of describing this.
Our team has been very good and very switched on. I am incredibly proud of the effort that has been put in by our team to spot things and to have a zero-tolerance approach to the development of the product.
I clarify that we have not launched the product; we have deliberately taken steps to ensure that we do not launch anything that is not fit for purpose. We talk about the software engineers making errors, and Accenture has teams of hundreds working on the product at various locations. When Accenture started experiencing problems, the i6 programme board, which I am a member of, was fairly robust with it.
We received assurances and Accenture drafted in external quality assurance people to take a look at the quality of the product that it was delivering. We have gone through several changes in the senior management people at Accenture, who have taken different approaches to running the project. It is more than just software engineers making errors—
It is coding errors—
I personally think that the environment in which they are working could be the problem. By that, I do not mean the person-to-person environment; I mean the technical environment—the actual structure that has been set up for them to do the coding in and the testing within that. The standard of that environment varies from what we would probably have been happy with. I will pass you on to Hamish Macpherson to provide a little bit more clarity on that.
Yes; I can probably add a bit more context as well. After we began to discover the defects, which were not getting sorted during user acceptance testing, we asked Accenture to carry out an independent assessment. It did that—it brought in an independent team. It is obviously a large organisation with hundreds of thousands of people and it brought in an independent team of people. I am completely happy that they were very knowledgeable about their subject.
They shared stuff with us without prejudice, which is awkward for me. They will eventually come back with a with prejudice update, which I will take back to the i6 programme board. However, they definitely identified some deficiencies in the management of the software production, which have caused some of the issues.
That is one of the problems within the industry—people move a lot. There is a shortage of software engineers and it is relatively easy for people to move from one job to another. The company may be losing expertise because people have moved on to somewhere else.
Are you confident that, if you have an on-going contract, the company will be able to cope with the changes that are bound to happen as a result of new laws that are passed? Different data could be required to be collected and so on. It might be okay when you launch i6 but as things develop, you will need to ask the company to add to the systems that it is producing for you. Are you confident that those additions will work?
I repeat what I said earlier. Personally, my level of confidence in the supplier at this stage is such that I would be very wary of extending the contract unless it can prove that it can up its game.
Are there other companies out there that could take on the work?
Yes.
You say that there are other companies. Can you explain in a simple way why Accenture was picked to do the job? What criteria were used? You said that you gave it all the specifications and so on, and yet—notwithstanding the movement of software engineers or anything like that—it has failed on lots of things. You have been left dealing with the flak when it was the developer that failed.
What process was gone through to give Accenture the contract? What other companies were looked at? Are there better people out there? You look absolutely fed up with it all, Mr Leven, and I am grateful to you for being so frank because I think that you just want out of it. I may be putting words in your mouth but I think that you could do without all this. What process did you go through, who did you look at and why was Accenture chosen?
No one at this table was involved in the original process that was gone through. I joined the policing family and the Scottish Police Services Agency towards the end of the process.
But you will know how it happened, so could you describe the process?
I know exactly how it happened. The process was probably the best procurement process that I have witnessed. It involved incredible levels of detail; it required guarantees from the companies; it required companies to demonstrate their ability to deliver; and it was based on a system that was already live in Spain, with something like 100,000 users. After the first stage of the process, there was a shortlist of four candidates. After an in-depth diligence process, that was reduced to two candidates. At that stage, Accenture was chosen as the preferred supplier: it had a track record, having successfully delivered the system to the Guardia Civil in Spain; it demonstrated an absolute understanding of our requirements; and, commercially, it was an attractive proposition.
So, if it was so good—top of the class—what went wrong?
Sometimes, you can get so much information out of a procurement process that, until the project is delivered, you do not know what you are getting. That is not just an IT thing; in any area of life, you can get something that is not exactly what it says on the tin.
But this is a big tin that costs a lot of money.
I fully understand that. Again, the project differed from other major IT projects in the history of policing in Scotland, in that the governance and diligence around it were extremely intense and focused. Accenture has been held to account at every stage.
I appreciate that, but that was not my question. You have told me that it has been held to account. However, I am also hearing you say that, when the Accenture contract is finished, you will get someone else in who will do a better job. My question is, if there were better people out there, why did you not use them?
I would ask the committee what else could have been done. We carried out one of the most thorough procurement processes ever. It was heavily audited and went through repeated gateway reviews. When the contract was signed and we went into production with Accenture, Police Scotland developed a zero-tolerance approach in terms of the commercial and technical coding standards that Accenture brought forward. I think that the process was managed as well as it possibly could have been.
I will let Graeme Pearson ask a short question, but then I will let Kevin Stewart come back in, as he has to leave soon.
On the point that Kevin Stewart made about expertise, I personally led the teams that delivered the national intelligence database for Scotland, which is still in existence. Thereafter, for my sins, I delivered the criminal histories system, which also still runs right across Scotland. In that context, I can commiserate with Hamish Macpherson with regard to his current responsibilities. I know that he is not responsible for what happened, but he is now accountable for where we are.
In May 2013, Mr Leven, you told the sub-committee that the longer that systems are in place without being replaced, the greater the risk of failure. In the subsequent meeting, in June, the chief constable gave us an assurance that i6 would begin in phases in 2015, and that it would commence operation from those phases. No doubt we will do a post mortem if we get to the stage at which this thing grinds to a halt. My problem is with the time that it has taken to call the company to account and know whether we are going to get a system delivered.
You mentioned a summit meeting. I know that you cannot give us the confidential figures but I take it that one of the options that was discussed at that meeting was that, in the event that Accenture was unable to deliver, a number would need to be considered and negotiated so that you could move on and deliver a system. Am I right in assuming that that was discussed at that meeting?
Again, I was not at that meeting. Hamish Macpherson can answer that question.
Yes, that was an option.
There was a degree of brightness in your presentation this morning, and I hope that the system can be delivered. However, I note what you have subsequently said, and your apparent lack of confidence in some of the people concerned.
I do not want to go through the quotations—I can do so later, if I am given time. The system was presented as being absolutely crucial to the reform of policing back in 2013. At that time, we were told that the team had already been led for four years, so we are now seven years into the project. I have been involved in such national systems, so I know that there comes a point when patience runs out. The delay is costing the service daily, in terms of commitment. Why has it taken seven years to get to the point at which we are beginning to get tough about it?
Those timelines are probably harsh, and are maybe not fair regarding the history of what happened. The contract was awarded to Accenture in 2012, and in August 2013 we got approval from the SPA to sign off the contract.
The chief constable reported to this committee that in June 2013 the team had been led by Alec Hipman for perhaps four years, so we are almost seven years into the police family’s iteration of it—albeit that Accenture came to the project later in those seven years. By the time it came into it, one would have thought that you could have said, “There’s a package. Just give us it.” We thought that we would get it in 2015. Why has it taken until now before we begin to get it? For nearly four years, we have expressed our concerns about the way forward, and we have constantly been told that it is in the bag and is coming.
I will pass on to Hamish Macpherson in a second.
The four years prior to awarding the contract were spent delivering the absolutely correct model that was required. They were spent on research and learning lessons from other public sector projects, some of which you were involved in and which I am sure that you remember. In the middle of that four-year period, Police Scotland appeared. Bear in mind that it started off as a project for Strathclyde police and was going to tie in the forces across the country; Police Scotland was not on the horizon at that stage.
It was still a single system for the whole country.
There was not one national police force, so there was a little bit of extra work involved at that stage, which meant that we went to market with a thoroughly mature and complete model of what we wanted to be delivered. We went on to award the contract to Accenture as the supplier after a lengthy procurement process. Procurement is not a fast thing; it took well over a year and half to do this—
When did they—
Please let the witness finish.
If we take a more realistic look back, from the point when Accenture came into the game it has been 18 months or two years, as opposed to the seven years that you mentioned. In that time, stuff has had to be delivered in a particular order. There has to be an in-depth dive and an in-depth technical spec of what will be delivered—how the business rules will be made to work technically. That is not an overnight thing; it is months’ worth of work, and there are several hundred people involved in turning the vision into reality.
Once that happens, it is passed on to Accenture to turn it into a technical reality, which we then audit.
When did that—
We do not see the product coming through until pretty late in the game. As soon as we see the product coming through, we can identify whether or not it will work. With any software development you will only discover whether it will be a goer towards that software’s launch date, when it goes through advanced validation of whether everything works.
I understand the point behind what you say. You have been involved in this, but I completely disagree with the way that you have turned the figures around, in terms of time.
I will pass you to Hamish Macpherson, because he is nudging me to say that he wants to come in.
I am not trying to defend myself; I am just trying to paint a picture of reality. Many of the issues that have been identified are issues or risks that, for some time, we have raised with the supplier. We received assurance and reassurance from the supplier on those very issues.
On 15 June, we eventually got the product for acceptance testing. That was as a result of its passing the product test. During the product test, the most that we did was witness some testing, because it is a supplier-led activity. We got reports that reported a success rate for the product tests in the high 90 per cents. The supplier came to the August meeting of the i6 programme board and increased its confidence assessment for going live in December from 90 to 91 per cent. Within four weeks of that, we were in high levels of defects, which is when we identified the problem.
I have looked to see what opportunity there was prior to my appointment, or even during my time, to intervene, but the reality is that we did not have that opportunity. If the supplier is saying that the unit test, assembly test and product test are being passed, and it is giving us the paperwork to support that, it is very hard to unpick that.
I will let Graeme Pearson ask the next question, and then I will bring Kevin Stewart back in.
When was the contract awarded to Accenture? When did it start the actual work?
14:00
I believe that it started in August 2013.
I have not got the date in front of me, sorry.
I believe that the Scottish Police Authority approved the contract at its meeting in August. I may be wrong, but that is my assumption.
As I said at the outset, I do not think that there is any need for Mr Macpherson to defend himself. I think that he is the tail-end Charlie in the whole process of delivering the i6 project.
We made a big play about the performance platform regime and the lessons to be learned from the loss of that system. Each member of the committee very clearly rehearsed that point. At one stage, I asked every member of the team whether they had a copy of that report nailed to their desk, so that they would learn those lessons. This seems to be almost a replay of many of the lessons that came from that report, in that the contractor has delayed your ability to test and challenge what they have offered to you.
I strongly disagree with you again, Mr Pearson. In the way that things have been managed and run, i6 could not be more of a polar opposite situation to platform—
I can see that the governance has been strict.
You raised the matter. Can you please let the witness deal with it? What does Martin Leven want to say?
The committee is well aware of the main lessons learned from the platform report and we all know them inside out. Those lessons were about the lack of governance and lack of specification. The scope of the project kept moving and, therefore, what started as a seemingly simple project turned into bells and whistles as every part of policing across Scotland—the eight different police forces—wanted a wee bit added on at different stages of the project.
As Hamish Macpherson said, the scoping of i6 was nailed at the procurement stage. To the best of my knowledge—and Hamish Macpherson will back me up on this—the only changes that we have put into that scope are the result of legislative changes that have come out of this building, but part of the contract was that that was likely to happen. As Hamish said, for us to immediately start challenging Accenture represented quite a fast turnaround, since our first chance to look at that was in June of last year.
I will just make this comment and then let you go on—
Hold on a moment. Did Hamish Macpherson want to add something?
No, I was probably going to add something very similar.
I will just leave it there. I know that Kevin Stewart has another appointment to attend.
I just want to say one thing.
No, gentlemen, please. It is just that Kevin Stewart has to go off to something else, so I will let him come in. Graeme Pearson can come back in after him.
Let Graeme Pearson make his comment, and I will come back in afterwards.
Okey doke, that is fine.
You said that it was a fast turnaround, but I will go back to the commitment that was given to the committee, in June 2013, that the system would start operating at the beginning of 2015. Back then, we were given the indication that it would begin to be launched just after Christmas 2014. We have lost over a year, and we are getting to the crux of it now.
Indeed, but you are going back to a committee meeting that happened a couple of years ago. I think that this is my fifth appearance in front of the committee over that time. Through written submissions, we have given the committee regular updates throughout the process.
I know that you have been giving full updates.
The committee knows the full history, and I am worried that, if I start replaying the entire history to you, we will not have time for any other questions between now and when we finish. That history has been played out repeatedly.
As Mr Leven just said, he has appeared in front of the committee a number of times. One of the things that we discussed during the course of those meetings was gateway reviews. By the signs of it, at each of the gateway reviews, everything seemed to be hunky-dory. Is that correct?
That is correct.
This is extremely important. Chief Superintendent Macpherson said that the test results from Accenture showed 90 or 91 per cent success. Am I correct?
No, sorry. The 91 per cent figure referred to Accenture’s confidence. The figure for success was actually higher than that. The product test results were in the high 90s—it varied from module to module, but they were generally in the high 90s.
So, all the information that you got from Accenture about testing seemed to show that the system was working.
Yes, it did indeed.
And it was not until you actually went live that all the failures became apparent.
It was not on going live, Mr Stewart, sorry. I know that—
Well, the test pilot went live. You know where I am coming from. I am getting the language wrong now.
Basically, can you tell us whether the information that you were getting from Accenture about the success rate of the testing was factual? Do you think that they were actually achieving that success rate, or were they maybe being a bit disingenuous in what they were telling you?
Being careful about what I say here, I think that it has identified that the coverage of its tests was probably insufficient.
What does that actually mean, chief superintendent?
It probably needed to test more.
It probably needed to test more.
Yes.
Was its testing, then, somewhat selective?
I have no evidence for that. Watching what I say again, I think that its team has identified that it could have had better coverage in the tests across each of the modules.
I am not going to be gentle about this. Was Accenture covering up difficulties in the process?
I have no evidence to suggest whether there was a cover-up or whether there was just a lack of coverage in the process.
Basically, we are saying that Accenture’s testing was not enough to show what would really happen if that system went live.
The very fact that there were 76 defects of severity level 3 or 4 when the project was handed to us and that we very quickly had a number of defects of severity level 1 and 2 certainly tends to suggest that more testing was required during the Accenture phase.
Accenture has obviously failed to do this properly. Can it therefore be trusted?
I can probably comment on that. Since we identified the issue, Accenture has gone away and carried out its own internal assessment, much of which I cannot share with the committee. On the back of that internal assessment, however, it has come back to us with a remediation plan—which, to be fair, I should say has not been presented to Mr Leven, although I believe that that is in the diary or about to be put in the diary. That plan would address most of the things that I believe are wrong with the application. I have some confidence in Accenture’s remediation plan, but Mr Leven has still to see it. Of course, it all relies on that plan being done accurately, but it would definitely remediate the issues with the application.
I am a very simple man in these regards. If someone did not give me the full facts about the outcomes of a situation, perhaps because it had not been done right—and this testing seems to be an example of that—I would find it very difficult to do business with that person again. I say that without prejudice.
I want to finish on an extremely important issue. You said that Accenture delivered in Spain. Is the Spanish system less complex than the Scottish one with regard to what is required?
I will pass that question to Hamish Macpherson, but I will point out that his evidence has already gone through the nuances of the Scottish criminal justice system. The way we handle information once it comes into the system and the way in which it is then categorised and translated into other parts of our criminal justice network are significantly different from the approach taken in the Spanish system. We knew that it would never be a like-for-like, but it was the template that we operated from.
In some ways, Scotland should be very proud of ISCJIS, as it is probably one of the most integrated criminal justice systems dealing with all the documentation that flies around. Even down south, you could have paper documentation moving between prosecution services, whereas in Scotland it is delimited marked-up data that passes between them. The system has been in place for some time now and, as you will be aware, it is now in its second big iteration; however, that has introduced complexity into ICT systems that, to be honest with you, was not necessarily the case in Spain.
It is partly the complexity of ISCJIS that has caused these issues. Its power might allow us, for example, to get a standard prosecution report from the police to the procurator fiscal very quickly, but it also makes the ICT systems more complex.
Do you think that Accenture thought—I am asking for an opinion here—that it would get away with delivering something similar to the Spanish system, with the addition of the bells and whistles that were necessary to deal with ISCJIS?
I do not want to answer that, if that is okay.
I think that you have answered it by saying that you do not want to answer it.
I cannot comment on what Accenture thought that it could get away with.
I wonder whether Chief Superintendent Macpherson can comment on that.
It is hard for me to comment. All that I would say is that starting with an existing asset that does not necessarily fit into the police domain does not necessarily make that journey any easier. In some ways, starting with a blank piece of paper might have made the journey easier.
I will ask the ultimate question. In terms of the software used to develop i6, was the vast bulk of that code exported from the Spanish system?
No—the vast majority of it was bespoke software for Scotland, because of the issues.
But was the skeleton of it—
To be honest, I think that the vast majority of the software has been rewritten. Potentially, the original intention might have been to use the existing asset, but as it has become obvious how complex the system is, a lot of the software has been rewritten.
So the original intention might have changed when the complexity of what was required was recognised. Would that be fair to say?
Excuse me for a minute, Kevin. I am just trying to follow the flow of this. This is the story as I understand it. Police Scotland prepared a thorough specification that took in the complexities of the criminal justice system. It was a very rigorous procurement process. Accenture won the contract because of its track record in delivering another system, which it was not really possible to use, so it was necessary to start with a blank piece of paper. Therefore, I take it that it was Accenture’s expertise rather than its system that led you to choose it.
Accenture proceeded to do the gateway tests and it told you that the system was okay—I am just making things simple for myself. However, when you did your test run to activate the system, you found that it was anything but okay.
That is correct.
So either we have people who do not have the skill to do it in the first place or we have people who made mistakes, or a mixture of the two. What I cannot understand is the fact that Accenture did not know before you did the full test run that you would find out that stuff. I do not understand that, because if I was a developer who was developing such a system, I would have a wee run at it myself. You are saying that it was only when you pressed the button that all 81 serious and substantial defects surfaced. Have I understood that correctly?
I think that your summary is fairly accurate, but there is a bit that you missed out, which is the intense functional and technical design element that takes place after the contract is awarded. That would have painted an exact picture of what had to be built thereafter.
But that is pretty well the story.
In fairness, your summing up of events and Mr Stewart’s summing up are not far off—and probably a lot politer than—the questions that I asked Accenture at our September board meeting.
So Accenture would have done what you did when you pressed the button. It would have tested the whole system before you did.
I am sorry—I do not understand you.
When you did the test run to activate the system, all the mistakes came to light, despite the fact that you had been told that everything was all right following the gateway tests. You found that the system was really bad.
Yes.
Surely Accenture would have done that before the system left the factory, if I can use that metaphor.
We had no sight of the internal quality testing that it did. That is the way in which such projects work. We get the system handed over to us, we play with it and we find out how it is working. My honest opinion is that the senior managers in Accenture, who are the people we face up against, were probably not aware of the issues until we highlighted them to them.
Crumbs—I would not have them fix my pipes and radiators; they would be telling me that they were all right when I still had leaks.
It would seem that there has been a breach of trust. At the very least, people do not appear to have acted in good faith. I am not necessarily asking you to comment on that.
The report that we have mentions
“the findings of the Root Cause Analysis Assessment”,
which we are told
“were presented to Police Scotland on 12 November 2015.”
It says that the assessment
“was carried out by an independent Accenture team”.
Presumably that was an in-house team.
14:15
Indeed.
The team was independent in as much as it was unconnected with the initial project. Is that the level of independence?
That is correct. It was an Accenture team, but no one on it was involved in i6. I have had an update from the team of people who carried out the assessment. I have no doubt of their professionalism in carrying out the review. We got back an honest review. I am not in a position to share that because it was given to me without prejudice.
This may not be a good parallel, but in the update you told us that, on Storm Unity, Cambridgeshire Constabulary undertook
“an assurance review of our readiness”.
Was there anyone else representing the police interest that could have looked at this?
With regard to?
I think that I understand your question, Mr Finnie. We asked Cambridgeshire Constabulary, as an independent third party that has already delivered a product that we are trying to deliver, to have a look and tell us where we were going right and where we were going wrong. That was different from the independent root cause analysis test, which was an internal Accenture model. We cannot influence that—we cannot ask Accenture to get in its rivals to look at its own analysis.
No, no. I understand that. However, just as you used Cambridgeshire to look at the project, was there any opportunity to have anyone unconnected with the immediate team that had been involved for Police Scotland to look at it?
Are you suggesting that we should have got an external team to look at the supplier product that we were already very critical of at that stage?
Yes, indeed—to have a fresh look at it. I am not saying that to be critical of Police Scotland but, sometimes, you can be so immersed in something that—
The root cause analysis was sent to my team internally and presented to Martin Leven’s team. That was the opportunity for Martin’s team or Police Scotland’s technical people to have a look at that independently and coldly and ask the same questions that I did of the assessment and then, on the back of that, look at the remediation plan for how Accenture was going to fix the issues that were found. To be fair to the Accenture team, I found its review to be honest and independent.
Right. Will the remediation plan be published?
We are discussing the remediation plan. As it stands, the numbers do not work out—it is too long for me. Therefore, we are looking for a modular approach. That has been put on the table, and we will be looking at that over the next eight weeks. As a result, we will, I hope, end up with a plan that we can work to that would get out the modules to our officers as quickly as possible. That is the goal for me.
Obviously, the modules have to go out to your officers. What is the consequence for the training of staff and officers? People may already have been trained on some aspects of the system.
We had not commenced the training. We updated you on that last time we were at the sub-committee. After the hard-disk issue, we cancelled the training. It was shortly after that meeting that we identified the issues. If we go for a modular training plan, that will require modular training, obviously. Over the next few weeks, we will be working through the detail of that.
What happens if Accenture does not get it sorted? What are your contingency plans?
I have undertaken to the SPA to provide an options appraisal, which will include continuing the journey in various guises. There are two options in the paper—the one presented by Accenture and the one that we now have, which is the alternative roll-out plan. I have also looked at other contingencies, including having other frameworks or rolling out interim solutions across the country. We are doing that hand in hand with Martin Leven’s team.
To a certain extent, there must be lessons for the public sector. Every IT project that is commissioned by the public sector seems to run into trouble, whether it is to do with common agricultural policy payments or whatever.
If I can say anything to reassure you, it is that as identified in the Audit Scotland’s reviews and in the gateway reviews, the project—this has nothing to do with me; most of it pre-dates me—has been managed contractually very well from the onset. We have used external expertise and, as a result of that, we have a robust contract. That one thing that you definitely want to take from this is that the contract is robust. There has been no other cost for Police Scotland. I think that there will be no other cost for Police Scotland. That is an absolute positive to take.
Can we always engineer a position where we will never end up in a similar position? My suspicion is that such things happen not just in the public sector but all the time in the private sector—we just do not see it as overtly.
I think that it is a problem in the industry.
This goes back to the point that Mr Pearson made about the platform project and the lessons learned from it. We looked at every major public sector IT project. We looked at the high-profile failures that members will be very aware of, not just in Scotland but across the UK. At the heart of every single one of those public sector IT failures was a private sector company delivering it. That is why we went into the contract discussions very robustly, and we think that we have a watertight contract that protects Police Scotland and its interests.
I think that it is a problem in the industry.
Everything that we have discussed so far is very important. I want to come back to a quote from two years ago, when the chief constable said:
“It is worth emphasising that i6 will cover 80 per cent of the current operational police activity, so it is absolutely massive as far as the organisation is concerned.”—[Official Report, Justice Sub-Committee on Policing, 27 June 2013; c 144.]
Operational officers are waiting for i6 to be delivered. We have heard a lot about governance and all the rest of it, and I am satisfied that there are a lot of good checks and balances. When will we deliver to the officers the radical changes that are required? Elaine Murray mentioned options. If there is a plan for the journey ahead, when do you anticipate that we will deliver the system for officers so that they can do their work better?
You are right that i6 covers 80 per cent of the data flow in operational policing. Since Police Scotland launched, we have successfully launched 23 national policing systems that have significantly changed the day-to-day working environment for officers across the country. We are in the middle of a roll-out of a brand new desktop environment and a brand new network across the country, involving desktop collaboration. We have not just stopped because of i6. Although i6 covers 80 per cent of operational policing data, it is not the be-all and end-all and it is not the case that the police force cannot operate without it.
My concern about i6 being delayed, which is where Mr Pearson is probably going on this, is that, because of the delay, we have some legacy systems that are past the end of their scheduled life cycle. We are working through remedial plans to replace some of those systems in the very short term with systems that we already have in other parts of the country and that are more up to date.
I presume that those remedial plans to deal with the threat that you mentioned of systems failure will have a cost for the service. Will those additional costs be borne in mind when it comes to dealing with whatever the future is for i6?
Yes.
I think that I raised that when I asked about indirect costs, and whether things will have to be written off or moderated.
I am going to stop there because, as the witnesses will know, we cannot sit after 2.30 and we have other business. Unless you have anything further to say to the committee off your own bat, we will move on to our next item.
I would absolutely like the sub-committee to be aware that this has been a difficult session for us, as I am sure members can imagine. The focus of the sub-committee appears to be on i6, but the session was originally going to be about the progression of IT systems, and i6 is one of several hundred IT systems that currently operate in Police Scotland. I produced an update for members, which you all have.
We have read it.
It shows the tremendous success that we have had with the other IT systems. I came to the meeting in 2013 that Mr Pearson has referred to and discussed the blueprint with members. I can absolutely confirm that we have a tremendous track record of delivering the key commitments that we put into that blueprint.
We have launched 23 national systems and we retain our IT staff. We have done that against a challenging background of difficulty with retaining the skill set in the IT department. There is an inability to recruit in the public sector in Scotland because of the salaries that we attribute to some of the key skill sets.
I appreciate all that, but you will understand why the committee wishes to focus on what is a very large project that has had very large problems.
I absolutely understand.
The session has been helpful, but I am afraid that I will have to cut short your pitch on behalf of Police Scotland.
I will suspend for a minute to let the witnesses leave.
14:24 Meeting suspended.Previous
Attendance