Pendragon PLC

Good morning, everyone, and welcome to our FY '25 results presentation. I'm Bill Berman, CEO, and I'm joined today by our CFO, Ollie Mann.

2025 was our second year as a stand-alone AI embedded automotive technology company, and I'm really proud of the progress we've made in that time. The past 12 months have been particularly significant for us due to a number of key transactions and important operational milestones.

Ollie will take you through the headline financials shortly. But before that, I'm going to take you through an overview of our strategic and operational progress delivered in the period.

We have grown our revenue by 30%, driven by strong growth among new customers, successful upselling to our existing customers and the revenue added from our Seez acquisition in March 2025. In July 2025, there were 2 events that were key to Pinewood's future.

Firstly, we bought Lithia out of the share of the North American joint venture. This removed the competitive overhang that existed and risked impeding our growth prospects in the key market.

Secondly, we signed a long-term $60 million contract with Lithia to implement our system in all their U.S. dealerships across the U.S.

and Canada. Meanwhile, we've been carrying out testing on the system in North America, which is progressing well.

The largest piece of work that we have been focusing on is the North America OEM integrations and is progressing as planned. Almost all of our OEM partners have now been engaged with and integration work is on target.

Another key milestone in the period was the acquisition of Seez, the automotive AI company in March of this year. Since then, we have made great progress bringing Seez into the Pinewood Group and integrated the 2 tech stacks to significantly enhance our AI capabilities.

The contract we signed in March 2025 with Global Auto Holdings included Lookers, one of the largest auto dealerships in the U.K. The system rollout with Lookers started in July 2025 and is progressing well and on schedule.

It is due to complete in Q4 of 2026. I'll now hand over to Ollie to take us through the financial review.

Thanks, Bill. Good morning, everyone.

We delivered strong revenue growth of approximately 30% to GBP 40.5 million. This increase was due to a number of factors, including the impact from new customers, upsell to our existing customer base and the revenue added from our AI acquisition, Seez.

We also saw the impact of FY '25 being a 12-month period and FY '24 being an 11-month period. Gross profit of GBP 34.7 million was 23% up on last year.

The gross margin rate of 85.7% fell as expected and reflects Seez gross margins being slightly lower than the legacy Pinewood business. The key profit metric that we use both internally and externally is underlying EBITDA.

In 2025, this was GBP 16.4 million, up 17.1% on FY '24. Our recurring revenue of 83.2% is a key metric for us.

The slight drop from last year reflects the mix with revenue from Seez now included. Our net customer churn of 2.5%, while higher than FY '24 is still at a very low level and highlights how much customers value the Pinewood platform.

A new metric that we've introduced is total contract value. This is the amount of incremental annual recurring revenue from signed customers that have not yet started their Pinewood system implementation.

If the customer is partway through their implementation, the revenue included in total contract value is just the element from dealerships not yet implemented. Our total contract value of GBP 64.5 million on top of our existing revenue will deliver us the majority of our FY '28 EBITDA target of GBP 58 million to GBP 62 million.

On to Slide 7, where I'll take you through the key movements in our cash flow during the year. During FY '25, we generated GBP 6.5 million of cash from operations.

There was a significant gain of GBP 60.8 million on the buyout of Lithia's share of the North American JV that was noncash. Other key movements in cash included GBP 1.1 million of bank interest received in the period and GBP 10 million received from Lithia for the settlement of a tax debtor.

The capital expenditure figure of GBP 11.4 million in FY '25 includes GBP 10.5 million of capitalized development spend. During the year, there was a net spend on acquisitions and reseller buyouts of GBP 13.5 million.

This was made up of GBP 26 million spent on the Seez acquisition, GBP 2.5 million spent on the South Africa reseller buyout, offset by GBP 15 million of cash received as part of the U.S. JV buyout.

Alongside the Seez acquisition was the equity raise that we undertook in March 2025, where we raised GBP 34.1 million of cash. All of these movements led to an end of December 2025 cash position of GBP 34.1 million.

On Slide 8 is the end of December 2025 balance sheet. There were closing shareholders' funds of GBP 204.2 million, with the main driver of the increase from the end of 2024 being the March 2025 equity raise and the July 2025 North American JV buyout.

We now have goodwill of GBP 51.5 million on our balance sheet due to the buyout of Lithia's share of the North American JV, the Seez acquisition and the South Africa reseller buyout. The increase in our other intangible assets from GBP 16.3 million to GBP 168 million was driven by the North America JV buyout and the Seez transactions with GBP 125 million of the year-end balance relating to the North America customer contract.

The majority of the tax liability we have relates to the U.S. JV buyout and will unwind as the intangible assets are amortized.

Finally, we have GBP 34.1 million in cash. In addition to this cash, we also have a GBP 10 million RCF facility, which remains undrawn.

On to Slide 9, where we show the non-underlying items for FY '25. There was GBP 4.6 million of costs related to FY '25 acquisitions.

Most of this expense was legal fees related to the North America JV buyout and the Seez acquisition. Following the buyout of Lithia's share of the North American JV in July 2025, we have incurred GBP 4.2 million of costs in our U.S.

subsidiary. As guided previously, these costs will be treated as non-underlying until the Pinewood system is installed in 20 North American dealerships.

Our share-based payment charge was GBP 3.6 million in FY '25, and there was a GBP 4 million amortization charge relating to intangibles arising on acquisition. The final item to call out is that there was a one-off gain of GBP 60.8 million that we recognized on the North American JV buyout.

On Slide 10, it's confirmation that our current FY '28 guidance is unchanged. This is underlying EBITDA of GBP 58 million to GBP 62 million in FY '28.

Approximately GBP 50 million of this figure is covered by existing customers and signed contracts. In terms of FY '26, 2 of the 3 analysts that cover us released notes following our recent business update on the 25th of March 2026.

The consensus of their FY '26 underlying EBITDA numbers was GBP 21.3 million. We expect underlying EBITDA for FY '26 to be in line with this.

I'll now hand back to Bill to run through the operating highlights and strategy.

Thanks, Ollie. I thought it would be helpful to take the opportunity to remind everyone about the fundamental strength of Pinewood.AI and what sets us apart from our competitors.

Firstly, being 100% cloud-based with one code base is extremely unusual in our industry. Our customers are working on the same version of the Pinewood system, whether they're in the U.K., U.S., Europe, Asia or anywhere in the world.

This gives us a number of big advantages. From a security point of view, it's much more secure being in a cloud-based platform than being self-hosted or using on-prem servers.

Additionally, we are able to release multiple software updates a week during customer working hours with no disruption at all, which is certainly not the case for the majority of our competitors. Our customer churn is incredibly low with customers generally only leaving due to M&A or network consolidation.

The majority of our revenue is recurring and is generally around 85%. The main part of our revenue that is not reoccurring is implementation income.

Another key differentiator from our competitors is having a fully embedded AI solution as part of our customer offering. Although having an AI offering is now standard, in most cases, it is a layered app on top of the system or basic API.

What we have achieved with Seez is very different from this and is a fully embedded AI offering across our entire system, whether you're in vehicle sales, after sales or back office accounting. With margins being squeezed across the auto retail industry, what we can offer with our AI-powered solution is incredibly powerful.

While we continue to innovate and recruit new talent, our experienced workforce underpins what we offer. We have 40 years of experience in the automotive industry, which has made a huge difference in us developing the system we have today.

Finally, our deep integrations with our OEM partners remains key to us. We remain committed to being among the best partner for OEMs through our cutting-edge technology and our ability to adapt to the ever-changing auto retail landscape.

I'll now take you through the progress we've made against our strategy during 2025. In the U.K.

and Ireland, we started their Lookers implementation in July 2025, and we'll continue the rollout through 2026 when we expect to be completed. The combined Pinewood and Lookers teams have done an excellent job on the rollout.

At the request of Marshalls, we have moved the start date of their implementation from Q1 2026 to the second half of 2026. This is so they can align their timing of the Pinewood rollout with a number of other projects they are working on.

In our international markets, the Japan Porsche implementation started successfully in December 2025 with the first dealerships going live. We look forward to getting all the Porsche Japan dealers onto the Pinewood system and then starting the Japan Volkswagen implementation.

We continue to be in discussions with a number of potential Central European groups. The buyout of our South African and Netherlands reseller is now completed, and we are delighted to welcome their teams into the Pinewood team.

These acquisitions enable us to fully control our sales and customer service functions in these markets and put us in a stronger position to drive our continued growth in both regions. The acquisition of Seez in March 2025 has made a huge difference to our customer upsell offering as well as diversifying our revenue streams.

The Seez team put a huge amount of focus into the rollout of Seez products into the U.S. and U.K.

dealers as well as growing their wider customer base. The final pillar of our strategy, North America is comfortably the largest automotive retail market in the world.

And I'll take you through our progress on this on the following slides. On Slide 14, we set out our expansion plans into the North American market and the progress we have been making.

With 20,000 franchise dealers in North America, this is comfortably the largest market in the world with a total addressable market of over $9 billion. We think we are in a strong position to scale across the U.S.

and Canadian markets. As a reminder, we now have a significant foothold in the key market following the $60 million contract signed with Lithia to implement the Pinewood system in all of their dealerships in the United States and Canada.

Testing on a number of areas of the system in the U.S. is well underway with the full rollout expected to start no later than in 2026.

The largest piece of the development work that we have for North America is integrating our system into all the U.S. and Canadian OEMs.

We have now engaged with the majority of the OEMs that Lithia represent and integration work is well progressed with a number of these. We are targeting to have the majority of the North American OEMs integrations completed by the end of 2026.

Having now bought Lithia out of their share of the North American JV, we are well placed to now sign further customers in the U.S. and Canada and have a number of leads on our official U.S.

launch at the February 2026 NADA conference that we are now currently working through. Moving on to Slide 15 and an update to our recent U.K.

system implementation. Our team have done a brilliant job so far on the Lookers implementations, working alongside the Lookers team to ensure the Pinewood platform rollout goes as smoothly as possible.

It started in mid-2025 and will continue through 2026. We are pleased that the Lithia U.K.

team are seeing the benefits of having Pinewood.AI in all the dealerships, and we continue to work with them to help drive their business forward. We are looking forward to starting the Marshalls implementation in the second half of 2026.

The Marshall and Pinewood teams have worked together extremely well in the planning phase of this project. We continue to expand our range of products and offer all of these to our existing customers to help them drive their business forward by both increasing productivity as well as increasing efficiencies by using the Pinewood products.

On Slide 16, we set out our progress in our priority growth markets outside the U.K. and the U.S.

We set out some key geographies in our strategy at our Capital Markets Day in 2024. Japan and Southeast Asia, Central Europe and South Africa.

The Porsche Japan rollout started well in December 2025, and we are working well with the Japanese Porsche dealers to work through the rest of the implementations during 2026. Once the Porsche dealers are completed, we will move to the Volkswagen dealers in Japan.

We have ongoing discussions with a number of European customers, most of which are based in Central Europe. We have fully integrated our South African business into the group following the buyout of our South African reseller in mid-2025, and we are looking at different ways to grow our revenue streams in South Africa.

We bought out our final reseller in the Netherlands in February 2026, and we're currently in the process of integrating the team into Pinewood. To wrap things up on the operating highlights, I just want to repeat some of the characteristics that make our market so compelling.

All car dealerships need not only a DMS, but also layered apps associated with it. At Pinewood, we offer all of this, and we are in a position to offer car retailers all the software they need.

Our average customer tenure is 15 to 20 years is a testament to how embedded our system is within our customers' work streams. Switching DMS is a long process and works in our favor for existing customers.

But for potential new customers, we offer something no one else does. Being part of an auto retail group for over 25 years has given us a competitive advantage over everyone else, something that no amount of money or investment can replicate.

In addition, our now fully embedded AI features differentiated us from everyone else. We are in a unique position to help our valued customers navigate a time of increasing demanding environment from a security and compliance viewpoint.

Pinewood.AI offers something no one else is capable of providing. Finally, on to Slide 19.

In the U.K., we will continue to work closely with the Lookers team to manage the rollout through completion in Q4. We are also looking forward to starting the Marshalls rollout later this year.

As we've talked through, we have made significant progress in North America on both OEM integrations and on testing the system in the U.S. The opportunity in North America is considerably larger than any other part of the world.

So scaling here as quickly as possible remains a top priority for us. Finally, our FY '28 guidance is unchanged at an underlying EBITDA of GBP 58 million to GBP 62 million, which is underpinned by our strong visibility from existing contracts, including the $6 million contract with Lithia North America as well as a number of other signed contracts.

Thank you, all members of the Pinewood team for driving us forward as a group. Thank you for joining us today.

We welcome any questions.

Can you hear me?

Yes.

I've got sort of 3 different questions. The first one relates to -- I think you said Lithia is going to sort of go live in terms of revenue at some point in 2026.

So I just wanted to understand exactly how far through the capital investment being made to make that product sort of fit for market in North America we are and how we expect sort of CapEx to move over the next 2 or 3 years? And are there any tools internally you're using that's improving your development efficiency?

The second question is that broadly my understanding is the U.S. is built on a per rooftop basis.

in the U.K., there's a larger per user element. So I was just wondering what the rough split was for the Lookers and Marshalls deal between per rooftop and per user.

And then lastly, I noticed that there was a court case. It wasn't involving you guys, but Asbury and CDK went to court about their switch to Tekion, and they won that court case.

Tekion and Asbury won that court case. Do you think that precedent will make it easier for you to win share in the U.S.

in the future from the sort of incumbent North American suppliers?

So Ollie, I'll kind of go in reverse order. So we really can't comment on somebody else's lawsuits and stuff like that.

What I can tell you about that is -- and obviously, if you would assume the ruling is -- they're going to go at it again. So they're going to appeal it.

But at the end of the day, the dealer owns their data. We manage our dealerships that way as well.

So in our case, dealerships data is their data to do what they want, and we allow them to facilitate whatever they need to do on that piece. As far as being able to open it up, just because you have open access to the data that doesn't necessarily mean it's in an easily digestible way to move over.

We have created technology more specifically through our AI offerings to be able to facilitate that real time, whether we're pulling from a dealer group's data cloud on their own data stack or directly from a different core system DMS provider. But I do think going forward, you're going to see more and more open data stacks and as such.

On the pricing models, obviously, it varies greatly by which part of the world you're in. Europe still goes primarily off of SaaS-based pricing.

All of our pricing is going to be going to a rooftop and modular basis over time. And there's multiple positive reasons for that.

Some things are just specific to a brand, and there's no way to actually charge it out on a per user basis, OEM integrations, certain updates, certain facets of it. A lot of AI functionality wouldn't be charged in a typical SaaS pricing.

So it would have to be a rooftop charge. And one of the key metrics that are key attributes that we bring into the marketplace is being able to help dealers reduce their cost.

And oftentimes, that is by automating certain non-revenue-generating noncustomer-facing processes, and it is such lowering a per user count. And it would be counterintuitive to sit here and continue with a per user basis when one of your key tenets is to try to reduce the number of users that exist that way.

As far as -- we're not disclosing the exact amount that we spend in one market for development. But I'll just keep in mind that we are on a single code base.

And oftentimes, things we do for one market transition and go into another market. and/or it can just improve the current functionality to the system that exists today.

So we've got some great accounting platforms and different things that we've used -- that we found up in the Nordic countries and that we use in different places of the world, dynamic accounting and as such. There's lots of different tech solutions in North America and the ways of doing business that we could -- we would think would help other parts of the world, and we'd like to bring that over.

Obviously, going into North America is a big endeavor. So while our CapEx has grown to us, we've just replatformed over the last 24 months.

We've now put hyperscaling in. We've got all the certifications in progress with the U.S.

OEMs. That would be something specifically that would represent North America.

But say that piece, I think our CapEx is going to be pretty much where it is with the exception of being opportunities to go into new markets and/or investments into new and more dynamic products as we go forward.

Sorry, Peel Hunt is asking too many questions probably. The thing I wanted to ask you, Bill, is that, obviously, I think you guys have done the right thing by making that Seez acquisition at the right time.

It's interesting to see Keyloop loop also making or trying to make a similar acquisition or has done so in Motortech.ai. And so obviously, everybody is kind of trying to move towards AI, embed AI.

I was listening to your Full Throttle podcast when you were over in the U.S. for the NADA conference.

In there, obviously, you were kind of reiterating the fact that you don't want to do what everybody else is doing, which is just putting chatbots on top and the AI has to be embedded deeply. And when you look at your product page, I saw there's a nice video about how Seez DNA is going to permeate through the full product stack.

The thing that slightly kind of confused me not listening to you, but listening to Jay at Tekion is in their NADA conference, it almost felt like he was playing down kind of Whizbang AI functionality, and he was trying to talk to the core of the single truth, the security layer, trying to move from -- trying to preserve the importance of system of record I kind of remembering, but I'm just trying to get to how you are trying to position Pinewood when you are talking to potential target opportunities when everybody is just talking AI, AI, AI all the time from Nextlane to even the constellation assets.

Yes. So listen, I can't really going to comment what any of our competitors would say out there.

I can tell you this, the one thing in automotive, whether it's on the retail side, the tech side or the OEM side, of things -- the best of ideas are often replicated. I will say this with Jay, I completely agree on system of record and the core data stack that goes along with that and the importance of that.

To -- the statement that you made maybe are different things here. I look at AI being able to utilize that core data stack and then being able to do new and exciting things with it.

So the one thing with a platform like ours is without having to have a plethora of layered apps and disjointed data stacks and disjointed functionality and having 20 windows open simultaneously and to the best state possible having the fewest number of non-OEM integrations to be able to facilitate a smooth and easy transaction for both the customer and the dealer, I think, is a real differentiator. Where I see AI coming into it is AI is going to be able to do things like a system of record, a core enterprise-level accounting platform.

Even as simple as ours is to operate, they still could be very complex. And we've already embedded our AI into that.

And now the AI can answer questions for a user and be able to show you how to use the system. Where I see -- I'm starting to use AI is being able to do predictive indexing, being able to help a customer be able to schedule an appointment to maybe get the car service, but be able to do it in a way where not only do we -- does the shop have the availability on the side of individual writing up the service adviser and as such, but also being able to make sure that we have a technician that's available that has the appropriate skill set to work on that car.

Oftentimes, appointments are made because there's an hour open in the day, but that technician might not have the qualifications to work on an electric vehicle, for example. I'm starting to do things like that.

Do predictive indexing to engage with the customer when their car is going to need their next servicing or being able to directly when it comes to recalls and as such without having to use phone rooms and operators and humans to see here and try to facilitate things like that. I can see it being able to sit here and assist the customer journey.

But at the end of the day, that core data stack, that system of record, that enterprise-level accounting platform with full integrations into the OEMs is the nucleus for everything. And any AI is only going to be as good as the data that it has access to.

And to your earlier point, that's why I'm not quite as big of a fan of the layered apps approach and sitting on top like a chatbot because at the end of the day, they're pretty similar and they're accessing the same data streams when you have a data set like we have and the plethora of information that goes along with that, it really changes the operations and the value proposition that AI can bring. But AI is going to be additive.

I do think it will be transformative. And I think you're going to be pretty wow by what's going to come out over the next 12 to 18 months when it represents to that with Pinewood and Seez.

I just have a follow-up question. So obviously, your biggest opportunity is the U.S., but I'm also really interested in the European opportunity you have.

I think you've always talked about the fragmented nature of the European market. There are a bunch of products that are coming to end of life.

But now as you described, AI is actually making use of these tools a little bit more easier. So the learning curve of adopting a new piece of software might be easier.

Then on top in the last couple of months, obviously, people are seeing scary headlines about cybersecurity. So do you feel like the ability for you to probably run harder into the European market is now stronger than it was 12 months ago.

I mean you obviously bought in Netherlands the Dutch distributor. And I know you have hiring plans in place from Germany.

Just some commentary around kind of competitive landscape and if it's become easier in your head to crack the European opportunity?

Well, I don't know if I'd say easier, but the demand coming out of Europe, Asia, Africa and South America is growing exponentially day by day. OEMs are looking to go about things in a different way.

Several franchises that -- sorry, several OEMs that we've engaged with have talked where they have upwards of 200 different OEM integrations, and it's just impossible to manage all those effectively. OEMs want fewer integrations into DMS systems, and they want more advanced systems that are out there.

Cloud-based systems like ours are definitely appealing to them. So they want fewer.

They want people that can sit here and operate on a worldwide basis. And to your point about -- and we call it assist that's built into our system where you can really just ask the system how to operate itself is a big differentiator.

Our system is also language agnostic. So you can work multiple languages within a single dealer group even within a single store.

So yes, I do think that. I think into the example that you bring, the European market is very fragmented.

You have lots of local players in there that solutions are built specifically for a geography, and I just don't think those are going to stand the test of time. And I think you're going to need -- kind of like you've seen with consolidation within the retail network within dealers, I think you're going to continue to see consolidation within the tech stacks that the OEMs are going to be willing to integrate with.

And ultimately, I think that will help drive opportunities for platforms like Pinewood.AI to be able to engage into there. And more specifically within Europe, you're right, there are several burning platforms, end-of-life, unsupported systems out there.

And I think that gives us a huge opportunity. Now that said, North America, by far, is the biggest automotive retail market, especially on the technology side.

So it's going to be a key focus, but we're never going to forget our roots and the geographies we already operate in. And in a perfect world, we want to continue to grow in our key markets close to our head office, the opportunities in North America, but then to be able to continue to grow in Africa and Europe, Asia and maybe even with the timing right, maybe even into South America.

Well done on what you've achieved over the last 2 years. It's great to see a vertical software actually company thrive in the age of AI.

Cheers.

I appreciate you saying that. The team has done an absolutely wonderful job.

So Ollie and I get to stand upfront, but there's 400 people that are doing the heavy lifting. So I appreciate the kind words.

A quick one from me. You talked about sort of competitive moat versus generic AIs in your RNS there.

Just be interested if you could flesh that in particular, you sort of talked about the benefits of 20-plus years of experience and you talk about the challenges of integrating with the OEMs. So if you could just talk a bit more about why that is a challenge and perhaps reference some of the -- some of what you had to go through in the U.S.

to get it done to the extent you have out there.

Yes. So Andy, so it's a good question.

So the weird thing is you would think that if you integrate with a certain OEM that, that integration would work anywhere in the world. And what you find is that's the farthest thing from the truth.

You can be in parts of Central Europe and have 4 countries touching each other and have 5 different integration levels with a single OEM. And so what you find is like going into North America, we're going to probably end up running nearly 1,000 different integrations to get all of the key OEMs fully integrated.

Now lots of them have multiple integrations. So it's not 1,000 different OEMs, obviously, but it can be quite complex.

And there is a huge differentiation even in some cases, between Canada and the U.S. So it's not the difficulty of doing it.

It's just the sheer volume of work. Currently, we're pacing to have by the end of this year, 90-plus percent of the volumes that exist within North America have to be fully integrated on the OEM side and already have the certifications.

And that really opens up the pathway for us to be able to expand outside of the Lithia deal and add additional U.S. dealers on in a relatively short period of time.

You talked about the moats, and I'm glad you brought that up because obviously, everyone saw kind of the SaaSpocalypse on February 4 with Anthropic and Cloud coming out there and kind of disrupting some of the legal platforms that exist out there. A platform like ours puts us in a really unique position.

You talked about the OEM integrations. That's one of those things where as big of an advocate I am in the company is for AI, I don't see how AI is going to be able to sit here and build OEM integrations.

If it's 1,000 integrations for the U.S. to replicate that one worldwide.

I also don't see the OEMs open sourcing or letting AI agents come in and start doing integration work. Having an enterprise-level accounting platform once again that has the integrations with the OEMs that is multi-jurisdictional in 40-plus countries to be able to do that way.

These are things that at least in the near to midterm future, there are just things that an agent, an AI agent just isn't going to be able to do. And we have quite a few moats that kind of work around us and protect us on that piece of it.

But there are also competitive advantages as we engage with new potential customers as we continue to work with our OEM partners, and we continue to expand our footprint worldwide.

Two questions from me. First one, obviously, some really useful new disclosures to help us understand how much of that GBP 60 million target is already underpinned.

I guess my question is around capacity and what -- how much Pinewood has around current and future implementations. So on one side, has the delay to Marshalls made things tighter in FY '27?

Or conversely, has that allowed you to get ahead on Lookers and in North America? And on the other side, there's obviously a lot more potential market share out there currently held by legacy peers.

Would the company have the implementation capacity in the event of a new contract win, say, in a new region where you need new OEM integrations?

So it's a good question. Not specifically within Marshalls or Lookers, our implementation team that's based in the U.K.

for -- at least for the U.K., we're able to flex our capacity and the ability to sit here and put additional dealer groups on. So while the delay in Marshalls doesn't really change that much, might free up a little bit of resourcing to sit here and go a little bit quicker with the Lookers.

But we have other contracts, obviously, besides Marshalls, Lookers going on. So it just opens up capacity on that end.

Normally, the biggest thing that limits the speed that we can do an implementation is normally on the customer side and working around their time lines. In the U.K., for example, you normally don't do integrations in March and September because they're big retail months.

So in that case, we might take some of that resourcing in the U.K., maybe move it into Central Europe to help with implementations on that site. I'd say the one thing as we go forward and we continue to grow our business and our capabilities and our tech stack is how we've started to use AI to help simplify the implementation.

So the biggest piece of implementation is actually transferring the first is laying out the accounting stack that the current group is working on. There are certain nuances to every group in the way they operate.

And while not every group has a bespoke platform that we help build for them, but then at the end of the day, there are nuances. But the biggest piece is extracting the data out of their current system or systems and then putting it in.

We've just now started to utilize AI to do something that used to take a couple of weeks to do to get it down to taking a couple of minutes to do. I talked to you about on the call at least about using Assist, our AI agent that's embedded in the platform that's able to help you operate the system where you can be asking a question, how do I stock in this car, and it will give you a detailed pathway to be able to go do that.

And in the relatively near future, that same capability, the system will be able to do it for you with a handful of prompts, be able to do it that way. So it really reduces the training time down significantly on new implementations as well as the way the system has been designed.

It's very intuitive. It's all menu driven.

And if you've used any Microsoft products in the past, and can read use a mouse and use a menu, you can operate our system in a pretty short period of time. So we're using our experience, our current tech stack, some of the innovative things we're able to do with the Seez tech stack and get it where implementations can move very, very quickly.

Great. Maybe one more, if you don't mind, perhaps, Ollie.

We obviously have clarity around the FY '28 EBITDA target, but maybe you could run through in a bit more detail your expectations for how cash margins trend through to '28 and the sort of CapEx and working capital dynamics around that. I appreciate you don't want to get too specific on the CapEx side.

No, no, that's right, Alex. Good to hear from you.

So yes, from a working capital point of view, I think we've touched on it before, as we grow our customer base, we do get an advantage from this because we invoice our customers quarterly in advance. This is for the recurring revenue and the majority of our revenue is recurring.

So we will get a benefit from that. So as we get into FY '27 and FY '28, there will be a significant working capital benefit.

So in FY '28, we're talking GBP 12 million, GBP 13 million of benefit there in that year. From a CapEx point of view, I think Bill touched on this.

The underlying CapEx or standard CapEx is very much being one code based as is. There is -- exiting this year, we have been doing the OEM integration work with the U.S., which we're in a good place with.

I think we've mentioned that we're anticipating having the majority of that done during 2026 or by the end of 2026, which is great, and it allows us to effectively sign any U.S. customers.

So we're in a good place with that. So they will once we annualize the CapEx, there will be a slight tick up for '26 and '27.

But the net impact of those movements is of the GBP 58 million to GBP 62 million underlying EBITDA, we expect sort of high 40s of that to drop through from a net cash position. So we're talking GBP 45 million to GBP 50 million of cash in FY '28.

Thanks, George. Yes, we've got 2 questions from Investec with regards to total contract value.

So Roger says, thanks for the total contract value metric disclosure. What's the assumption being made on average customer lifetime, please?

Presumably, this can end up being many years, if not decades. So what is the cutoff being used?

And then the second question on TCV is, broadly speaking, what's the subdivision of total contract value number between subscription maintenance type businesses and implementation/service business?

Yes. Thanks, Roger, for those questions.

Yes, just going in reverse order. So in terms of the split between recurring and implementation revenue, it's all recurring revenue.

So we don't put any implementation revenue into that figure. So it's from the contracts, the customers that we signed and they are not implemented.

It's just the recurring element of that revenue. And in terms of a cutoff or lifetime, the assumption is we haven't put a lifetime on it.

And the reason being for that is the only time we really lose a customer is generally through M&A activity. So if we've got a customer with, say, 5 dealerships and they get bought out by someone else with 50 dealers who are on a different competitor system, that tends to be the only time that we lose a customer.

So we're working on the assumption that most of these customers in the total contract value are big enterprise size level customers, the majority. So the assumption is that they're sort of lifetime customers and there isn't a cutoff.

We haven't put a 10- or 15-year lifespan on. Hopefully, that makes sense and answers the questions.

I think that's it, Bill. So good to...

Perfect. Listen, thanks, everybody.

Kind of like I said earlier, when I was talking to Andy, I just want to give a reminder that even though Ollie and I could sit here and take the lead here, this is all a testament to the hard work of the 400-plus employees that we have worldwide. And this is probably the oldest start-up ever, and we look forward to even better performance in the half years and the years to come.

Thanks, everybody.