Welcome to the Public Accounts Committee on Thursday 14 May 2026. For more than 20 years, the Government have tried to cut costs by sharing Departments’ digital platforms for back-office operations, such as finance, commerce and HR. While also reducing costs, such shared services can enhance decision making through standardised processes, as well as improving the consistency of data. In March 2021, the Cabinet Office announced its shared services strategy with the aim of introducing five cloud-based shared service centres. That strategy was allocated £1.15 billion in funding from HMT. The Committee last reported on the strategy in 2023, when we warned that the longer it takes for the Government to get on top of the situation, the greater the impact will be on the ongoing cost to the taxpayer. Today, we look to examine what progress the Cabinet Office has made with shared services since the Committee last reported on the strategy in 2023, and to understand what the Cabinet Office is doing to ensure that all Departments formally commit to joining their assigned shared services cluster. We will also explore how the Cabinet Office plans to align the shared services strategy with other digital transformation programmes, and what it has learnt from the failure of previous transformation programmes. We are very pleased to be joined by our team of witnesses. Jerome, I am going to ask you, please, to introduce yourself and tell us when you were appointed to your current role. Then we will go to Nathan, Dianne and Marcus: please, could each of you do the same?

I am Jerome Glass. I am the director general for future civil service in the Cabinet Office. I think I am about to end my third or fourth week in post now, so I am relatively new, but I was previously the chief operating officer in the Home Office, so an active player in one of the main clusters. I will hand over to Nathan.

Good morning. I am Nathan Moores, shared services strategy director in the Cabinet Office, and in post for five years.

Good morning. I am Dianne Jeans. I am the SRO for the Synergy programme. I have been in that post since January last year. Previously, I was the director of group corporate services transformation in DEFRA so, again, a player in the Synergy programme.

Good morning. I am Marcus Mason, director of public spending at the Treasury. I have been in post in this role for six months. Previous to that, I had a series of roles across public spending—growth-related roles—in the Treasury for the past 10 years.

You are all very welcome, but particularly Dianne and Marcus, because I believe it is your first time before the PAC, so a warm welcome. I hope that you find the experience interesting. To start our questioning, Rupert Lowe, please.

Thank you, Chair. It is rare for me to start, so thank you for picking me first. Good morning, everyone. I will start by reading a few lines from the Report—just a few little lines I picked out from the summary alone: “significant gaps remain”; “buy-in uncertain”; “no clear owner”; “Engagement…has been inconsistent”; “cannot enforce”; “fragmented governance”; “quality…varies significantly”; “functions…at a completely different place”; “limiting the usefulness”; “absence of a strong technical lead”; “remains some uncertainty and disagreement”; “inconsistent…configurations and data convergence”; “still some funding gaps”; “hindering…the planning for the value for money”; “Inadequate governance of independencies”; “unable to make key…decisions”; “timeliness at risk”; and “fundamental issues continue to jeopardise”. That is just from the summary. In paragraph 23, too: “Unless these gaps are addressed urgently, the government will not achieve the full benefits it set out to achieve and may incur increased costs.” All that is just from the summary of the Report, so my first question to you is: are you happy with progress?

Thank you, Mr Lowe, and good morning. The first thing to say is that this is clearly hard, and delivering this is difficult. The scale is huge, and wherever we look around the world, migration to the cloud in the private sector is difficult and problematic. The scale of this one is enormous. That said, I would emphasise the fact that there has been a huge amount of progress since this Committee last examined the programme. Even since then, we have established the clusters; we have three shared service centres already live; there has been an enormous injection of funding, £846 million in the last spending review; and we have overhauled the governance relatively recently, in the past six months, in response to the frequent scrutiny that we have had externally on it. There has been a ton of progress—all the procurements are now live and more than 7,000 people currently work in shared service centres. This is happening. This Report and this hearing are incredibly well timed because we are at a really important point where, for a lot of the clusters, the Departments are going to be moving into the shared service centres properly, and that is a point of difficulty and challenge. We need to look at the Report and reflect on all the progress we have made, but think about whether we are set up in the right way. I think we are, but that is something we need to constantly think about. I want to emphasise that this is absolutely foundational. I am director general for the future civil service. This is all about how we make the civil service fit for the 2030s. The shared services approach is fundamental because if we do not have the ability to use AI to bring processes together and bring some of the benefits that the new cloud-based systems will bring, we will be on the back foot. I recognise all the points in the summary. I would emphasise the huge amount of progress that has been made, but we are at a point where this is hard. I do not shy away from that.

So you are happy with progress. I think being efficient means doing things well, regardless of importance. And being effective means doing important things well. I read this Report, and all of the witnesses that were going to be here today—they have been changed for various reasons; Cat Little is not here, although I did not realise until just now—have built their careers on driving efficiency. Either they are not the right people to deliver the project, or the project is wrong and doomed to fail. It sounds good and it sounds great, like all civil service initiatives: let’s get shared services; let’s get the equivalent of private sector efficiencies. That sounds great and it’s a great strapline, but, again, I read about the inefficiencies: applicant tracking system (ATS) shelved; shared services portfolio incomplete; project ownership vague—I want to talk about that in a minute; risk management patchy; managing interdependencies concerning—I would say impossible. That is my view on that. How can you deliver efficiency across shared Government systems when the entire approach to the project itself is inefficient? I do not understand. I run lots of private sector businesses. You can give me the verbal baby food, Jerome, but at the end of the day I want to drill down and understand. This has been going on a hell of a long time. A lot of taxpayer money has been spent on this. Our job is to protect the taxpayer, not anything else. The taxpayer wants value for money, so I want an answer and I have not had one yet.

Two things. First, if I can just come back on your previous point that either the strategy is doomed to fail or we do not have the right people in place, I think the strategy is not doomed to fail and we do have the right people in place. I might let Nathan and Dianne speak, but these are people with deep expertise in delivering this sort of transformation. This is a huge cross-government operation. There are five clusters, five accounting officers and five SROs across all the clusters and right across Government. There are about 1,000 people working on this, all of whom are experienced in delivering this. Coming back to your question, this is already delivering efficiencies. That is the important thing. We are already seeing benefits from two of the clusters that have already moved into delivery. The potential benefits of this for the £900 million that we talked about for the spending review are of the order of £4 billion, or over £4 billion, so there is a huge prize to get. We are confident that we have set this project up in the right way. Obviously, as I said, it is quite hard and delivering this sort of change is difficult. That is why this Report is welcome. We have been scrutinised by NISTA, by the IPA, and by this Committee twice. We have reflected on that feedback because it is important to be confident we can deliver, but also be humble enough to listen to feedback. I do not know whether Nathan wants to come in.

I can add a little bit of colour to that. We recognise the progress we have made over the last five years. We have now got three service centres live and running, delivering services to civil servants and public servants every day. We have a fourth stood up in UK SBS supporting three Departments and one large ALB—that is UKRI. They are live and working. We know we have to deliver the three big ERP change programmes—we will come back to those. We recognise the Report that the NAO has delivered, and we value the recommendations in that. It calls out areas that we can work on further. Let me just explain a bit about what we have been doing over the last few years. Since we have had the IPA review and the NISTA review, we have developed an enterprise portfolio approach. We are managing about 18,000 data parcels across a portfolio of about 21 programmes. We are tracking dependencies through those. We have working groups at a portfolio level and a benefits level, so we have built that enterprise portfolio approach. We have started to work further on our governance with our shared services board and the technical design authority. That is the engine room where the SROs come together and work through the gritty delivery challenges, using the insights from the portfolio team and the portfolio working group. We have embedded a risk management framework that is Cabinet Office standard using RCoE tools. That is for the Cabinet Office, and we have amplified that to work with the five core clusters delivering the change and the functions. That is starting to tease out our key risks around delivery, data integrity and dependencies. We have also been working with live service centres to try to share best practice across the five service centres. We think we have made significant progress over the last five years. To support the clusters and live service delivery, we have done something that the functions are really proud of: we have created a suite of data standards. Those mean that we—

Nathan, I am hearing you. Thank you for that but, to Jerome’s point, the clusters don’t always agree with what you are telling us here. If you read the Report, the clusters contradicted what was said. The NAO uncovered that, as they usually do in their excellent Reports, so that is not entirely consistent. If you can’t even build and maintain a routine dashboard—which is in this Report, as you probably know—I fail to see how you can deliver resounding success and value for money on this project. If you can’t even deliver the dashboard, how can you tell?

I would question your assumption. I would say that the dashboard is one of the big points of progress that we have made, actually.

I don’t think the NAO would agree with you, from the Report.

That is a development since the last Report. The quality of the returns is now much better, so we are getting much better information and are able to use that to look at where the independencies are. On your point about how we know what is going on in Departments and how that impacts the overall programme, that is getting much better. The refreshed governance that we have put in place allows us to do that. We are learning and improving as we go. For something of this scale—which is, I would say, probably unique, but certainly unusual—that is probably the right approach.

Thank you for that. You were quite right when you said earlier that this is a time for major reflection; I think it is. You should look, as I did, at the case of the Office for National Statistics—I don’t know whether you know what happened with the Office for National Statistics—because I think they are a microcosm of what you are trying to achieve. They set out to do exactly what you think you are going to do—this top-down, centrally planned nirvana, which very seldom ever appears; it goes on forever, like the Forth Road bridge. The IDS was a multi-year project. The intention was to build an integrated data system, and £240 million was spent on it before the project was shelved. I will tell you what I think. All of you need to think seriously hard about this, in the interests of the taxpayer, because I can see a lot more money being wasted here. That is what I smell. I smell the tower of Babel. All you are going to find is what they found when they built the tower of Babel: the oxygen gets thinner, and God doesn’t live up there. You have not got the structure that the private sector would have. I heard talk in the Report of you wanting to emulate the private sector and bring everything together, which is what people do when they buy businesses and put them together, but it is quite clear to me that you have not got the structure to do it. More importantly, you have not got any one person who is in charge of this project. Nobody gets fired for not delivering. If you have a project in a private sector company, somebody is given the job, and if they don’t deliver to timelines, they get fired. Historically, you have not delivered to timelines. We are reviewing progress, and I can’t see any progress. You mentioned AI, Jerome. I think you have got to sit down with a cold towel on your heads and decide whether you are going to be overtaken by the digital revolution and if you are eventually going to do what the IDS did. Are you going to junk this when you get overtaken by the digital revolution, which I can see happening? I do not see that this is fit for purpose, and I think that you are going to waste billions more of taxpayers’ money on this—as I call it—dream. It sounds great, but you are not going to be able to deliver it. I am seriously concerned, sitting here representing the taxpayer, that you are going to fail, and I think you need to think about that seriously now. It is tragic, because I gather that a lot of the Departments are relying on legacy services before you deliver this.

There is quite a lot in that. Let me pick up a few points. The first thing to say is that, in designing this, we made the conscious choice that we would take a cluster approach.

You did not, because you were not there.

We—Government and Ministers—made a conscious choice to deliver this via five clusters, each of which have an AO, an SRO and so on. That is a choice that we made because we thought that was the most efficient and best way to deliver this.

But who is in overall charge? Who is going to get fired?

There is someone in overall charge of each of the clusters. My role is to be in charge of the overall strategy to make sure that those clusters are delivering, but each of the clusters—

But Jerome, you have been there a month—you just said so. You are in charge now, but this has been going on a long time, has it not?

Yes. That is precisely my point: each of the five clusters has an AO and an SRO who are responsible for each of their programmes. The way that we did that was a conscious choice. In the previous hearing, we discussed that there were alternative ways of doing this. We could have had one huge project for government; the decision we took was that that was much more likely to hit the sort of buffers that you were talking about. But this was a pragmatic way of getting the benefits of shared services and trying to do it in a way that delivered best for the Departments concerned. It is important to say that some Departments have already moved to cloud-based systems. My former Department, the Home Office, moved on to an Oracle cloud-based system—

But moving to cloud-based systems is not going to solve this problem, Jerome.

That is exactly what is going to solve the issue, because when we talk about AI, the best way to make sure that we are ready for it is to move to cloud-based systems. That then allows us to analyse the data better. We are transferring the responsibility for upkeep of the systems to the providers rather than having to do it ourselves and have in-house teams and so on. It is precisely because we want to be ready for AI that we are clear that this is the right approach.

I have used my allotted time. I can hear the Chair getting nervous that I am burning too much of it, but I would like to see you seriously reappraise whether you are going to fail here—now is the time to do it. If you are, you have to be honest with the taxpayer, because otherwise they are going to end up footing this bill. As you know, in a private company, it is people spending their own money. Here, it is people spending other people’s money. That usually leads to disaster, so you all need to sit down and reappraise the entire thing, because, I am sorry to say, I think you are going to fail. I think that you are going to make the same mistakes that have been made in the past. You will come back and tell us that progress has been made—a bit here and a bit there. If you plough on with this, I want to see somebody in overall charge. I do not want clusters or lots of different people who can all blame each other. I want one person, and when they fail, I want them fired. That is what we need to start having in government, and then we might get some action. That is my view.

Thank you, Rupert; we will reflect carefully on the questions and answers. This is quite a technical subject, so can we try to avoid acronyms? Here, we all know that “AO” is accounting officer and that “SRO” is senior responsible officer, but let us try to keep the technical bit clear so that we can get what we are talking about across to the public; otherwise, we are just talking to ourselves with acronyms. Rupert queried the dashboard system. The NAO are very used to designing and eventually running dashboards, because we did it during covid. Siân, the effectiveness of this dashboard is being called into question. Your report covers this in some detail, so can you give us a brief assessment?

Of course. Paragraph 9 in the summary and paragraphs 1.15 and 1.16 refer to the issue. We say that the Cabinet Office has developed a shared services dashboard, which gives the team and clusters quarterly information against a set of performance indicators, and it has other information around interdependencies too. However, the team cannot enforce or mandate timely and accurate returns, which means that they are not getting complete and consistent data from across the clusters. Our conclusion in the summary was that the quality and completeness of data provided by the clusters varies significantly, which limits the usefulness of the dashboard.

There is an immediate question there, Jerome. If we cannot even rely on the data going into the dashboard, how can we possibly rely on its results and recommendations?

I might ask Nathan to talk about this in a second. I would emphasise that this is an improving picture. We put a lot of effort into making sure that we are getting better data, and then using that data to take decisions and share with clusters what the programmes are delivering in the Cabinet Office and how those relate to the clusters. It is a conscious choice for us to be delivering this through clusters, each of which has an accounting officer, a senior responsible owner and so on. Our job in the Cabinet Office is to make sure that the overall strategy is delivering and that we are sharing information as well as possible; the delivery and the dashboard are an innovation that we have put in place to do that. We have been improving the data on that. I don’t know if Nathan wants to add anything on that front.

The dashboard is a brilliant tool that we have invested in. It is available 24/7, and it has all the cluster plans in as well as the plans we can get from the dependent programme. The tool does its job, and more. It is not a static production that happens once a month; you can go and view it. We have about 1,600 people who use it day in, day out to manage between the different programmes. That is just the tool, though. We have built a methodology around that, through the working groups around the portfolio, to start to surface insights around challenges, but I recognise the dependency point that has been raised. We can flag dependencies and work together through collaboration, but we do not have accountability for other allied projects and programmes, and that is pretty usual in an enterprise portfolio. Where we have to go further is in using those working groups and the tool that gives us the insights to be able to make decisions where we surface any tensions or differences in programme delivery.

I have an overall, encompassing question that I will leave to the end, which will cover some of this, but in the meantime I would like to get into some of the detail. Can you confirm that this dashboard is not publicly available and is for internal use only?

It is for use by people working across the portfolio of shared services, yes.

All this effort is going in. How do we, on behalf of Parliament and the public, know that it is producing some useful information?

It is a tool for us to use at an operational level to co-ordinate programme delivery and cross-programme delivery. When issues are flagged that cannot be resolved at a project-professional level or SRO level, they are escalated and feed into the governance we developed, which is two-tier governance, as recommended by IPA, NISTA and NIC, which involved setting up a service design authority—the engine house I talked about before—for resolving issues. If they cannot be resolved there, it moves into a shared services board that is chaired by Jerome and has the the lead DGs on it for the clusters. Ultimately, if it cannot be resolved at the shared services board, the issue would escalate up to the transformation board. That is how the governance based on decision making works.

Your reply plays straight into Rupert’s question: this is far too complicated, is it not? The whole structure is far too complicated. George Bernard Shaw said that genius is simplicity; any downfall can make it complicated, I say. Reading this report, the eyes boggle at how complicated it is. Why is it so complicated? Why, as Rupert says, is one person not in charge, with a clear organogram as to how the thing works?

By its nature, this is going to be complicated. We spend half a billion pounds a year on shared services. We are delivering services for 500,000 civil servants. Providing back-office services is, by its nature, an enormous operation right across Government, because we are talking 18 Departments and 300 arm’s length bodies. We had 286 legacy systems that we are going to be collapsing down into five. This is a vast transformation programme, and that is why I opened by saying, “This is hard,” but I am confident in the progress that we have made so far. There are ways of delivering this that would be appear to be simpler, which would be to have one shared service centre across Government. We made a conscious choice not to do that, because we wanted to make sure that it was going to deliver. Ministers’ view was that the best way to deliver that was via clusters. We brought Departments together by their existing software, or we brought Departments together who had worked together before. It may not appear to be simple, but we were trying to inject a pragmatism in the way that we develop those clusters, so that they could deliver. Now, each of those clusters has its own programme structures—they are their own programmes themselves. The role of the Cabinet Office is to make sure that the overall programme is delivering. The governance that we have put in place is essentially two boards. One of them is the Technical Design Authority and the other is the shared services board. The role of those is to bring the clusters that are delivering their programmes together to make sure that everyone is on track and to map the inter-dependencies. This is, by its nature, complicated—I am not going to shy away from that. But to Mr Lowe’s point, we have tried to design this in a way that is most likely to deliver benefits, by being pragmatic in the way that it is set up.

We need to move on and get into some of the detail, so I will ask Catherine McKinnell to come in with the next question.

I am very confident that we have looked both at the private sector and indeed at international examples. Nathan?

We have worked with other Governments around the world to try to understand their journey and to benchmark where we are compared with them, including countries such as Singapore, Australia, Canada, the Netherlands and Germany. We have linked in with Euritas, which is a big shared services community across European countries. We have also reached out to learn from the public sector, which has been leading the way in some areas, and docked in with private sector organisations to learn from their journeys. We have tried to understand our maturity relevant to their circumstances to shape the strategy and our approach to delivery. In addition, we have taken external advice to benchmark what we are doing, our maturity and whether we are pushing in the right direction. Everyone is advising that we aggregate. First, you must converge your data so that you can aggregate at scale, and that is why we invested in the NOVA data standards. We have those built into everyone’s business cases so that we all speak the same language and we can get the systems to do so. We have then moved beyond the data into looking at how you can scale, aggregate and go from the 286 systems down to the five. Others are telling us that it is also about maturity. You have to get your stable platform with your core systems and services working, and then you can start to move up the value chain into other areas—move into the middle office, so some of the work that maybe functions are able to do in commercial HR or finance. We are following others, not leading the way.

We have 92 in the first wave. Those ones are closest to core Departments and have more of a need to move to a central solution more rapidly, so we are giving them continuity of service. That leaves about 200 that are not in that initial wave. We should not underestimate the heavy lifting needed to bring those valuable arm’s length bodies on to those systems. We have been working with clusters—

The hard work to bring those arm’s length services into the core service centre.

Because we want to bring their HR, finance and commercial systems into that central service provision—and arm’s length bodies, by their very nature, have more unique service requirements.

Sorry; I was trying to give a full answer. It leaves about 200 that are not. We are working through some criteria, because we believe that some are so unique that they would not benefit from joining the shared service centre. Some will be so small that they will want local arrangements. We believe that there are about 50 to 70 in the middle, where there is value for money for them and for central Government in joining a cluster, and we plan for that to be in the next spending review.

No.

I can give an example about the Synergy ALBs if that would be helpful. In Synergy, we have 124 ALBs associated with the four Departments, so a significant proportion of those. Right from the start, we have ruled 38 of those out of scope in terms of those rules. They are public corporations or industry bodies; they are not appropriate to bring in. Out of that, another 55 are committees, authorities or advisory bodies that are just cost centres for their parent company. They will automatically come into our phase 1 delivery, and another 18 are included in the Synergy first phase. That means that they are already sitting on the same systems as their parent Departments, so they will be swept up into the first phase. That leaves 13, which are in the category that Nathan just referred to. They sit on separate systems, so will be more difficult to sweep up in the initial phase. We are working through how best to do that in line with the Cabinet Office’s timescale requirements.

It is an incredibly complex picture. The key test is, “What will bring best value for money for the Government overall?” Some of these arm’s length bodies are related to security services, so we have ruled them out of scope on that basis because of the security standards that they need. But another test might be that, if it is just a small body with a handful of people, it may not be in the best interest of Government overall.

Nathan may be able to say whether it is a shift in approach. I can tell you that, right now, we are going through all the arm’s length bodies to work out what the test is for whether they should be brought in in phase 2. We have talked about the first phase, where we have identified the 92; we are going through the remainder to see whether it will bring best value for money for the taxpayer, or whether it will just add complexity and stop them delivering in some way. We are putting those tests together, and then we will come out and say, “Right, these are definitely the arm’s length bodies that we think will be able to move or will be unable to move.” I am unable to speak to what the original intention was, but that is the intention from now.

It is a shift in approach. When we set the policy back in the day, we did it at a high level. As our understanding has grown, we think there is value for money in those that we have talked about. Some will not benefit, so we are targeting those where there is better value for money.

I think this is covered in the Report. It was part of the Matrix cluster. Obviously, with the shift, we are reviewing its membership of that migration. There is a question—again, this is referred to in the Report—as to whether there is a separate cluster that it could join in due course. Nathan may well want to speak to that. We have already covered arm’s length bodies, but the approach is a pragmatic one: rather than say that we are going to do everything and there is one rule, what we are trying to do is ask what is most likely to deliver best value for money for taxpayers. That is what has informed the cluster approach and the change of approach.

We have conducted some of the most extensive business cases scrutinised by the Treasury on this as part of the spending review. When this programme last came before the Committee, there was a question about whether this was going to be funded in the spending review. It has been funded to the tune of £846 million and we have robust business cases. The benefits that I was talking about—the £4.3 billion—are measured over a 15-year horizon. That sounds like a long time but, as I described, this is a huge programme and a huge shift—so 15 years is the right sort of timeframe to be measuring that.

I might ask Marcus to pick that up. I think it is to do with the distinction between the Treasury as a Department itself and the Treasury as funding the overall strategy.

Generally, the Treasury is very supportive of shared services because of all the points that Jerome made earlier: the big benefits over 15 years and the huge opportunity to maximise efficiencies and savings by cutting the type of duplication that can lead to increased costs. As Jerome said, it was funded at the last spending review with a huge amount of support from the finance function in driving functional convergence. As a Department, the Treasury is part of the Matrix cluster and is engaged and committed to the programme. Along with a small number of other Departments, the Treasury is already a cloud user: the decision making to onboard to Matrix is slightly different because of that. It is just a matter of timing whereby the Treasury as a Department requires some additional information about the benefits specific to cloud user Departments along with the service level provided for existing cloud user Departments, which will be provided by the programme in the coming months. At that point the accounting officer at the Treasury will make the decision on full participation. We are part of the cluster and are committed, but we require some additional information specific to us as a cloud user Department already.

To be clear, it is the strategy overall for 18 Departments and 200 or 300 arm’s length bodies. The key thing is, as I was saying, this is an incredibly complex landscape. Some Departments are already—

I know, but the reason I am emphasising it is because some Departments already have software in place that is of an equivalent standard. It is therefore a more nuanced decision for accounting officers there. The Home Office is another example that is brought out in the Report: it is already on a cloud-based system; it migrated in 2021—it is the Department I was the chief operating officer of—so the value-for-money judgment is more nuanced.

The strategy is that all is delivered through the clusters, so that those clusters are on the same systems. That is the first point. Secondly, my argument is that, yes, the plan is for everyone to be part of a shared service. The Home Office, even though it is already on a cloud-based system, is part of the Synergy cluster and is moving as part of that. My point was just that—going to Marcus’s point—it is a slightly more nuanced question for an accounting officer if you are already getting a lot of the benefits by being on a cloud-based system. If you are already getting that and you are moving into a cluster, it is just a more finely balanced accounting officer decision.

Your question was about the Treasury, and the answer we tried to give you explained—

Because, out of the 18 Departments, there are 17 others. As I say, the decision has not been taken for the Treasury, but there are 17 other Departments for which it is absolutely essential. And the overall strategy for the Government is to do this.

No, it has been deemed essential for all Departments. I am saying that, for the Treasury, it is just a slightly more nuanced decision for its accounting officer, as Marcus explained. But as I say, the overall strategy—

Just to clarify, I think you are right: it is a question of timing. As I say, the Treasury is committed and is engaged in the Matrix cluster. The Treasury system is more mature, and therefore it is a question of phasing. There are different waves. The second wave in the Matrix cluster involves Departments, such as the Treasury, that are already on these more mature cloud-based systems. Therefore, that decision making is being made as the actual Matrix programme is further matured, because there are just more complex interactions in that.

Apologies if you are getting that; I thought I had said the opposite, which is that the signal to Departments is very clear that the Government are moving to shared services.

No. Marcus has explained that better than I have, in terms of the judgment call for the Treasury, but it is clear for Departments. We are doing this; it is happening. The value-for-money case is one for the Government overall. It is £4 billion-worth of benefits, it is going to deliver better services for everyone, and it is going to deliver better services for citizens, for civil servants and so on. I do not think there is any question about the strategy, so apologies if my answers are giving you that impression.

Yes.

No, that is not the total cost. The £4.3 billion is the benefits that we have profiled through the business cases. There is funding through this SR to the tune of £846 million from the Treasury. Then the programmes will need to look for—

There are costs to be factored in separately to those.

I do not have it to hand. I can send it to the Committee with the detail of that with the business cases.

On that latter point, clearly you cannot claim £4 billion-worth of benefit without costing in the actual total cost, which you do not seem to know. We will want a note on that, and we ask you not to use that £4 billion-worth of benefits without fully knowing what the costs are, so that we get a straightforward, down-to-earth figure of what the real benefit is.

Perhaps we can write to explain the breakdown of costs and benefits. It is also worth explaining the costs of keeping up the 286 legacy systems if we do not move to the cloud-based systems. The cost of keeping them going—patching them up all the way through because they run out of licence or they break down or they are more likely to be subject to security risks—is something that we have factored into this.

I hear that. Jerome, you have just been put in charge of this, and you are in the hot seat today. Catherine was valiantly trying to get to the bottom of the Treasury’s role. To quote from paragraph 14 on page 9 of the NAO Report, “The absence of a firm commitment”—this is from the Treasury—“presents challenges for planning within the Matrix cluster, as the business case assumptions include participation from both DfE and HM Treasury.” If we cannot get a firm commitment from the Treasury, who doles out the money, it is pretty catastrophic for the whole programme, isn’t it?

I think this is precisely the conversation that Marcus and—

It is the conversation, but if we do not have a wholesale buy-in from the Treasury, it sends a pretty poor signal to the rest of the programme, doesn’t it?

Again, the Treasury has completely bought in and has committed to the tune of £846 million over the spending review period. That is them investing in the shared services strategy and saying, “This is the right answer for the Government, the 18 Departments and the up to 300 ALBs.” That is their commitment. What Marcus was describing is the timing question—he explained it better than I did—of when the Treasury joins the cluster. There is not any question, unless Marcus wants to come on this.

I am going to bring in either the CAG or Siân. The wording that I have just read out is very clear, as the business case assumptions include participation from both DFE and HM Treasury. CAG, Siân, were there question marks about the Treasury’s participation in this project?

As you say, it is clearly worded and it is agreed with the Department as well, so this is a factually accurate summary of the position.

There we are. It is agreed with the Department. There seems to be a question mark about whether the Treasury are fully signed up and fully participating, not only in their cluster but in the whole system, because they are the people who dole out the money. I don’t want a clever answer. I just want a basic, down-to-earth answer. Is it the case or is it not that the Treasury are fully signed up to the whole system?

I do not think it is a clever answer. The Treasury has invested in the shared services strategy—

They may have invested in it, but that is a different thing from them signing up to the project.

The Treasury as a Department is part of the Matrix cluster. It will have a decision about the go/no-go, which is based on additional information around benefits and service level for the Treasury as a cloud user, but we are heavily engaged with the cluster and we are working through all those components for us as a mature cloud user Department.

Marcus, I am going to stop you there, because I think we are getting the runaround. I think what I have just heard from you is that the Treasury are not yet fully signed up. Let’s have a straightforward answer: are they or are they not? I don’t want a lot of waffle. It is a simple question—yes or no? Are the Treasury fully signed up or not?

We are not yet signed up to the actual system that is relevant to cloud users. We are committed to the cluster. It is a matter of timing and sequencing, and us receiving that information in order for the accounting officer to make the decision. I want the Committee to rest assured that the Treasury’s commitment to this extends much beyond the funding that Jerome mentioned. The Government Finance Function is a central player in driving the functional convergence that we discussed earlier, and it is leading the way on that and supporting other functions across Government. We are committed across all areas of this programme. Hopefully that is clear for you, Sir Geoffrey.

About as clear as mud. Catherine, do you want to come back in on this?

May I come in at this point and explain a bit about Matrix, which may help the Committee?

I do not want to get too much into the detail on Matrix at this moment. I simply want to know, if the Treasury are not fully signed up, how we can expect every other part of the system to be signed up. No wonder it is not working if the Treasury are not fully signed up to it. If, as Marcus has made clear, the Treasury are not fully signed up to the cloud part of it, when are they going to be? Otherwise, we are all wasting our time. We might as well pack up and stop.

On that question specifically, we are expecting to receive the additional information for the Matrix cluster in order to allow us to participate and for the accounting officer to make that decision in summer this year. To your timing point, that is the period that we are working towards.

You said that it is for the accounting officer to make a decision, which implies that they have not committed. Otherwise, there is no decision to make, is there?

Sorry—I know I keep coming back to this—but we are talking a lot about the Treasury, but there are 17 other Departments, with hundreds of thousands of civil servants. The Treasury has invested in that, and they are part of the cluster. There is a judgment call for them about when they participate in which part of their cluster. That is for them as a Department and the one or two thousand—

It is the DFE as well, and one or two of the other Departments. All those Departments signed up to this at the beginning, didn’t they? Without reservations or conditions, they all signed up—all of them.

Including the Treasury.

That is my understanding, yes.

So why are there now questions about whether the accounting officer feels it is right to go ahead? Surely everyone agreed to it; all Departments signed up without conditions or reservations. Has anything significantly changed to allow any of those Departments to say, “Wait a minute—we’ve got conditions and reservations now.”

I would make a distinction between all the relevant accounting officers signing up to the strategy, which is absolutely the case, and the go/no-go point at which money is committed, when there is a second accounting officer judgment. It is absolutely correct that we are committed and engaged, but it is also correct that the final decision to move to this Matrix system from our already cloud-based system will be made off the back of the additional information that we will receive in the summer.

Shouldn’t that have been thought about at the beginning?

Others may want to come in, but there has always been the idea of having different waves and Departments sequenced in, which I hope you will agree is a pragmatic way of dealing with Departments at different stages and levels of maturity.

Yes, but the Treasury is not just any Department, is it? The Treasury is the Department that grants the money, so you would have thought that before it signed this off, it would have thought of all these parameters. I repeat that it sends a terrible signal to the rest of the system—I might bring David in here—that the Treasury are not yet fully signed up. The project was started in 2021, and five years on, the Treasury are still not fully signed up. Does that not send a terrible signal?

David might want to come in on this, but I would say that the signal from the Treasury that the programme and the rest of Whitehall receive is of huge commitment. The Government Finance Function is at the heart of driving forward the functional convergence agenda. The fact that the Treasury happens to be on a mature cloud-based system already makes the decision slightly more nuanced, and therefore it is operating to a slightly more delayed timing. But that does not undermine the overall commitment of the Treasury to this programme.

Thank you for your replies, Marcus. We have heard that loud and clear from both of you. We will digest what you have said, but—and this is not because of what you have said or because of the questions— I do not think we will get any further on this particular matter now. Unless you wish to come in, David.

If I may, picking up on a point that you made, the general system—not that related to shared services, but the general framework for managing public money—means that the Treasury allocates funding to Departments, but it effectively delegates to accounting officers to ensure that the money is spent in a value-for-money way. I do not see an inconsistency in the Treasury allocating funding but relying on accounting officers in individual organisations to then, as Marcus says, make the decisions at the point that you engage on whether the spend is value for money. There is a distinction between the strategy and where you are going, and the final decisions about whether it is the right time to spend the money, which are made by the accounting officer of the individual organisation.

The VFM decision is whether it is value for money for the Exchequer as a whole over the long term. Obviously, there will also be a short-term point, so that if we are locked into a contract for the next three years and cannot get out, of that, you would probably make the decision to join in three years rather than today, because it would be pointless to break the contract.

May I emphasise something that David just said about value for money for the Exchequer as a whole? That is a really important judgment. You have to take into account what will happen—what the impact on the cluster, and the overall impact on the taxpayer, will be when you are making your accounting officer judgment. That is part of what binds the clusters together. To go back to Mr Lowe’s point at the start about value for money for the taxpayer, that is right at the heart of the decision the accounting officer needs to make when they are making the sort of judgments Marcus was talking about.

My criticism is that if the Treasury were going to express reservations about any aspect of this, that should have been done much sooner, to give a positive decision for the rest of the system that the Treasury was on board. We will want to reflect that in our Report.

The DFE, too.

Unless you want to come back on that, David—and this is not because of what you have all said—it is the habit of the Committee to take a break in the middle of the hearing. I will let everybody take their breath and digest what they have heard. The clock is now at 11.06 am. We are under fairly tight time constraints, so if we can take a break of no more than four or five minutes, that would be hugely appreciated. Sitting suspended. On resuming—

To start off the next set of questions, Sarah Olney.

Mr Glass, there are obviously different levels of preparation to be onboarded to the shared services, but as of November 2025, the Government People Function, responsible for HR, is behind in all the clusters, excluding Overseas and Defence’s civilian HR capability. Why is it the Government People Function data so difficult to converge and standardise?

I have two things I want to say. I will answer your question directly, but first I want to re-emphasise that in terms of progress since the last hearing, the NOVA thing referred to in the Report—basically the standardisation of data and processes—is a big part of how we are going to save money for the taxpayer on this. It is fantastic progress, for the first part, but it is also absolutely fundamental. To answer your question directly, I acknowledge, as is in the Report, that the Government People Function is behind the finance function. The finance function, which is my background, has decades of bringing data together to produce accounts. That is much less mature in the HR space. In addition, in terms of the standards themselves, in finance we are trying to standardise 95 areas; that is 370-odd in HR. By its nature, it is much more complex. Also, HR rules and standards change with respect to legislation. All of that is not an excuse, but it does explain why it has been a little bit harder in HR to get that standardisation to the same level as finance. It is definitely something that we are focused on right now and have to improve.

Can you explain a little more why it is difficult? Is it that it is embedded within the processes and that it is hard to do, or is it just that there has not been the attention on it or the will to bring it together?

It is more about the quantity, as Jerome was saying. Also, in the HR space, there are more policy impacts, and the policy ownership often resides in Departments, so there is more negotiation around agreeing a standard across the whole of Government. What we have seen, though, is that clusters, like Synergy here, have been working to consolidate, where there have been any gaps across the clusters—we have cluster consolidation. As Jerome said, there is a team in GPG that is working on bringing those departmental and cluster professionals together to agree the remaining standards for HR. We have a bit more work to do. We are not as far forward as the GFF—the finance community—but the team are working on it.

Put simply, it is that finance has had to standardise in order to produce accounts, whole of Government accounts and so on. Departments have much more scope in the HR world with regard to HR policies; those are delegated to Departments, and they have much more scope to design their own policies and processes. So you are trying to drive standardisation in a system that has, by its very nature, been very heterogeneous, as opposed to the finance system, which is much more mature in terms of standardisation.

But is it correct to say that the people function decided not to implement a functional standard and that that is one of the reasons?

We are committed to implementing the functional standards, and there are people working on them. I just think that it has been more work, and it has taken us longer to do it than other functions.

But it is not a resistance, and it is not that it is impossible?

As the person in charge of the function, it is absolutely not a resistance or a sense of, “We don’t want to do this in principle.” It is just that point about complexity; it is just that much harder to do it. We are making good progress, but it is absolutely something that I want to focus on.

What are the remaining obstacles to progress?

Nathan might want to update you on how much progress we have made recently. We are just working through it, basically. I don’t think there is an in-principle obstacle, other than the one that Nathan talked about, which is that legislation changes and Departments change policy as they change.

That is something that is happening in the future—

All the time.

From now on, you will have future policy changes. Can there not be a standardised implementation?

That is exactly what we are trying to do, but it is a slightly moving target. Nathan, is there anything that you wanted to add on obstacles?

We have the critical foundations—the standards—to go into the build for the three big systems that are being implemented. That is the first step. The second step is that we know we have to go further on the data standards, and that work is in train. Then there is the investment that we all need to make across the different functions, and the continuous improvement of the data standards. What we have learned from when Government has tried this before is that when we create data standards, they get put on the shelf and their quality erodes, and people diverge. What is brilliant about what GFF have led is that we have created the NOVA tool, which is a repository for all our data standards and process standards, with a glossary of terms and, really importantly, the key performance metrics for managing the function and service delivery.

One of the things that we are doing right now is spending, I think, about £5 million or so on bringing the processes together—standardising the processes. The way that one Department does recruitment follows one path, the way that another Department does recruitment follows another path, and that in itself brings inefficiency, so we are trying to bring those processes together—standardise the processes. On the question about AI, the other thing that these cloud-based systems allow us to do is to look across all these processes and say, “Okay, this one is much more efficient than that one. Therefore, why don’t you standardise according to this process?” It may sound dull and esoteric, but by its nature mapping out how these processes work and trying to spot the inefficiency is part of how we are trying to get value for the taxpayer.

I understand that. I just want to highlight paragraph 2.14 in the Report, which says that four of the functions—finance, HR, grants and commercial—consolidate their existing standards into the NOVA system, which you were talking about. However, it says: “HR participated in the design process but did not work towards implementing its revised standard.” Can we come back to that point? At that point, having participated in the design process, why wasn’t more work done to implement that standard, or why does the Report say that?

I am not the functional lead for the Government People Group or the HR function. From what I can see from my perspective, around the time of this Report there has been renewed interest in implementing those standards, and a renewed team focusing on that.

Okay. So the Report’s findings that it wasn’t doing that are now being addressed?

They are helping to get action.

Sorry, can we just stop there? That is a fairly shocking indictment, isn’t it? I quote from the paragraph that Sarah has picked up: “HR participated in the design process but did not work towards implementing its revised standard.” She quoted that sentence. Jerome, I believe that you chair the shared services board. Why wasn’t that immediately brought up to your board, and why didn’t your board immediately take action on it?

Obviously, this is not an excuse, but I have been in post for a small number of weeks. To go back to where I started right at the start of this hearing, we are at a critical point. This Report has been really helpful in flagging up some of the issues that we have identified. It is absolutely something that I am focused on and want to get to the bottom of. My understanding, as Nathan was explaining, is that we are now making much better progress than we were when the fieldwork for this Report was being done. This is an area that I might also update the Committee on because, as you say, it is fundamental. This is where there are going to be a lot of savings for the taxpayer, so it is something that I am focused on right now.

Can you tell us a little more about what exactly you are doing to implement the functional standard in the people function?

The Government People Group has a dedicated team working with clusters and Departments. The functional standards are adopted and are being used in the clusters. What we have to do is embed those with Departments. We have a recently appointed senior leader across GPG to work with that team to embed them.

What specifically needs to happen in order for that to be achieved?

That is a convening across the functions, bringing together Departments to agree one way of doing a specific process or data element. It is about reaching consensus and agreement—using best practice to agree on that—and then that is published as a standard that we all use, so that we can share the data and can all understand what the data element does.

What sort of timescales are we talking about here? How long do you think it will take before everyone is signed up to and using the new standards?

We have the core elements to be able to build the new systems and services, but as I was hinting at before, this is a continual process, so rather than a once and done—

I understand about future legislation, policy change and so on, but presumably a baseline is needed, at which point you say, “Right, here we all are together, and we are going to move forward as one.” Let us separate out the fact that this is obviously not a fixed state, but at some point you will have to get to a point where you have agreed that all future changes are going to be adopted collectively, at the same time.

Yes, and that moment has already passed for the core essentials, because we built them into the requirements when we purchased the new systems—the big three systems, the big three ERPs—and that will then be baked into the early service design.

How long will it take before everyone is using the same shared standard?

Unless Nathan happens to know the answer, it might be helpful if I come back to you on that as part of my update on where we are. It might be helpful for me to set out where we are in HR functional standards and where we think we are going to get to. The point is that this is a continuous process, and as new things come on board we are going to be continuing to iterate. It might be helpful for us to set out exactly where we are now and where we are going to get to by, say, the point where we are onboarding some of the major Departments by the back end of this year. Would that be helpful?

I think so, yes.

Thank you.

I have a final question, Mr Glass. How are you going to ensure that your function is not going to hold up the implementation of the overall shared services project any further?

I do not believe that that is holding up the implementation of the overall programme. This relates to maximising some of the benefits to the taxpayer as we go through, but there is total commitment from HR directors right across all the clusters. Again, I would emphasise that these clusters are chaired by senior responsible owners within Departments, so we are all working together to get this set up. The HR community has completely bought into this, can see the benefits and actually really wants this to happen. It can see from the other Departments that have done it, such as the Treasury or the Home Office, how it can make their lives easier and how we can get better data, so there is no resistance culturally. But as I say, for some of the reasons that we have talked about—all the different processes and policies within Departments—it is just a bit trickier to get there. Sorry, in terms of what I am seeing, it is not resistance; what I am seeing is more work that needs to be done.

To turn to the governance and management of the different interdependencies, paragraph 18 of the NAO Report states: “Inadequate governance of interdependencies is still hindering clusters’ planning for the value for money and delivery of shared services”. I would like to understand how you are managing your five interdependent programmes to deliver the shared services effectively. What are you doing to address the governance of those interdependencies? At the beginning, you mentioned that you were laser-focused on the governance, so what are you doing to make sure that those are actually working properly?

There are two layers of governance. The first is the clusters themselves. I know that is not what the paragraph refers to, but I re-emphasise that the clusters themselves have huge, extensive programme architecture in place, ultimately reporting to the accounting officer responsible for the overall cluster who holds the money for that implementation. There are governance and programme boards for each cluster. On top of that, bringing the clusters together—again remembering that most of the responsibility for delivering sits with the cluster—we revised our governance in, I think, October 2025 so that there are two layers, one of which is a more technical group in which we have—my apologies; the acronym escapes me right now—

The National Infrastructure and Service Transformation Authority.

Thank you, Chair. NISTA is on that technical group to make sure the strategy is delivered. On top of that, there is the shared services board, which I chair and which has the DG and SRO levels across each of the clusters. So the governance sits within each cluster. A technical group meets every month and then a quarterly board on top. That is a new governance that we have put in place.

If there is a problem, what is the route to a decision?

It is escalated. Monthly, clusters can say, “We are worried about this.” We use the dashboard in that meeting and we share what we see. They then challenge as part of that group. It then gets escalated quarterly. On top of that, ultimately it goes up to the accounting officers, the heads of Department across Government. That is the level of commitment that we have: it goes all the way up to the top of Departments. That is how things get escalated. There are different groups beneath that that work on technical issues, but that is the foundational governance.

Figure 7 on page 31 of the Report shows a host of other Government digital change programmes that also have interdependencies. I am keen to understand what you are doing to ensure that those programmes are compatible with what you are trying to achieve here, because there are so many of them, and the impact of 14 of them, according to the Report, is unknown. How are you addressing that challenge?

This is another area where we have made progress. In terms of learning, I do not think this is an area where we could have done better in the past. The Report refers to a particular applicant tracking system, where I think we could have learned better. We are now tracking, I believe, 25 programmes that central Government is rolling out that could have—our assessment, based on the dashboards, is that they do have—an impact on clusters, either all of the clusters or particular clusters. We use the governance that I have described—the Service and Technical Design Authority and the shared services board—to talk through where those programmes are at. We share information about where the programmes are at and the Departments and the clusters themselves question us on that. That is how we track all of these. You may well come on to this, but I think that is something we have not been good enough at, so we are now trying to get much better. I think at the root of the point on the applicant tracking system referred to in the Report was the need for better sharing of information between the Cabinet Office and the clusters.

If you are sharing more data, where does the ultimate responsibility lie for making sure it is unlocked and a solution is found? It sounds quite complicated to me. If you are in a specific arm’s-length body or a Government Department and you have identified a problem with the interdependency with another change programme, where does that decision lie? How do you unlock that? If the applicant tracking system was presented to you again, how would you unlock that?

We refreshed our governance around the same time that we reset the applicant tracking system programme. Nathan will know more than I do, but my understanding is that we were not confident that the structures we had in place were spotting things well enough. The direct answer to your question is that it is in one of those levels of governance—either the Service and Technical Design Authority or the shared services board. That is where decisions are taken across the clusters, as opposed to within the clusters.

The two-tier governance has been refreshed. We have embedded it. It is actually under review at the moment to make sure that it is still fit for purpose. We recognise what is in the NAO Report and we are tracking these programmes. To your question exactly, we are getting much more information on some of them. That is not because people are not willing to share; it is just that in some projects, they are at an early stage of maturity, in that scoping phase. We have some gaps where we need to work harder with those teams to make sure that we can see their project plans, see their dependencies and map the dependencies. Then, if there are conflicts or tensions, we can use the governance that Jerome has walked through to try to resolve those. The ATS was worked through our governance, but also through a separate governance; an ATS advisory board was set up to support the discussions around that programme.

Are you confident that, if the equivalent of the problems with the ATS presented again, it would be caught sooner and addressed faster?

I am confident that we have the right governance in place. I do not want to give you an assurance that it is perfect. Apologies, as I know that I have said a lot of times that it is complex, but this stuff is hard. The governance is much better. One of the reflections that I might take out of the Report is that—to the point that Mr Lowe has been emphasising—as we enter the most difficult part of this, which is when most of the clusters begin to onboard, we need to think about whether those structures are robust enough. I think that they are, but the challenge is there and we will think about that. I am confident that it is much better and that we are sharing much better information between the clusters. It is important for us to look at whether it is resistant too.

If you identify an issue, who ultimately has the lever to mandate a decision?

That conversation happens at the shared services board. The Cabinet Office is there and the cluster heads are there. Ultimately, this goes up to the heads of Department, so it is my boss, Catherine Little, who is the chief operating officer for the civil service. It is ultimately a ministerial decision overall. We would hope that the shared services board should be resolving and able to take those sorts of decisions, but if it cannot, it goes up to the accounting officers, who are the heads of Department—permanent secretaries. That is where the decision gets taken.

I have been listening to this exchange very carefully. We all know that if the governance of a project is not right, it is not likely to succeed. Let me take you to paragraph 3.5 on page 32—you have referred to this, Jerome. The Report says, “In January 2025, the National Infrastructure And Service Transformation Authority”—this is the governance body that is embedded in the Treasury, used to running big projects and has a lot of expertise—“urgently recommended the creation of a central design authority governing both shared services and ‘the wider functional and interoperability agenda’. In July 2025, SSfG—that is the lower board—responded so that recommendation by “applying for funding to resource a new Government Corporate Services Board”. What happened? “The SSfG’s request was unsuccessful.” They were told to go away and form a technical design authority. As a result, one cluster “expressed concerns that the SaTDA”—what an acronym, for God’s sake—“could not ensure alignment between shared services and functions sponsoring other change programmes.” I do not know how the two of you can sit here and say that your governance is correct. You reset it in October 2025. You, Jerome, said it is now fine. Nathan said it is being reviewed again. Have we got to a settled state with the governance of this project or have we not?

Apologies if I said it is now fine; I think I said that I am more confident that it is addressing some of the problems, but that it is important that we keep reviewing it and making sure that it is right, given that we are entering the most difficult phase. That is the point I was trying to make. I would never say that it is perfect. It is important that we reflect on this Report and think about it as we enter this difficult stage. The governance has been refreshed, following the NISTA recommendations, so that we have that shared services board. It is not as if we have been sitting there admiring the problem; we have taken action. With any programme of this size, it is important that we do not sit there, put our fingers in our ears and say it is perfect; it is important that we listen, reflect and think, “Actually, is it right?” That is the point that Nathan was referring to: have we got this right, given the stage that we are about to enter, which is the most difficult stage of these transformation programmes? We are delivering, but we are about to enter that difficult stage, and it is right that we reflect on that and make sure that it is right. It is better, but is it perfect? That is the question we want to answer.

You have not really answered the thrust of my question. NISTA, which is the Government body that understands how these things should be done, urgently recommended that you set up this central authority. Your sub-board did the right thing by applying for funding. I suspect that they would not have applied for it unless they had applied to your board, the SSB. They presumably got cover, as they were applying for funding assuming that the request would be granted. They wanted to set up this Government corporate services board, which sounds quite powerful and useful, but instead it was watered down to the creation of a service and design authority. As I said, that has been proven in that paragraph by one cluster expressing concerns that the sub-board could not ensure alignment between shared services. Nobody can seem to make up their mind about what the governance of this project should be.

May I come in here, Chair? The governance that we have in place is fit for purpose. You are right to question why we are having a review, but after six months of running the current governance, we thought it was useful to check up with the members on the terms of reference and membership to evolve it and make sure that it is fit for purpose. On your point about the Government corporate services board, there was an ambition to go beyond the current scope of the programme—the strategy we are talking about today—and the governance to see if it was beneficial to look at wider corporate services transformation. Our governance to deliver our remit is fit for purpose. There are other programmes, projects and portfolios like us; there may be opportunities for the Government to try to bring our collective efforts together into such a board that may have advantages for the Government, but our governance works.

Jerome, you told us that the Cabinet Office permanent secretary sits on top of this entire programme. You chair the board, not her. Does she attend the board? What is the reporting mechanism between the board and the permanent secretary?

I am responsible to her for delivery of the overall strategy, and she is responsible for delivering the Government’s overall strategy. She delegates that to me, so I chair the board to ensure that the overall strategy is delivered, bearing in mind again that the clusters are responsible for delivering their programmes.

I think you said earlier that at least five permanent secretaries were involved. Do they sit on the board?

They do not sit on a shared services board. Each cluster has an accounting officer overall, who is responsible for their cluster and for running their programme. As is relatively common, the programme is delivered by a senior responsible owner, and the senior responsible owners of the programmes are on the board. That strikes me as relatively sensible in terms of architecture: the programmes are delivered by the senior responsible owners, delegated from the accounting officers, and they are part of a board to bring those together.

I get the structure. What is the mechanism for the five accounting officers to make sure that their Departments fully participate in all the clusters and sub-bits of those clusters?

That is the transformation board, chaired by Cat Little.

Who sits on the transformation board?

The accounting officers for the five lead Departments for the clusters.

How often does that meet?

I would need to take that away and check. I am not a regular attendee, but I last gave an update for shared services pre-Christmas.

I think we have got as far as we can on governance for the time being.

This question is probably for Dianne Jeans to answer to begin with. You seem to be saying, “We have all these problems, but it’s all going to be all right in your cluster, isn’t it, because we got Capita to come in and sort it out?” Do you understand the amount of scepticism there is, given Capita’s performance on the civil service pension scheme? They just had the Royal Mail pension scheme taken away from them; I think the teachers’ pension scheme is subject to litigation too. It has been an unhappy experience at every single stage with Capita, but you have awarded this contract without any reference to that, apparently.

As you say, Capita has just been awarded a contract. That is a very different scenario than from pensions, but I understand the concerns about the pension scheme. The award decision followed a 12-month procurement process which commenced in mid-2024 and finished in the middle of last year. It followed all the Government regulations and processes. We also had strong legal and commercial oversight and subject matter experts from all four Departments assessing the competing bids throughout the whole process. Capita emerged as the clear preferred bidder under Government procurement processes. We are working very proactively and positively with them to mobilise in a different way. The important thing is that this situation is a very different landscape from that of pensions. For example, the technology for payroll delivery is not being delivered by Capita—it is being delivered by an IBM-Deloitte-Oracle consortium. It will be based on an Oracle platform, which is a very well-known HR and finance system. Capita will use that system, but it is not designing or delivering it. We have a much different relationship between our incumbent supplier SSCL. We have a very clear understanding of all the work that is going through current SSCL contracts. We also have contractual obligations between the two for a smooth transition. A very positive working relationship is emerging so far. An important point is that there is no hard stop in Synergy. We will not go live with a system unless it is absolutely tested and stable. We will have a very strong focus on payroll testing to make sure that it is operating stably before we go live with it. We are very conscious of the issues that have happened in the pension space and will be very focused on continuity of service. We are absolutely staying close to the Cabinet Office and learning all the lessons that they have from that pensions experience.

We had lots of assurances about the pension scheme before it went live—we were told that everyone had talked to everyone the right way and got agreements and assurances. They were not worth the paper that they were on. They failed. They gave assurances; they gave promises; they gave commitments, and then they were not delivered on. I know you are saying that it is a different part of Capita to the pensions part, but it is still the same company, the same culture and the same ethos, isn’t it?

My clear response is that they are not delivering the system. I understand there have been issues with the pensions.

Why do you need them then?

They are our business process service. They will be doing the transactional services—processing invoices and information.

Why can’t that be done in-house?

It has not been done in-house for 10 years. We are currently with an existing outsourced provider. It has been that way for 10 years. We have harvested all the benefits and savings from working with that provider, and we are building on that and continuing that process.

Why can’t it be done in-house?

That is not the strategy that the programme has adopted.

Why can’t it be done in-house?

It is not the strategy. My job as SRO is to deliver on the Government’s strategy and on the business case that has been agreed and endorsed by the Government and Treasury.

Was it considered?

Different clusters have taken different approaches. Some are delivering things in-house. The HMRC-led cluster, which is known as Unity, is doing some of its business process services in-house. The Matrix one, which is led by DSIT, is also doing it in-house. Again, that was a conscious decision based on what is the best value for money for the taxpayer. Also, pragmatically, as Dianne said, the Synergy cluster, which is led by the Department for Work and Pensions with the Home Office and the Ministry of Justice in it, is the most mature in terms of working together as a cluster. For a number of years, that has outsourced to a provider called SSCL. Given that we were outsourcing, the decision taken by the clusters on what is the best value for money was to carry on outsourcing, and that is the process that Dianne has taken forward. It is not that things cannot be done in-house. That is the choice that the accounting officers have taken when setting up this programme in terms of what is the best value for money, because that set of Departments has already been outsourcing those services.

Was an assessment done on whether it was best value to outsource on this occasion? Was there any assessment done against the possibility of bringing it in-house?

The decision on the direction was taken in 2021 very early on in the process and in a very early part of the strategic business case. It was assessed as the best value for money in terms of building on the existing benefits, services and savings that the four Departments had already gained from outsourcing. The decision was also taken in the context of the Government policy at the time to reduce the civil service headcount. A combination of the best value for money and Government policy meant that was the decision taken at the time and it has been pursued by the programme since then.

I was not there, but I assume that it was one of the options examined. We can confirm with you, but I would be very surprised if there was not a judgment made about whether to outsource or insource it. I will write to you to explain that.

We want to know whether that was an assessment about the value for money of that aspect of service, or whether the overall policy was to keep the civil service headcount down, so it was not examined seriously.

It was a combination of both. The savings had already been taken out of an outsourced provider and to reverse that would mean bringing a lot of costs back in house.

That is assuming that in-house provision would have been more expensive.

When you outsource and you have saved a lot of money, the assumption is that if you bring it back in, you will reverse a number of those savings.

Sorry, that seems a belief rather than a systematic approach to assessing what the actual costs are.

I take your point and, as Jerome said, we can provide that information.

May we have a note from one of you on what the assessment was between the in-house and the outsourced version? What was the actual business case that was produced?

Jerome just indicated that a different approach has been taken with different conclusions on, I think, HMRC. Could we have a note to show why they thought it was better to bring it in-house? I cannot understand why the two are so different.

We certainly can do that. We designed this to be as pragmatic as possible. For example, in Defence, in HMRC, and in what was the Department for Business and the Departments that came out of that, they were already provided by in-house providers, so it was a pragmatic decision to carry on with that. When the assessment was done, that was the right decision for those Departments. The Synergy Departments were in a different place; they did not have their own in-house providers because for a number of years they had been part of a cluster that was outsourced. Throughout, the way that the cluster has been designed is about pragmatism in terms of what software we use and whether it is insourced or outsourced, in order to maximise value for money. That has been the watchword, but I commit to sending you a note on the judgment.

Doesn’t the Cabinet Office oversee this, and haven’t they got a sense of direction about how these things should be approached?

It is for the accounting officer responsible for the clusters to take the decision about what is the best value for money. The Cabinet Office interest is about making sure that, overall, we keep the cluster strategy and that all the Departments are migrating at the right point.

Surely, the Cabinet Office’s role in overseeing this is to say, “Hang on a minute. They are doing this quite well here in-house. Why don’t you have a look at it over there?” These are not separate and completely independent of each other.

Chair, I think that note needs to look at both cost and operational benefits, because the two are entirely separate.

This has been a pretty torrid session, even by PAC standards. It is because, as you can see, the Committee is pretty sceptical about whether, how and when this whole project is going to work. I think it is partly because it is far too complicated and nobody is really in charge—my questions on governance prove that. Can I get a commitment from you that you and the transformation board will look at every aspect of the project to satisfy yourselves that it is going to work? As Rupert said at the beginning, there is no point carrying on with it and wasting more taxpayers' money if it is not going to work. Shared services should be a pretty simple concept. If you make it too complicated, of course there is going to be trouble. Can we have a commitment that your board, SSB, and the transformation board will have a complete look at whether this thing is going to work? When it has done that, could it then produce for us a road map, with timetables, dates and costs?

And responsibilities.

We know the responsibilities, but that could go in there. We need to see very clearly your intentions and the transformation board’s intentions for how to make the whole project work. Witnesses indicated assent.

At the moment, you can see that the Committee is pretty sceptical. We are also pretty sceptical about the £4 billion of benefit figure, which we have never seen before. The Report does not mention it. In fact, the Report mentions a figure of £1 billion, so we are on different planets here, it seems to me. We need to be very careful about what benefits—net benefits, after all costs—we are expecting. If you could go away and do all that, and if we could have you back when you have done it to query what you have done, the Committee will have a lot more assurance that it is all going to work. In the meanwhile, I thank you very much for coming. I am sorry that this has been a bit of a difficult session. I know that you have just been parachuted into the job. It is now perhaps incumbent on you to make it work, and therefore you will have an incentive to make it work and to produce the information that we require. An uncorrected version of the transcript will be available in the coming days. Following that, we will produce a Report with recommendations, which we hope you will study carefully. In the meanwhile, as I say, many thanks for coming to give us your views this morning.