Good morning. Great to, -- it's a really good video, isn't it? -- Just a good couple of minutes. Just seeing it yesterday for the first time. I thought, wow, that's good. I'm Adolfo Hernandez. For those of you who haven't had the opportunity to meet, good morning to everybody, everybody in the room, -- full room, and those of you dialing away, from remotely, from home and your offices. We're going to spend the next 40 minutes, Pablo and I, our CFO, covering the highlights of the first half of the year. We're going to do it, as always, we're going to do a very brief introduction. Then I'm going to hand over to Pablo. Pablo is going to go through the numbers, and then I'll get back, and I'll take you through strategy, executions, and plans going forward.
Just by way of recap, just over a year ago, we did a Capital Markets Day in June last year, where we presented our new strategy, a strategy that wasn't predicated on changing what Capita did. We were very happy, and we were hearing from our customers what we do really matters, matters every day, matters to millions of citizens and consumers all over Western Europe. What we do is really good. It was the how we did it that we needed to work on. We set out on a really simple strategy, which was we're going to build a better Capita. That better Capita was going to be predicated on making changes in four key areas: better technologies, better efficiencies, better delivery, and building a better company. I'm really pleased to see that as we have gone executing, we are now getting very granular into the execution.
The strategy remains the same. Thirteen months after we unveil it, we continue to work on the better and the four betters. Underneath, we have been able to map the six areas of where we have to work to truly get there. This has been a lot of work that we've been doing over the last six to nine months related to execution, pure sheer execution. Underneath that, we have come out and worked with the company to come up and create the new culture, a new high-performance culture, underpinned by new values that are necessary to create a different way of working to build that better Capita. What is really particular and important about these values is that they were not created top-down.
This is not the CEO or the Executive Team or the Board waking up one day and saying, "These are going to be the new values." This has been a grassroots exercise that we've done, asking all of our colleagues to participate and actually getting input from over 10,000 people to come and describe what are the values that we need to be successful in the near future. We are connecting the strategy of building that better Capita with better financial returns, delivering better outcomes for our customers into the execution priorities of the four betters. We've got the actions and everything powering through our values. We feel now that we've got a truly connected ecosystem to start gathering speed as we get into this inflection point of growth into next year. What about the first half?
I always say there's never a boring day in Capita, and there hasn't been a boring day in the first half. I think it's been brutal execution. We're happy with the solid results we got out of the first half. If you look across the four betters, as you read through the release we sent out this morning, we've seen significant areas of improvement across all four. I'm just calling out a few here across each of them because I think they're important. On the technology front, besides saying a year ago that this was going to be important, that working with the hyperscalers was going to be critical, we've managed to translate that into action now. Launching earlier this year, the capital, -- the AI Catalyst Labs, which I'm going to talk about a little bit later, and getting some real output out of this.
Over 212 use cases for AI, some agents live, to see some real examples already at work is quite, quite important. Looking to the future, the over EUR 4.4 billion in tech-related pipeline that we've been able to build, a pipeline that we feel we're in a very, very well-positioned way to go and address. If I look at the efficiencies, a year and a half ago, we used to talk about EUR 40 million to EUR 60 million. When I joined in March, -- and then six weeks later in March, I said, "No, we're going to do EUR 160 million." Then we said, "We're going to do EUR 250 million." As of today, we're over EUR 205 million. I think we have executed really well. We built a cadence of roughly EUR 10 million per month of efficiencies that this team is working really, really hard, finding really painful and hard ways to go and deliver.
We understand that is the only way forward. We're really pleased with where we are. Most importantly, as we're doing this, the thing that I'm the most proud about is that we are creating a culture. Now we are thinking about efficiencies in ways we were not thinking about before. We have created an operating system to stop the cost to creep back in unless it's part of our reinvested policies. If it's part of reinvesting, it has to go through a very thorough business case and milestone systems. On the delivery, I think last time I was standing here, I talked about customer satisfaction going up, even though we were going through a very difficult transformation. I remember six months ago saying to you, that is the foundation. In my experience, when you start delivering well for your customers, that's the first step to start getting to growth.
If you go through the release today, you see a number of leading indicators. You see that growth on the TCV around the group, which is great. If you look at some of our businesses, they're growing faster than the group. Pensions is an 88% growth. If you look at our public service, it's growing TCV north of 50%. We're starting to see that we're seeing good win rates. We're seeing improvement on the book-to-bill ratio. We're starting to see real momentum on the delivery. On the company side, on the employees, I said to you last year, we care mostly at this stage of the transformation about engagement. Engagement is critical because we are changing the how we do pretty much everything in the company. That's staying roughly flat.
We're starting to see a jump on employee Net Promoter Score as the changes, no matter how painful they have been, they're starting to take hold in the organization and people starting to say, actually, this is turning into a better place. I'm really proud of this sort of set of results. There is work to do. We expect a much better improved financial performance. If we look at where we are just 13 months after we unveiled the strategy, I think this is solid progress. I'm going to let Pablo take us through the numbers quickly, and I'll be back to talk about execution.
Thank you, Adolfo. Good morning to everyone. I'm going to cover, as usual, the financial performance year- to- date and speak about some of the operational improvements made during the last six months. In this slide, we can see the group-adjusted consolidated results, which have a number of moving parts and understates the improvement that we've seen in Capita Public Service that you can see in the bridge below, delivering strong margin progression. Revenue saw a 3.7% reduction, which is a mix of growth in our largest segment, which is Capita Public , offset by revenue decline in the contact centers that is now 24% of total revenue.
Operating profit has declined year on year by 21.8% or EUR 12 million, primarily reflecting EUR 10 million of contract headwinds in the regulated segment, the earlier timing of the pay award this year, the reinvestment in the business, and all of this partially offset by strong cost savings. Free cash outflow was EUR 26 million outflow, improving by 50%, of which EUR 22 million were payments related to our cost reduction program, together with an improved operating cash flow in the first half. From a housekeeping perspective, we completed our 1-for-15 share consolidation, the share premium cancellation generating distributable reserves, and we extended the maturity of our EUR 250 million RCF to the end of 2027.
The next slide shows the customer reconciliation between our adjusted and reported result. You can see that the main reconciling items were business exits, including our mortgage services due to exiting Q3, and the P&L costs of the ongoing cost reduction program. Moving on to Public Service, which is our largest segment and where we have made the most progress to date with 110 basis points margin improvement to 8%. Revenue grew by 3.8%. This was the best revenue growth seen for a number of years in the division, reflecting the go-live of both Health Assessment Advisory Service and Disabled Students Allowance contracts and expanded scope within our Royal Navy training contract. Operating profit increased 21% to EUR 57 million. We were able to offset the flow-through of the cessation of contracts, the larger than usual growth in staff costs driven by the earlier timing of the annual pay award this year, and the increased NI costs through the cost efficiency program.
Cash conversion was in line with the prior year. This slide, though, does not show the tangible progress made delivering our strategy, given that financials are usually lag indicators. The next slide shows some more tangible examples of the progress made. Public Service is now 62% of the group's revenue. As you can see, we have maintained our strong operational KPI performance at 94%. Total contract value wins were up 53%, including Educational Authority of Northern Ireland, Gas Safety Register, and NHS England’s Primary Care Support England contract. We have also had significant improvements on our win rate, particularly on mid-sized deals that are a strategic priority, with total win rate across the division of 81%, up from 39%, driven by a recently optimized growth function.
As we look to H2 and onwards, our unweighted divisional pipeline has grown over by EUR 1 billion, and innovation and adoption of AI is accelerating, with 90% of bids now including AI in the scope of the solution. Examples of progress from our technology investment were the Silvertown Tunnel document verification solution, the go-live on our debt collection with Appian, opening further opportunities with local councils. Finally, and not covered in this slide, you may have seen an announcement a few weeks ago that Ofgem are investigating Smart DCC, a Capita subsidiary which is not consolidated due to the arm’s-length requirement under the license terms and conditions. The investigation relates to its compliance with license conditions related to historic procurement activities. DCC has been subject to the Ofgem price control every year, and our results have always included the impact of this review.
Turning into Contact Centers. In the Contact Center segment, we have seen 20% revenue decline, partly driven by the annualization impact of the decline of volumes we saw in the second half of 2024 in our telco vertical. We have also seen the impact of other known losses and volume reductions, as we work with our customers driving volumes to our nearshore and offshore delivery centers, reducing revenue with some higher initial costs whilst becoming more efficient and competitive. Margin deteriorated in the first half to -4.1% with EUR 11 million losses, driven by the impact of revenues already explained, partially offset by the results of management action taken to right-size the cost base, which has not yet fully delivered the expected outcome.
Operating cash flow was EUR 21.4 million inflow in H1 2025 compared to a EUR 13.3 million inflow for the prior year, reflecting positive timing of EUR 8 million in cash receipts this year, as well as the usual phasing of significant cash receipts from a large customer that delivers inflows for the year in H1. Turning to some operational metrics, contact centers are now 24% of the group revenues. We have continued investing to transform the cost base of the business, shifted volumes to our offshore and nearshore centers, and invested in AI-enabled solutions, becoming more competitive in the marketplace, whilst increasing our customer satisfaction to four out of five CSAT in our South Africa global delivery center.
After many years of revenue reductions and low wins, we saw renewal rates of 84%, with 100% in the U.K., including Southern Water renewal with a TCV of EUR 92 million with expanded scope, alongside renewals with Borghese and Centrica. We are focused on growing our mid-sized contract pipeline, where we are making good progress to date, and we are being more selective on large-scale deals. There is more work to do to grow the pipeline and crystallize new wins. From a technology perspective, we roll out AI-driven gamification across 4,600 agents, and we now have AI solutions rolled out in 30 clients to 6,000 of our colleagues, allowing improved margins in free contracts.
I encourage you to look at the appendix in pages 49 and 50, showing a long list of contact center technology projects that are either already live or in proof of concept to go live that will keep us competitive and allow us to grow. Moving on to pensions, Pension Solutions revenue was broadly flat. The declining operating profit reflects lower interest rates, offsetting the progress made in the cost savings program. We expect good margin improvement in the second half, driven by revenue growth from key contracts and cost savings. Cash conversion was 70%, reflecting mainly timing of payments from software development activities in our Civil Service pension scheme contract that we expect to even out in the second half of the year. Pension Solutions now is 7% of the group revenues. TCV nearly doubled from H1 2024, driven by the increased win rate of 94%.
Material wins in the first half of the year included Scottish and Newcastle pension plan and the Severn Trent pension scheme, giving us confidence on a positive H2 and a robust full year for the segment. We maintained strong delivery for existing clients with an SLA of 95%, and we saw strong momentum from our investments in digital pension solutions, with 100,000 members signing up to our digital solution, 15 GenAI agents being produced, and a strong uptake on our data analysis and cleansing offering. Moving on to our regulated services business, this is an area we announced at the Capital Markets Day that we're managing for value. We made continued progress on exits in the period, agreeing an exit with our mortgages software business, which has provided a EUR 19 million revenue one-off through the release of deferred income and a net EUR 6 million profit and cash termination fee.
We continue an active dialogue with the last customer from our closed book life and pensions business, seeking an orderly exit from this segment. We continue estimating that this segment will generate EUR 20 million cash losses per year on an ongoing basis, running the services for the last customer. On cash flow, operating cash flow for H1 was EUR 56 million, with conversion of 70%, showing an improvement from the prior year. Timing of deferred income and working capital remained broadly in line with prior year, with slightly higher deferred income and lower adjustments from provisions. We discussed at the year-end the significant relative impact of timing of receipts in a large company like ours, where large receipts can cause significant swings in cash flow performance.
When I look at H1 operating cash flow, my view is that for the group as a whole, we landed in a broadly balanced position, with no significant swings distorting the delivery from the business. In terms of cash spent in our cost reduction program, we spent EUR 21.5 million in the first half, lower than our initial expectations, driven by timing and delivery of savings in the contact center business. To date, we have delivered EUR 205 million savings, and we remain on track to deliver the full EUR 250 million savings, with the previously announced cost of delivery of EUR 55 million for the full year. In terms of cash flow and debt, our free cash outflow has improved to EUR 26 million, reflecting the increase in our cash generated from operations, which includes slower than planned investment in savings, timing of CapEx investment, and the reduction in lease payments.
Closing net debt was EUR 412 million for the half, including EUR 86 million, EUR 87 million of financial debt and EUR 325 million of IFRS 16 lease liabilities, and as I explained in the full year results, does not include EUR 94 million of shop-led lease receivables. Turning into liquidity and net debt, as announced in July, we have extended our RCF maturity a further 12 months, and in line with the full year 2024 position and half-year 2024 position, this facility remained undrawn. This has given us a total liquidity of EUR 380 million at the half-year, including the EUR 94 million of debt issued in March 2025. Net debt to EBITDA remains below the 1x target. Finally, looking at the guidance for the remaining of the year, our outlook remains unchanged for revenue, operating margin, cash flow, and debt.
In terms of revenue, we have upgraded Capita Public to mid-single-digit revenue growth, driven by annualized benefits on new contracts and a strong H1 sales performance, offsetting contact center guidance that has been lowered to a mid-teens revenue reduction, reflecting continued volume reductions mainly in the telecoms vertical and the impact from losses and offshoring activities. We continue to expect Pension Solutions to grow mid-single-digit revenue, driven by growth with existing clients, and in regulated services, we continue to expect a further revenue decline as we hand back contracts.
In terms of margins, we continue to expect overall modest improvement with public service margin improvement, driven by the continued cost savings and revenue growth, contact center margin reduction, reflecting revenue headwinds and timing of cost savings, Pension Solutions margin at higher than the group average, expected to remain stable, and finally, regulated services expected to see a decline in margin given the ongoing exits. In terms of free cash outflow before impact of business exits, we continue to expect between EUR 45 million to EUR 65 million outflow for the full year, of which EUR 55 million relates to the cost reduction program and an improved cash conversion of 55% to 65%. We still expect to be free cash flow positive from the end of 2025. Thank you.
With this, I'll hand back over to Adolfo.
All right. Thank you, Pablo. Very thorough presentation of the numbers. Let me just quickly move into strategy and execution to show you what it is that we're doing, but most importantly, why we are excited about the opportunity of what's coming forward. When I joined 18 months ago, one of the things that I used to say, having spent many years in the world of tech and knowing fairly well what was possible, I talked about my deep belief that the BPO market was going to be fundamentally transformed by tech as a whole, particularly with the advancements around data and AI.
What back in the day seemed like, yeah, he might be right, I think 18 months on, it's very clear that that's definitely happened, and I am really pleased that as an executive team and as a board, we took on the challenge to move very, very quickly in that direction and really just harness that opportunity of seeing the shift from the traditional BPO market, that sort of EUR 50 billion market today, to what's going to become EUR 55 billion, but with a very different makeup. We're going to see the AI-enabled services, the EUR 10 billion of today, moving to 300% growth. I think therein lies the opportunity and the challenge for a company in this sector, and for us is the opportunity that we're striving for. How do we become better at the blue, at the traditional?
Because many customers will still need services that will lend themselves to very little AI, and there's going to be a continued significant market in Western Europe, EUR 25 billion, where we want to be really good. More importantly, how do we build, deploy, and harness the capabilities of AI to, A, migrate existing services to be more efficient and provide a better service through the injection of AI, and B, build brand new services on the back of AI? We're seeing that there is more adoption, and I think you're reading about it if you look at it in the private sector, and the private sector is the test bed for adoption and for innovation. Traditionally, certainly adopting a lot of things faster, and as we have seen in Corinne's world, the contact centers business is our test bed for innovation when it comes to AI.
Yet you look at our largest market, the public sector, now from the Prime Minister down through government funds, the AI playbook, a fully funded plan coming by the government, also the adoption, the work that DCIT is doing, the work that the independent individual government departments are doing. It is all about injecting AI in the public sector as well. It's less about cost cutting. It is more about delivering a better service. It's about working down the backlogs. It's about being faster to deliver citizens' value. Two very different dynamics in private and in the public sector. What we're seeing is the adoption is not going to be linear. There are going to be challenges when it comes to adoption. If I look at the public sector, one of the things the public sector is dealing with now is how do we procure this?
This is not the standard procurement that has happened for a while, and we are having many conversations with the Civil Service to just help them shape possible ways to go to market with these innovations that they're really trying to drive as an agenda item. Beyond that, there's some more generic barriers to adoption, which I believe are turning into great opportunities for Capita. Number one, skills. There's a lot of technology. You could sell a lot of technology. You could sell a lot of technology to a lot of our customers, both in the private and the public sector. These people, these users, these management teams will need to be trained. They will need to be trained on the art of the possible. They will have to be managed through skills. They'll have to be certified. You'll need to have a proper managed skills platform, which is something we have been doing in Capita very well for many years.
The second thing is the tech. Every day you read about a new model. Every day you read about a new application. Every day you read about a new AI agent. Who can keep up with all of that innovation? Who can test all of that innovation? Who can ensure that the innovation works in the right place, that it has the right controls, that has all the right implementation parameters that are acceptable in a regulated industry? I think the role that we can play there on the pre-testing, certifying, prioritizing, and identifying what technology is better suited for what particular business problem, I think that's unique for Capita. Obviously we've got the data maturity.
I think if you look at both, you know, what we have in the local public sector, what we have in central government departments, what we have also in the private sector, data maturity is a challenge. We had Pablo talking about data cleansing services and offerings, and some of the work we're doing with our newly onboarded partners in the data space, Snowflake and Databricks, towards that address. How do we help our customers clean and build and structure and protect the data that they would need to build AI capabilities on top? It's a big shift, and this is why I've spent so much time on this particular slide, but I believe it is a transformational shift for the BPO market and the single biggest opportunity this company has had for many, many years.
As we do that, we've got to do, in short, the sort of Navy much better and continue to do it really, really well, defend our positions and grow our positions on the not-so-AI-rich services, but build more and more of the capabilities on the dark pink so that we can go and disproportionately take a share in that market as that market grows. Which brings me to a little self-assessment on this. This is one of the key slides I shared. At Capital Markets Day, we did the strategic review that was done by OC&C. They categorized our services offerings, bucketized them into where do we build star positions over the years. Star positions, places where we win often, we win well, we deliver well, we have a strong reputation, we have good demonstrable skills, expertise, and differentiation, and we have to just do more of.
I think you're starting to see through some of the numbers that Pablo shared, we're starting to see a lot of that progress, in particular around pensions, around the public sector. There is now real clear evidence that we're doubling down on that space and we're doing really well. We had a number of areas we identified as transformation potentials. Good markets, nothing wrong with the market, but there was something that needed to be improved in our execution. That's all the areas where I think we're making a lot of progress in most of them. I think in contact centers, Pablo reflected on, we are applying the right medicines. We are doing what we should be doing in this market. There is nothing wrong with that market. If you look at the market leaders, they're getting above 10%, 10% to 15%, 17% net margins in that space.
They started a lot earlier. They transformed the business a lot earlier. We are four or five years late to do that transformation, and it's taking us a little bit longer to catch up. We are seeing already on those accounts where we deploy the tools and the solutions that we've created, we are already seeing significant operational improvement in average handling time going down, increasing the first call resolution time, increasing CSAT, increasing Net Promoter Score. We're seeing account margin already improving, and it will take a while until that cascades through the P&L. It's definitely an area where we needed to do more work. We got the manage for value bucket, which I wanted to sort of double-click on a little bit because there was a lot of nuance. Remember, manage for value was there are areas where we just simply are in the wrong place.
Either we had the wrong place, we were in the wrong place, or we had the wrong offering altogether. It was going to need some work. There was going to be some more transformational work, like it was the exit of Capita One. Full transformations, like the one we're doing, for example, on our networking services. We're going from the traditional system integrator role that we did of a lot of third-party equipment and trying to provide networking services to our clients as part of larger outsourcing contracts to now fully migrate into effectively what it is, networking on the cloud type services in deep partnership with AWS so that it becomes a more soft offering on the cloud with very cost-efficient security from the outset, asset light, CapEx light, networking services. That's just been a strike. Similarly, on IT managed services, it was heavily manual.
When we started, we were managing ourselves internally heavily manual, and we were managing our customers heavily manual. That was an absolute inefficiency. Since then, we have already moved over 25,000 employees for ticketing inside Capita. Now that we've got our experience with ServiceNow as client zero, we are now migrating our existing contracts over to that, as we did in the call center business with Amazon Connect, where we have been migrating most of our customers' telephony stacks over to that platform. Obviously, there is still work to do on closed book, like on pension. As Pablo mentioned, we continue on the handback of the great exits, but there's still one more that we're working on to agree and move forward.
Overall, as you look at that sort of bucket of manage for value, it's taking quite a bit of management time, but it's important to see that execution is happening. From an agentic, and apologies for those of you who might be fluent in the topic, I thought it would probably be good to give you the Capita perspective of why AI, what type of AI, what's our role in AI, and why is it relevant ultimately to the value creation of AI for Capita. I think if you look at three different types of AI, you've got sort of the predictive AI that started a little bit earlier. It doesn't mean that it's aged; if anything, a lot of value still today derives from predictive AI.
It's the capability of getting existing data that the company has and getting it to work together so you get early insights and actionable insights. It might be less sexy. It might be something that we got really good at a few years ago, but do not underestimate the raw and sheer power of that tool. We continue to invest in that area. For example, you look at Call Sight, which is a GenAI-based tool that we've deployed there to get real-time visibility of the quality of every single call that we make on behalf of our customers, no longer sampling maybe 5% of the calls. Now we can do 100% of the calls, just a particular example. You've got the generative AI, which is the ability to generate content based on content. That could be generating code based on existing code. It will be generating answers based on knowledge bases.
That's something that became very prominent with the arrival of GPT-4 back in December 2023. There we've seen a lot. It's the ability to use all the content you have to generate new content. It's too big an opportunity not to use. It's no longer just getting data to give you insights. It's now getting sheer values of content, whether it's in the knowledge space or in the actuation space, just to get you to the next best action or the next best content. We're already deploying that. We talked about the five clients where we deployed this, for example, on the Agent Assist tool and the Capita Accelerate, as we're doing it with the Army Recruitment Program and a number of others. Definitely a play.
The next frontier is not only I'm going to predict data based on existing data, not only I'm going to generate new content based on existing content, but actually now I'm going to act. This is sort of the agentic AI wave. It's the sort of the programming of the activities and on the actions that will happen. It's a more automated, it's a more independent. It can be as independent as you want it to be. We have decided in Capita that we don't want it to be 100% independent. We work in regulated industries. A lot of the public sector, very sensitive materials, a lot of financial institutions, a lot of regulated industries. We want to specialize in a human-in-the-loop type of agentic capability. Everything that has to be customer-friendly.
Internally, for some of the work that we do internally, that is purely efficiencies-driven, that maybe we can just have less of a human in the loop, but when it comes to external, it's very important that we do that. You see massive investment being made by hyperscalers to deliver on those capabilities. Very important things. Very few, actually hardly anybody, is delivering endpoint agents. What they're doing is creating productivity environments, development environments where anybody who understands a business process can map up and create that particular agent. I think if you look at that from a Capita perspective, we don't have to worry about any of this, because that's already coming. All we have to worry about is for what we do really well.
What we've been doing for 40 years is understanding that customer expertise, understanding what the business process is, and then just go and map it. Map it to the right agentic platform or to the right data platform at any point in time. I actually think, I really want this sort of sentence to sink in, is I think in the future, my vision is that we will be very well at orchestrating agents and humans to work together to deliver those outcomes. I'm going to show you some examples already where we are orchestrating multiple agents. If you think about the concept of multiple agents independently with humans in the loop, and then Capita creating that master capability to do that. We're blessed. I think we were blessed and lucky that we don't have to do anything that it is in the bottom part of the stack.
There are trillions of dollars of investment going into sorting out the enabling layers of AI, whether it's at the microprocessor level, at the computer architecture, data layer, all the way to the application layer. There is a lot of money, a lot of billions. It's probably the first wave of investment that I see in my tech years where everything is in the tens of billions or the hundreds of billions. Otherwise, you don't even grab the headlines. Fortunately, we don't have to do anything like that. All we have to do is leverage all of that investment and understand what is it that we do well. It's mapping those customers that we know well, the business processes to the existing capabilities that are coming there every day. It's ensuring that we select the best, more suitable, more secure, more agile, more price-effective solution at any point in time.
I think it's that how do you use all of these tech innovations with a human in the loop? How do you create the guardrails that are important? Working with our customers to understand what is it, the compliance requirements that they have, what data needs to be made available under what circumstances? What actions need to be logged for what purposes? How do you audit? How do you limit all of these agents? I think that is at the core of what we do because that's what we've been doing for decades. I think that's where the delivering of a managed outcome, not only with people as we have in the past, but a managed outcome leveraging people and leveraging tech becomes the secret sauce for us. We are going to be doing this across the whole of Capita.
Bear with me with this slide because I was trying to just explain a number of concepts which I think are very important, and it will be very important going forward. If you look at what we deliver for customers, you could oversimplify it by saying we do front office work, everything that touches the end user, citizen, consumer. There is middle office work, which is sort of doing some reconciliation or some processing that needs to happen to support the front office. There is some stuff that happens somewhat disconnected from the front office, but they go deep into process and systems.
You will find that in some customers we do all three, but in general, maybe oversimplifying it a bit, everything that we do in customer experience, both in contact centers and pension solutions, is very front office heavy, with a thin veneer of middle and back office, depending. For example, think of the work that we do for water utilities. You will find in the water utility front office, we engage with email and call centers. We will do some work around payments. We will do some work giving them connections and disconnections of consumers. In some areas, we might even take responsibility to notify leaks and things like that. That is sort of a very thin veneer. Most of the innovation at the moment is in the front office, right?
In Corinne's world, we're looking at how do we get productivity out of the agents, how do we get them to do a better service, how do we get them to have better data. We're looking at quality assessments of 100% of the calls, as I talked about. Now we're looking at, from a sales and an outbound perspective, how do we give them the tools so that they can be more effective in converting leads. That is kind of the work. If you look at our public sector business, it's actually the opposite. A lot of what we do is middle and back office. In many of our customers, there is a front office citizen engagement.
A lot of what we do is just really complex middle and back-end processes that will have a very different AI investment profile, somewhat different tools, but we will apply some of the same tools when it comes to citizen support. In every one of these interactions, we are going to be working with a human in the loop, right? We will not let our agents lose when it comes to medical assessments or when it comes to social assessments or when it comes to the leader of different things, but we might just get the agents to provide some solutions. This is important because I think what you're going to see over months and years to come, the profile of investments are going to be different. The profile of opportunities is different.
In today's world, as you can see from Pablo's world, we're seeing more traction on the Capita Public Service and the work in the middle and the back office through a big automation, big streamlining of middle and back office. We're doing a lot of the test bed, both in digital pensions and in contact centers. Digital pensions, we're already seeing the pull-through that's working call centers, we're seeing the pull-through at an account level, but not yet at a divisional level. This gives you. In the front office, we talked about agent suites, something like we're doing, we're already doing on Southern Water, 17% improvement rate on average handling time. Similar work that we're doing now on the Department for Work and Pensions, where we're looking at proof of concept to do call transcript and just getting some of the quality done there. That's very typical sort of front office activations.
In the video, you heard about the discount verification scheme that we put together for Transport for London on the Silvertown Tunnels. How do you deal with all of these neighborhoods where there's so many different cases as to why you might qualify for a discount? There are so many documents that need to be verified, and you need to look at fraud. Traditionally, that would have been a purely human looking at all the documents one by one, and then eventually in a few days coming back. Now we've got that capability built in, so that is the technology that does the fraud verification, the document correlation, and it can do that.
On the back office, if you think about a concept like the debt collection, how do you go about collecting debt? How do you identify the most likely to pay? How do you prioritize the right payers? How do you build a workflow associated with getting the best possible yield of your debt collection efforts? There's something that is already working in Bexley, and we've got a good pipeline of authorities. It's trying to show you that AI is a lot more than just talking to ChatGPT-4 and getting there. This is technology at work today solving customer problems. The final thing is, how is that going to benefit Capita? How are we ultimately going to get the financial return? If you think about the foundational layer of what we know, the people we got, the process, the experience, all of those projects at work are generating data. The data we've got lends us to start working on automation, automation opportunities. The automation opportunities will give us two benefits.
They'll give us one internal benefit, which is because it's automated, you'll do less manual work. Because you do less manual work, our colleagues benefit from a better work experience because they don't have the tedious retyping and typing things in two or three screens. These things are done. As a result of that, obviously, they have a better work experience. They have a better work experience and are more likely to stay. They're more likely to stay. We'll have lower attrition, and we have lower attrition. We have lower costs of attrition. That improves our internal productivity numbers and our numbers. Externally, it allows us to build repeatable offerings, things that we can reuse in other parts of the portfolio. We don't have to be bespoke. We don't have to have innovation cost of one every single time. That will give us obviously a much better productivity.
I think as we get all of this value flowing through, we are a cheaper provider because you're a cheaper provider. You win more because you win more. You build more understanding of processes. You get more data. This is where the flywheel starts gathering speed. This is how you start making money. This is more than AI's an ingredient, and it is a lot more than AI's justified. This is the flywheel of how we are going to become one of the leaders in business process-enabled AI with a human in the loop. We couldn't do this without our hyperscaler partners. I remember when I started talking about them 18 months ago, there was some talk about, are they really going to really help you? Yeah, they are being absolutely critical, fundamental to everything that we do.
I've talked about ServiceNow. We've talked about the work. You saw some video there from Darren on Copilot. We're going now to 260,000 monthly Copilot interactions inside Capita. We're saving now about 9,000 hours of work per month inside the company and just growing very, very quickly, just to name a few. As we go forward into the second half, I'm really excited about the stuff that we've got. I'll call out. We have been selected by AWS to be the EMEA reference partner to launch something that is the process to agent automation. You see the P2A program. That'll give us the ability to shape what goes into the offering and be able to be one of the first movers there. Similarly, there is a lot of interest from our Microsoft Copilot adoption program for the public sector and some of the work that we're going to be doing to build some hyperscaler-based offerings in some areas.
Good progress, really excited about what's coming forward. I talked about the Catalyst, something that we launched in February, an internal capability where we have recruited a lot of external people who came from this world to engage first internally and then now with customers on where can we adopt this thing. We are doing interviews. We're getting ideas. We're getting submissions. Over 300 ideas have been submitted already. They get validated. They get prioritized. They get technified where it makes sense. We already have 70 agents there. We're just getting started. Many customers have expressed interest and are already engaging with the AI Catalyst Labs to look at, okay, help me out. Do you know this process? What could we do? What is the art of the possible? You saw Sarah earlier from Salesforce talking about the AI recruitment agent based on Agentforce.
Just think about this was an idea we had in November, December that we went live with the early version end of May. We already have 10% of all the engagement of our recruitment going through it. It's just really important to see that the 19,000 candidates have already engaged in that very little time. We're part of the solution. The solution continues to be built up. By the end of the year, we expect it to be 100% of the internal candidates. Over time, once we got this to work at scale internally, we'll be taking it out to market. All moving very quickly, delivering already. As you can see, there's some benefit in the few weeks to recruiters who are seeing the benefit. It will get a lot more sophisticated. I talked about earlier my vision of Capita being the best orchestrators for agents and humans.
I think if you look at what we've got on the left, that gets pretty. This is something that we're doing in the government space where you have a multi-agent capability, where multiple agents talk to each other. You have a generalist agent, you have a specialized agent, and you have a QA agent and a human, all of them talking to each other, ensuring that when this particular work in assessment is performed, a human is assisted with a recommendation in real time when it comes to rules, policies, but also to a particular specialized knowledge all the way down to performing quality checks. Super important, very, very applicable everywhere. They're a bit more mundane, but equally important as we do fire assessments, and this is a part of the work that we do in fire service assessment. Very important for productivity. They're very long pages.
They need to be submitted. They need to be addressed. They need to be verified. You need to be validated. You need to make sure you check against standards. You need to make sure that you do it timely and consistently. We've already got that going on. We've got another one in the pensions world, which is really important, is how do they improve the outbound response to the inbounds? How do you catalog, understand what's coming in? How do you start bucketizing the points of the information, the areas of information? How do you start drafting responses? You can have the responses sent automatically. You can have the responses being coursed through a human. It is effectively just your pulse and your ability to get really good response times to your customers with a human in the loop.
You are starting to see all of these things already working in different parts of Capita. Pablo talked about efficiencies. I will only just say that we remain very committed to it. It was EUR 190 million at the end of June, EUR 205 million at the end of July. We have got a run rate of about EUR 10 million a month historically. Yes, it is hard. It is hard for our colleagues. It is requiring a lot of sacrifices, a lot of hard work. We are now finding that sort of EUR 10 million per month run rate, and we will continue to do that in multiple areas. I think mostly it is in the areas of fixing the basics. Right? How do we look at all the processing efficiencies? That is the EOS project.
How do we do more shared service centers? How do we go, particularly now, into the enterprise data platform space? How do we go and get more sophisticated around the data that we have got, where we have got it, so that we can drive savings and efficiency when it comes to quality, when it manages to manage information, and a number of others? We will continue tackling legacy. Very important, because I think we still have legacy in the system, moving legacy to the cloud. I talked about the process to agent work that will be happening in the second half, as well as the culture around cost efficiency. On delivery, Pablo touched on most of these. It is important to see momentum across all the key areas.
When it is TCV growth, or when we are looking at win rates, or when we are looking at the book- to- bill, or when we are looking at the size of the pipeline. Again, some of the great wins, and we have also got some good opportunities going into the second half. Talked about in the intro, very important to me, the values, the high-performance culture that we are trying to build as we transfer into this new world. I am in love with the four values. I think they really tell the story of who we want to be and what kind of behaviors, from me down, we all have to display every day. This is the customer focus, the innovation, the delivering together, both internally and with our partners, and the sense of inclusion.
As I said earlier, I think we just started to see early signs that it's catching on with the organization. I'm very pleased by that because that works hand in hand with a very contentious topic, which is, is it humans or is it AI? I am very clear this is both. This is humans working with AI. There's a lot of things that humans do a lot better than machines, and there are machines that do things a lot better than humans. How do we get them to work together? A lot of what we've done has been first disinformation, then is sharing the successes, then see they're seeing that actually their work becomes a lot more palatable. Most importantly, you have to invest in your people.
We have been investing all the way down from the leadership team, where we've got 600 people taking the leadership and the management academies, all the way down to data academies, AI academies. We've got 10,000 content modules being accessed. This is, -- we're just getting started. It's fundamental to our value proposition that yes, there is an impact of AI in jobs in certain areas, that can be negative, but also positive, right? Enhancing some of the existing jobs and creating other jobs. Excited about catching that opportunity. Then something that you would expect us to do really well, because we've been doing this for 40 years, which is managing well the deployment of this, managing the ethics associated with AI. How do you manage to make sure that the same level of compliance that we have to security, national security, regulatory requirements, ethics are captured from the AI perspective?
What we've done is rather than create something ad hoc, we're building on the existing solid Capita processes. We've enlarged them to also cater for the same levels of control and governance to include AI-based solutions. We're in a position to have the same level of conversation and reassurance to our customers. Last but not least, bringing up a good old slide from the past because I said this is what we're going to do, and I think this is what we're doing. I talked about three different waves of transformation. I talked about funding the journey. I talked about fixing the basics, and I talked about investing for growth. I did say they're not sequential. They're going to happen in parallel with a bit more of one at a given point in time.
We continue to be heavy on the cost transformation to fund the journey, but we are already moving into the innovation in terms of ways of working, fixing the basics. As you have seen already, building capabilities and seeing early signs when it comes to investment for growth, all of these three things have to happen in parallel. It's not trivial. Sometimes it's complex to drive this amount of change in the organization. I am pleased and proud of how the management team, the leadership team, and the executive team are driving this. To Pablo's point, good progress, solid progress in some areas. In some areas, the financials are already telling the story. I'm particularly proud of seeing two of our three operating groups already ahead of the operating margin ambitions with one to work on.
I think as I look into the second half priorities, it's the focus on that third one that is missing, which is contact centers, which obviously needs to be there. We continue to work on manage for value and the last remaining bits that we need to solve. We're just cracking on with the cost savings, the reinvestment, and just building on this refreshed culture. I'd also like to take the opportunity to thank all of our customers who have supported us as we continue to go through the transformation, our patients, shareholders, and our colleagues who are working through really hard, very difficult months of transition in the company.
I think hopefully all three stakeholder communities would agree that we are on that journey to build a better Capita at the time when it mattered the most, which is at an inflection point in the market that we're trying to leverage.
Thank you very much. Now we'll take any questions. Helen or David?
Good morning. It's David Brockton from Deutsche Numis. A few questions. Firstly, starting with the contact centers, you touched on how the peers are generating 10% to 15% margins with the cost savings that you're pushing through and the AI that you're now implementing within that business. Are you confident that even on the scale of this business that you have, which is lower than some of those bigger peers, that you can achieve a comparable margin? Do you see a route to that now?
I think if you look at what they have going for them, right, if you look at technology investment, tick, we can do that. Moving to offshoring, nearshoring, and get that value proposition. Yes, we're doing it at a revenue cost, but we're doing it because it's right for our customers and it's right for the profitability. If you look at the industry specialization of the offerings, we're already doing that as well. The only one where I can't tick is scale. We do have a contact centers business that is mostly Western European with two really good delivery capabilities, an extremely large one in South Africa and one in India. I've got some nearshoring in Eastern Europe. That appeals really, really well for customers who are just like that. Corinne always says, you know, we're really good at selling to customers who are like us.
If there's a global player who needs global solutions, somebody who needs delivery out of Central America or delivery out of the Philippines, we are not in that bracket. What we're seeing today is for customers where we look like them or they look like us, we can actually match that performance. The answer is yes, but we're not there yet. It's going to take us a while to get to that. I would say first I need to get that business to six to eight and then to see what's the art of the possible because these other guys had the scale and they started four to five years ago.
Thank you. The second question area relates to the cost savings. I think you referenced that there was a slower phasing of that through the first half. Can you just touch on why that was the case? Secondly, when you unveiled the EUR 250 million, you talked about reinvesting EUR 50 million. Can you give an update on where that reinvestment is, please?
Sure. As you get going, the transformation gets harder and you need to be more intentional. As technology evolves and we see the art of the possible on agentic AI, remember nobody was even mentioning agentic, not once, when we were in the same place a year ago. As these solutions become apparent, you start looking at the art of the possible. You start looking at, okay, how do I map this? All of a sudden that has opened different windows. We would have gone the wrong way. We are now evaluating differently. We also have to execute these changes in a responsible manner with regards to our customers. You take all of these three things combined, we have decided just to be more paused. Don't get me wrong, we're at EUR 205 million versus EUR 250 million. We are actually in a really good place.
I think Pablo and his team are actually being very intentional at checking that every investment that we make of cost to achieve is the right investment and we've got the right insurance around it. That sort of speaks to why we were there. The reinvestment of EUR 50 million, I think if you sort of look at this presentation and compare it with the presentation we had a year ago, you will see a lot of it is around this area of working with the hyperscalers, building the innovations, bringing in new teams, getting the AI capabilities. There have been areas where we had to fortify execution capabilities as well in the business as usual business so that we can protect the sort of the Blue Navy base.
There were also some red projects last year where we had to invest in them to get them out of the red space. That is the space where most of the money's gone to.
Has the EUR 50 million been reinvested or have you only done some? [crosstalk]
The plan is, by the end of the year, it will be reinvested.
Thanks.
Hi, it's James Rose from Barclays . I've got two, please. The first is on public. It's good to see book- to- bill back over 1x . I think within that, could you comment on how much TCV or your win rates are on the business which has that higher technology underpinning? Essentially, are you winning more of what you want to win? The second one is on contact centers. Could you elaborate a little bit on why revenues are behind your expectations from the start of the year? I think there's a broader question in there as well. As the market is changing and AI is being deployed, how does that affect revenues in that broader business?
Okay. Let me take it and then Pablo can chime in. On the first one, on CPS, where we have seen the biggest uptake. First of all, as you can see, the TCV at 53% is way ahead of the group's average. First thing out there. Second thing is we've seen the biggest uptake in what we call the mid-size deals. You might remember a year ago when we presented this strategic review that OC&C did with our team, that was identified as one of the singular levers of work where we had traditionally not engaged really well as a company. We took an action a year ago and now it's already flowing through. Many of them, those tend to be shorter sales cycles and they tend to have a technology underpin. As such, you're already seeing it at work.
With CE, the contact centers, the issue is not AI. AI is actually part of the solution. Structurally, the issues that we have seen, we have seen offshoring. As you offshore, yes, you save costs, but you reduce revenue. We've had the under. Realization Of the telecoms business that has come through it, we've had some known losses that are now going through the P&L. I'm looking at AI as part of the solution. If you look at Corinne's get-well plan, effectively, there is a menu of medicines that we apply account by account. We look at things like, what are we going to do on average handling time, where that is our priority. What are we doing in terms of first call resolution? That's our priority in terms of tech. What are we doing in terms of gamification with this in tech? What are we doing with digital agents, whether it's in tech? What are we doing with offshoring or nearshoring? We're using AI as part of the solution.
This is one of those where the financial numbers are telling a story and the operational numbers that we're seeing and the early indicators are telling a very different one. I don't think there's going to be any. There is an opportunity, certainly a massive opportunity to leverage AI for contact centers. If I look at our peers in the industry, they've all sort of been flattish or growing. They all have 12% to 15%, 17% net margin in those areas. There is no reason why over time we can't get it to be much better than it is today. We will get to the scale question in the future.
Thank you.
Anything else?
Thank you. I've got a few questions online. First couple are for Pablo around cash flow, which is, do you have an early view of what the free cash flow run rate could be for 2026? Is it possible to give an indication on that metric once the severance headwinds are out of the way and more costs are removed from the business? That's the first part. The second part is probably related, which is, what would it take to get cash conversion higher than the 55% to 65% rate that we've guided for for this year that you're currently expecting?
Perfect. Thank you very much, Helen. In terms of views of cash flow for 2026, we've got on the website where consensus is. It is around the EUR 20 million free cash flow, excluding any other investment. We're not continuing with any headline so far on big restructurings. It will become part of our business as usual. I've got nothing to say to that number at this stage. In relation to cash conversion higher to 55% to 65% this year, that is where we expect it to land. That cash conversion of this business is going to improve and it's going to be a matter of over time.
I think I explained at the half year results that one of the things that we're covering in this business is the consequence of IFRS 15 adoption back in 2017, I think it was, where we are forced to spread the big transformation of the large contracts over the life of the contracts. To spread it, when you spend the money, you create an asset for the transformation you've built and done. When you get the cash from it at the beginning too, you book deferred income. What happened is that the asset was written off because the contracts were not profit making, and we got the lag of a deferred income flowing into revenue and profit every year since then. That used to be around EUR 100 million up to 2023. 2023 was EUR 100 million. 2024 was EUR 80 million.
I suggested that it would go down to EUR 50 million, EUR 50 million, and then continue to tail down to EUR 30 million. New contracts with transformation come with a deferred income, but also with a contract asset and both broadly depreciate and are amortized over the life of the contract. The issue is this historical legacy that is slowly coming down.
Thank you. I have a final question, which I think you partly addressed, but maybe you'd like to just put to bed. Can any information be given regarding the Ofgem investigation and also what could be the financial implications for Capita?
In relation to Ofgem investigation, we've got very little to say. Ofgem, we spoke about it in the year-end results. Ofgem is subject every year to a pricing control review by the regulator, the DCC business, sorry. We've been going through it year after year after year after year. This is a specific point in relation to historical purchasing procedures in that company. The company has its own board, it's self-regulated and independent from Capita as required by the license conditions. Other than that, we know that the company DCC will collaborate with Ofgem and provide everything that is needed, but we've got nothing else we can add to that.
For those that don't follow that, obviously as part of that governance, these processes have all been audited in the past. Obviously, they're there, and as an industry, the procurement processes are guided and followed by the government procurement process. It's a standard, well-audited, governed entity and just looking forward to its resolution.
Thank you, sir. I have got one more question popped up from Chris Bamberry. The contact center renewal rate was better in the U.K. than elsewhere. Were there any particular reasons behind that?
I think we're doing a good job for the contact center customers. If you look at not only the renewal rate, but if you look at CSAT, a four out of five, the preliminary data that we got around customer NPS significantly improving. If you look at the menu of options that now we give them with what do they want to do in the U.K., what do they want to do nearshore, what do they want to do offshore. If you look at the pilots, the quality of the pilots that they see already, they see that, okay, we got this menu that I was talking about earlier. Many of the U.K. customers are already looking into those and they're driving those.
People say, when you get such a high renewal rate, it speaks about the quality of the relationships, of the account management, of the delivery, of the innovation, of the right shoring strategy. I think we are onto something, which unfortunately is not showing in the numbers just yet, but there is no other way to get the financials in right until you get the operations right, you get the customer satisfaction right. We are seeing that already. That's one of the indicators.
Okay, thank you. There's no further questions online.
Any other questions in the room? If not. O kay. I wasn't quite sure. Once again, thank you very much. I think this concludes our first half 2025 results session. Thanks to all of you watching remotely and all of you in the room. Thank you.