Well, good morning, everyone, and thank you for joining us this morning. We're here today to mark a definitive shift in the RWS story. Over the next hour, we're going to walk you through how we are accelerating our evolution from a service-led company into a technology-first future. This isn't just a plan. It's a shift that's already underway. I'm Ben Faes, the CEO of RWS, and joining me today is Candida Davies, our CFO, who's been instrumental in navigating this year of change. And I'm also pleased to have Christina Scott here, our Chief Product and Technology Officer, who's driving the innovation that you will hear about. In terms of our agenda, Candy will kick us in with the financial details for FY25. I'll then return to deep dive into the new growth strategy we've put in place, including a closer look from Christina at our product technology roadmap, which is central to this pivot, and finally, we'll wrap up with our outlook and guidance for FY26 and beyond. Before I hand over to Candy, let me set the context for the journey that we've been on, particularly since we last spoke in June. Looking back, FY25 was a definitive inflection point for RWS. While we navigated a complex environment, the headline here is our trajectory. We saw a strong exit rate and a distinct positive shift in dynamic during the second half. We haven't stood still. We fundamentally reshaped the company to be fit for the future. First, we refocused our strategy. We redesigned our operating model to be more agile, accountable, and efficient. This new structure is not just a concept. It's already been fully operational since the start of FY26 in October. We've also injected fresh energy into the leadership team with new talents and successfully relaunched our brand to reflect on new direction. Crucially, we see strong momentum in AI services, a trend that we see accelerated into FY26. Our acquisition of Papercup is fully integrated, giving us leading capabilities in the vital AI for video space. To win in this fast-paced tech environment, you cannot work in isolation. I'm proud of the deep collaboration that we've established with major tech players. Partnerships that really allow us to build, to develop, and to distribute superior products. Finally, this is really key. Our SaaS revenue is continuing to climb. It now accounts for 46% of our license revenue, up from 39% last year. This shift is vital because it builds a predictable recurring revenue base, and Christina will show you shortly how our robust product pipeline is fueling this transition. Looking ahead, FY26 is all about accelerating that momentum and demonstrating a clear return to profitable growth. Our focus is simple and threefold. We're targeting high growth opportunity. We are executing an ambitious product innovation roadmap, and we are driving efficiencies to fuel this transformation. To put it simply, we're focusing our sales effort on the most attractive market. We are building products that our clients need, and we're running a leaner and smarter organization. With that context, I will hand over to Candy to walk you through FY25. Thank you. Thanks, Ben, so let me take you through the financial performance for the year ended 30th September 2025. Overall, it was a challenging year financially, but one where management took decisive action on cost control while also focusing on building a refreshed operating model designed to lead the shift to a technology-first AI solutions partner. As you can see, we reported revenue of GBP 690 million, reflecting an organic constant currency decline of 0.7%. Gross margin declined year on year by 350 basis points to 43.4%, primarily impacted by adverse mix. I'll give you some more color on that shortly. Adjusted profit before tax was GBP 60.4 million, with a strong delivery of GBP 42 million in the second half, following GBP 18 million in the first half. This translates to an adjusted basic EPS of 12.1 pence. Adjusted EBITDA was GBP 101 million and adjusted EBITDA margin 14.6%, down 500 basis points on the prior year. Capital expenditure was GBP 26 million versus GBP 46 million last year, which reflects both lower spend related to the transformation initiatives and the change in the capitalization treatment of some of our technology R&D. Operational free cash flow was GBP 80 million, a 46% increase year on year, supported by improvement in our working capital as well as the lower CapEx. Finally, I can confirm that the board is recommending a final dividend of GBP 0.046, bringing the total dividend per share for the year to GBP 0.0705, following a reset of the dividend in line with the adjusted profit performance in the year. As I'm reporting on FY25 results, this slide shows the legacy structure. Ben will talk to the new segments and their performance later in the presentation. The FY24 OCC revenue adjusts FY24 reported revenues by subtracting the predisposal PAT base revenue and restating these numbers at FY25 FX. Year on year in organic constant currency, the FY25 revenue at the group level was broadly flat, with repeat revenues stable at 95%, a slight headwind from price, and promising new logo wins. From a divisional standpoint, IP services revenue was flat on a constant currency basis, with strong growth in renewals offset by the decrease in Eurofile on the back of lower grant applications. Language services grew 3% in constant currency, with our data services AI-led proposition TrainAI performing strongly within enterprise clients, and we continue to win new business across the division. From a core localization perspective, we saw clients grow more confident in machine-first translation for non-specialized services. In regulated industries, the OCC decline of 10% is primarily driven by our linguistic validation business due to a reduced number of projects and client delays, as well as some softer trading conditions and spending cuts in both life sciences and finance and legal. Fresh management and the new operating model, which places RI within the broader Transform segment, was put in place from October 1st, is expected to resolve some internal operational issues whilst also moving forward on a leaner cost base. Finally, our language and content tech division was flat year on year in organic constant currency terms. SaaS revenues grew 14% versus prior year, and as Ben's mentioned, this now represents 46% of license revenue versus the 39% last year. Moving to gross margin, which declined 350 basis points year on year to 43.4%, driven largely by some unfavorable mixed dynamics. I'll highlight the four main areas. Firstly, we have a number of new growing services which are currently lower margin, such as renewals and IP services, TrainAI, and the geographic area of APAC within language services. As these areas grow, we expect to see margins improve through both automation and economies of scale. Secondly, as mentioned at the half year, we experienced some teething issues as a couple of larger clients moved to newer automated delivery models. The majority of these issues have now been resolved through tech and process improvements and, to some degree, increased pricing. We also continue to drive SaaS revenues to secure recurring revenue that come with slightly lower margins, and finally, this year, we also experienced a lower-than-expected revenue in the higher margin areas of Eurofile within IP services and linguistic validation in regulated industries. Looking at price, price was a tailwind in our language and content tech division via increases through renewals. We did experience modest downward price pressure in our translation services as clients expect continued efficiencies from AI and automation. But you can see that in-year efficiency supported 150 basis points margin expansion, which more than offset this pressure. As machine translation consumption increases, we continue to adapt our processes, tooling, and overall cost structure to drive the benefit of tech first and protect and grow our gross margin. As mentioned briefly on the previous slide, we're actively pursuing operational and functional efficiency through a combination of process improvements, automation, and simplification, leading to headcount offshoring and reduction. The total FTE at the end of FY25 was just over 7,600, a reduction of over 400 or 6% versus a year ago. Please note these figures exclude contractors, freelancers, and other casual workers. This, alongside continued offshoring, equates to annualized overhead savings of about GBP 14 million, 50% of which was realized in the year. As a result, revenue per FTE increased 5% in FY25 to 90K. So moving to the adjusted profit before tax bridge, where the waterfall summarizes the main impacts year on year. And firstly, looking at the trading items, where the decline in volume mix and price has been fully offset by the efficiency initiatives. These include the 150 basis points seen in gross margin from headcount reduction, offshoring, and other non-comp and ben initiatives, the reduction in overhead headcount and offshoring, and some additional one-off items such as reduced sales commission and other discretionary cost control measures. And then the non-trading items for GBP 32 million. There was a GBP 10 million overall impact from increased amortization related to historic investments and the change in the capitalization treatment of technology R&D. While gross R&D investment levels remained broadly flat year on year, as explained at the half year, we have expensed more technology investment this year rather than capitalizing it, and the offsetting impact is seen in the CapEx number, and overall, FX impact in the year was GBP 20 million adverse, with the strengthening of the pound versus the dollar. The FY25 average of the cable was 131 versus 127 last year, and the euro 118 versus 117. Adjusting items totaled GBP 160 million in FY25. This is primarily due to a non-cash goodwill impairment charge of GBP 88 million, relating to our core localization services within both RI and language services and the weaker performance of linguistic validation. It also reflects the shift in direction towards a technology-first offering and more technically, the increase in WACC rates. Exceptional items are mostly restructuring and integration costs, enabling the new strategic direction and leaner operating model. And finally, acquisition costs in FY25, which were primarily the deferred consideration for PropelOn acquisition. Moving to cash, the business remains strongly cash generative with operational free cash flow of GBP 80 million, up GBP 25 million year on year. This increase reflects working capital improvements and tighter CapEx discipline, more than offsetting the decline in adjusted EBITDA. The group had a net debt position of GBP 25 million at the 30th of September, an increase of debt of GBP 12 million versus a year ago. Working capital focus delivered a GBP 13 million improvement driven by enhanced receivables collection and payables optimization and the increase in deferred revenue as SaaS licenses grow. Capital expenditure reduced this year to 26 million of revenue, as I said, reflecting both the reduction in the spend in internal transformation initiatives and the change in capitalization treatment, and the acquisition exceptionals and other includes primarily exceptional items and acquisition costs as discussed in the previous slide, whilst the dividends reflect the interim dividend for FY25 and the final dividend for FY24. Finally, I'm happy to confirm that we completed an amend and extend of our RCF in early October, extending the facility to $285 million through to October 2029, and with that, I'll pass back to Ben. Thank you. Right, so turning now to our growth strategy, I want to share the vision we have for RWS and the specific pillars that we are executing to get there. RWS has evolved significantly over the last few years, and it's a shift that isn't always obvious when you look at the numbers quarter on quarter, but with the strategy we're presenting today, we are accelerating that evolution to become a true technology-first leader. The best way to understand this evolution is to look at our revenue mix. Today, more than a quarter of our revenue is already derived from AI-related products and services, whether that's from data services to fine-tune client models or AI workflows to adjust the output of AI. Sorry. We expect this to become our dominant revenue stream in the future, exceeding 50% of our revenue share. We see the same trajectory in SaaS. We've moved rapidly from 24% to 46% today of our license revenue, and we project this will climb to over 75%. To give you a sense of scale, this year, our technology processed 1 trillion words. That's roughly the equivalent of half the text of all of the public internet out there. And to do that manually, to give a sense also of that scale, you would need about 1 million linguists working full year, full-time for a year to do this. Yet even 1 trillion words is a fraction of the demand that we see coming. We see a clear path to our technology translating over 5 trillion words in the future. Conversely, traditional localization, where humans handle every step, used to be our largest segment. Today, it's less than a third, and we expect this to settle just below 20%. Now, we do see a natural floor there. There is a stable baseline of regulated work that will always require expert human validation and this creates a natural high-value floor for this segment. Plus, there are clients who actually restrict our use of machine translation for policy reasons and there's also a long tail of rare languages where AI isn't quite ready. The direction of travel is inevitable, but we see this floor, so why is this shift happening and it all starts with our client. Their fundamental goals haven't changed. They still need to reach new markets, new audiences, to protect their IP, to ensure compliance, to do this in a safe and secure way. What has changed is the speed and scale required to do this. Speed is now really the currency. Content breaks daily and if localization lags, that content becomes irrelevant. Time to market is really everything. At the same time, the environment that they're operating in is exploding. We are past the tipping point where actually half of the internet is AI-generated. Content is becoming multimedia-first. It's hyper-personalized. It's hyper-personalized by audience and by channels like TikTok or Instagram. We are also seeing shifting in trade routes. We see, for example, a surge in demand from our Chinese customers who now export directly to other Asian markets. And that creates demand for new language pairs that were previously niche. And finally, what we hear from our clients is they're facing a flood of AI innovation. New models appear every week, and it is incredibly complex for a CMO or a CTO to know which tool to use. They're looking to us to help them navigate this use of AI efficiency without risking their brand reputation or their data security. And this is really what makes RWS a strategic partner rather than just a service provider. That brings us to our vision. Fundamentally, our mission at RWS is to build the cultural intelligence layer to enterprise AI. Let me break this down. We know that very soon, already today, our lives will be surrounded with AI-powered experiences. Imagine your alarm clock that briefs you in the morning, your autonomous car that drives you to work. Probably very soon, your robots that will make your dinner. Can't wait for that. But for those experiences to be truly adopted, for them to be wonderful rather than just functional, they need to adapt to who we are. They need to understand our culture, our nuance, and the context they're operating in. The current enterprise AI focuses heavily on maximizing the IQ. They want to make large language models as smart as possible. But there is a deficit, especially as soon as you move to a non-English market or to a complex cultural context, that intelligence drops off. Our mission, to put it simply, is to bring the EQ, the emotional intelligence, to that tech stack by solving three critical deficits in particular. Those three critical deficits that hold AI back today are, sorry, first, the data deficit. There is a gap between generic raw data and the high-quality domain-specific data needed to train AI for mission-critical tasks. Second, the culture deficit. Generic models today struggle to build genuine relationships. And we bridge the gap between robotic output and culturally resonant communication. And third, the trust deficit. Speed is useless without safety. And we close the gap between automated speed and the expert validation required for regulated industry and IP protection. If AI is going to play a central role in our lives, it must understand how we are, how we behave, how we communicate, and this is an immense opportunity ahead of us. To capture this opportunity, our strategy is simple and focused on three growth pillars. First, a new go-to-market where we're focusing our firepower on large enterprise and high-growth verticals. An ambitious innovation roadmap where we're really building that cultural intelligence product at scale, blending our tech with human expertise, and operational excellence. We're implementing an efficiency plan to deliver a faster, leaner experience to our client. Let's take each pillar in turn, starting with our go-to-market. As I signaled during our interim result in the summer, we've moved this strategy now into full execution, and we're shifting from transactional sales to solution selling, where we're integrating directly our technology into our client tech stack. That creates stickiness, scalability, and recurring relationships. We're also being much more selective, and we're focusing on industries that need us the most. Those industries that are global by nature, that are highly regulated, or that are managing high-value brands. We do this from a position of strength. RWS already works with half of the Fortune 500 companies, including the world's largest tech company out there, and they trust us because of our global footprint and our heritage of quality. To leverage this, we've reshaped our sales team. We now have a dedicated focus on nurturing our most valuable existing clients and hunting for new strategic logos. We're expanding also our reach through partners, and a prime example of this partner-led approach is Microsoft. I'm delighted to share that RWS is one of the very first partners integrated directly into Copilot. At their Ignite conference a month ago, less than a month ago, Microsoft outlined a vision where Copilot becomes the orchestrator of the enterprise, the single entry point for all the tasks that you need. So instead of toggling between different applications, a user simply asks Copilot to get the job done. And I'm proud that all localization technology was part of that launch. This means that if you're a Trados or a Language Weaver customer, you can now access all technology directly within Copilot. You don't have to leave your workflow, and you find right there a secure, approved translation tool. It is right there using your company's tone of voice and your company's glossary. Now, of course, with over 500 million professionals using Microsoft 365, this is a formidable distribution channel for us. As of today, we're the only, the sole third-party translation agent, including in Copilot. That, for me, stands really as a powerful testament to the quality of our product, but also to the agility of our tech team. And on that note, I'll hand over to Christina to take you deeper onto our product roadmap. There you go. Thank you. Good morning. I'm Christina Scott, and I joined RWS as Chief Product and Technology Officer in May. I've been in technology my whole career, and my passion is really driving business growth and value through technology and building fantastic products for customers. I've had the privilege of building world-class digital products at companies like the BBC, Financial Times, and Pearson. I've used AI and automation to drive efficiencies in companies like News Corporation and OVO Energy, and the latter whilst doing a very complex transition away from legacy technology for five million customers. And I also spent 18 months doing an AI fertility startup, which had the challenge of AI in a regulated industry. And I spent a lot of time speaking to my counterparts in different industries, and there's some very key trends that everyone's talking about. Gen AI has driven the explosion of content, the explosion of different formats. It's never been easier to create content, but actually, it's also driven a change in consumer behavior, where and how consumers look at that content. We've had the hype curve. Everyone's talking about AI, and people are beginning to come down into the trough. CEOs and boards are really thinking, how do we drive the investment from AI, and how do we get those efficiency gains that we were promised? At an AI roundtable I went to recently, it seems to be a consensus that actually the efficiency gains will come from domain-specific use cases as opposed to rolling out generic AI capabilities, and accountability is becoming more and more important for customers. It's no longer good enough to say you're not responsible for the AI that you're using. Last year, we saw Air Canada was ruled against by a judge for saying that they weren't responsible for a rogue chatbot on their website, and finally, the thing that people are really grappling with is it's not actually creating more content. It's how do you stay globally relevant? How do you connect with your customers in an environment that is changing so rapidly in technology, geopolitically, and in a regulatory landscape? I've also had the privilege of speaking to many of our customers since joining, and they're giving me some very clear messages. They want partners to help them for the long term and to make sense of the AI journey that they're going on. They want to launch faster, quicker. I spoke to a hotel group last week, and they're using our capabilities in Trados to launch new websites. They launched in Japanese and Korean. And they have no hotels in those regions, but they saw actually a huge growth in sales, meaning that actually experiences that resonate culturally are really important for buying experiences. LLMs have promised scalability in content creation and localization, removing humans, but actually, there's a lot of ongoing investment you need to do, and it's not trivial, and it's not core to a lot of our customers. There's a real shift of focus to cost, so cost per word translated, to actually what is the business outcome you can drive? What is the customer conversion? How do you engage more with your customers? And technology teams like mine have been building systems for years where actually we've made you learn the technology. But with the new generation of AI, it's more human-like, and actually, it means it's now more accessible to non-experts. So actually, the content creators and not the localization teams can do the content transformation. And all companies are having to think about how they verify and explain their AI output. It's particularly pertinent for our customers in regulated industries. So if you're producing instruction manuals for hazardous equipment, you need to make sure that it's not only correct but understandable in all the different languages you're producing it in, and it's not excusable to say that something is lost in translation. Listening to our customers and understanding what's happening in the market means that I've come up with a number of very strong design principles upon which we're building our new products. We need to be embedded where our customers work, so Ben talked about Copilot, actually move and stop making people flick between lots of different capabilities. We need to use technology to deliver what it can, and then humans become the value add on top, removing the mundane from the human workflows. We need to make sure that we are orchestrating those technology and human agents in flexible workflows, and in doing that, we're really focusing on the business outcomes for our customers. It's no longer one size fits all. We need to make sure that we're truly multimodal and working across all content types. And as we build out, we need to make sure that we remain open and agile and that we are looking at, as the technology changes, we are able to partner with the right companies and make sure that we avoid any vendor lock-in. And security is definitely a non-negotiable, and verification has to be built into the process, whether by machine or human. I'm delighted to say that we're already building our next generation of products using the design principles that I've just mentioned. We have a lot of key strengths at RWS, and so we're not starting from scratch. I'm taking the assets that we already have, and so we have years of investment in our language AI and including our AI dubbing acquisition. We have a fantastic expertise in our linguists and networks of experts. We have a real understanding of our customers and their domains and the challenges that they're working with. We have experience of creating and managing the content needed, excuse me, to fine-tune the AI, and we have a lot of experience at managing complex workflows on behalf of our customers. And security and compliance, we build in by design. So we're taking our RWS assets, and we're looking at the new capabilities that AI is now producing, the LLMs, the agents, the agentic workflows, and working with some of our key tech partners. We're creating our AI intelligence layer. So we are taking our existing assets and our AI intelligence, and this really builds in the efficiency, not just replicating what humans do, but going beyond human capability. These capabilities will then be surfaced in our products that are tuned for domain-specific use cases. In Generate, we're bringing an AI-native enterprise knowledge platform that empowers knowledge workers to research, draft, and manage change with accuracy, auditability, and governance at scale. For Transform, it's an AI platform that's delivering the future of multimedia content that really drives growth for our customers, the cultural intelligence that Ben mentioned earlier. And for Protect, it's an AI-powered platform that improves IP attorney productivity and makes sure that we are automating the critical practice management functions for them. And again, we will surface these capabilities in products where our customers are working, whether that's in existing workflows, in our own workflows that we're producing, or in places like Copilot. And we're building this in a slightly different way. With me joining, we've brought all the technology teams together so we can build out this AI intelligence layer once and surface it through all the products. We're working really closely with customers, getting feedback as we're building out. And we're working in a fast and agile way, so these products will be launching in 2026, and then we'll be evolving from then. Sorry, I forgot to click. Okay. And so to give you some examples, why is this different? So we talk about scalability, efficiency, but actually, let me give you a few examples we're talking to customers about. With one customer, they want to launch a product to market, and it takes three to six months to currently localize. We're trying to do it in a day. With another customer, they have an uber spreadsheet where they capture all the latest compliance regulatory requirements for their products, and they have to manually update all their product manuals. We want to automate that. And so it's really delivering value to customers and the customer needs that we talked about earlier. So finally, why can RWS win, and how are we going to win in this space? We have a long heritage ourselves, but we're also very keen to look at the right partners. RWS have been building linguistic AI since 2017. We have a significant number of patents. I think we got a new one yesterday. We're constantly winning awards. And so we have that capability, but LLMs have come in, and they have a huge promise of the scale, the efficiency, but they need to be really tuned on appropriate data, and they need to be verified to be domain-specific. They're also really expensive to build and train. Cohere is a Canadian AI company. It's currently, I think, valued at around $7 billion, and they were founded in 2019 and have focused on LLM model building. They created their multilingual model in 2023, and we've been testing it for the last three months, and it beats competitors. So I'm thrilled to say that RWS and Cohere have agreed to an exclusive partnership by which we bring together Cohere's capability of model building with RWS's capability to do the fine-tuning, have the right data, and the linguistic expertise. And these capabilities together will be surfaced within our products, making sure that we're meeting our customer needs now and in the future, and the technology doesn't just create better products for our customers. It also drives efficiency that I believe Ben's going to speak about next. Thank you, so this brings us to our third and final growth pillar, strategic efficiency. The challenge our clients face today is straightforward. They only localize a fraction of their content in a subset of language, and this is because traditional costs are still a barrier to scale. Yet content volumes are exploding, and new market opportunities are opening up every day, so our goal is to break that constraint, and to do that, we're deploying a scalable technology-led delivery model that makes content expansion faster and more cost-effective for our clients. We aren't just cutting costs. We're removing that bottleneck to their growth. We're driving this efficiency to three specific levels. First, the rationalization. We're streamlining our systems, and we're migrating towards a digital AI-optimized operating system. Second, we're offshoring more. We're leveraging our global footprint more aggressively. We already have 20% of our colleagues in India alongside key hubs in the Czech Republic and Romania, and we'll continue to scale this center of excellence. And third, AI efficiency. This is about deployment. We're working in close partnership with two major tech companies to build agentic platforms for our production systems. And we're already seeing the results. We've recorded a 15% improvement in productivity in our developers through the deployment of AI coding tools. Globally, we expect these initiatives to drive a 10% productivity increase over the next 18 months. This creates fundamentally a faster, more valuable service for our clients while improving our margin. 2026 will be a year of acceleration. Before I share the financial guidance, I want to show you the operational KPIs that we are tracking to measure the success of this strategy. We're looking at this through the lens of our three pillars. In go-to-market, success means deepening our relationship. We aim to grow the net retention score with our top 100 clients to be greater than 100%. It means keeping all of them, cross-selling, growing those top 100 clients, which are so strategic. Crucially, we intend to achieve this while maintaining our exceptional customer satisfaction score and keeping it above the 46 that we are at today. For innovation, we'll measure the shift to tech-led revenue, and we expect our SaaS revenue to climb to 60%. For AI-related products and services, we expect this to grow to 40% of our mix. And for efficiency, we're tracking the revenue per employee, which we aim to increase to GBP 105,000. And ultimately, this flows down to the bottom line where we're targeting an EBITDA margin gradual improvement. This is really becoming cultural within the company, and this is a good measure of the efficiency of our work. So that brings us to the financial guidance. As we execute this strategy, transparency is key. And so we're moving from our legacy four-segment structure to three distinct pillars: Generate, Transform, and Protect. And these slides provide a clear map of that evolution, so you can see exactly how our historical reporting translates into this new simplified model. Before I dive into the number, at a high level, this is the dynamic of the group moving forward. Think of Generate and Protect as our high-growth engine. These divisions are operating in expanding markets and are primed for acceleration. We've isolated them with creating autonomy, and we want them to grow as fast as possible. Transform is the foundational core of the group. It is where we're pivoting and modernizing to shift from human-centric service to technology-led model. And underpinning all three is our group-wide efficiency program, which is optimizing our cost base to fuel this transition. Okay? So what does it look like for a number in FY 2026? In terms of revenue, we expect our new go-to-market strategy to return the group to low single-digit growth on an organic constant currency basis. This growth is driven by three distinct components. First, Generate. We expect good momentum here with growth in the mid-teens. This is our AI and data powerhouse. Protect, we see mid-single-digit growth. This business is robust. It's recurring. It's growing steadily. And we feel confident with our FY 2025 exit run rate. We see the benefit also of cross-selling our new product to our existing clients. Finally, in Transform, we expect a low to middle single-digit decline. And I want to be clear why. This is a managed consolidation. It reflects our strategic decision to step away from low-margin, non-core legacy work as we replace the traditional service with higher-value technology. Looking to the medium term, beyond 2026, we expect the group growth to accelerate. And the dynamic behind this is that in Transform, we will reach a crossover point where our technology-first offering and our new business lines with resilient pricing fully offset the decline of traditional services. For Generate and Protect, we expect them to continue powering the group forward with their underlying market and boost it further by our strategy of diversification across product and vertical. Now, turning to profitability and cash, the story here is one of gradual improvement in gross and operating profit margin and a normalization of our operational free cash flow conversion at a strong level, so for 2026, we expect to see immediate benefit of the operational changes we've already executed. We're targeting a gross margin expansion of about 150 basis points. We expect this to flow through to an adjusted operating margin expansion of approximately 100 basis points. Looking to the medium term, we expect this margin expansion to continue, driven by the three structural levers that unlock value as we grow. In Transform, the new technology platform and redesigned processes reach full velocity, and we will unlock economy of scale. We're moving from a linear service model to a scalable tech model, which inherently drives better margin. In Generate and Protect, we're modernizing the delivery platforms to drive efficiency. And also, as our newer high-growth business lines like TrainAI reach maturity, their contribution to the group profitability will increase. Finally, at the group level, we will continue to drive operational excellence, leveraging our offshore hubs and maintaining strict discipline on central cost. On cash, for FY 2026, we expect to deliver another year of strong cash conversion, aided again by our focus on working capital management and disciplined capital expenditure. In the medium term, as growth accelerates, we expect our free cash flow conversion to normalize to around 65%, which is the right profile for a growing company. And it reflects a healthy reinvestment in working capital and CapEx to support the revenue expansion that we've outlined. This brings me to our capital allocation policy. We are resetting the dividend this year, and I want to be clear why. While our cash generation remains robust, we're prioritizing the funding of this technology-led growth. We need the flexibility to deploy capital into investment opportunity as they arise. However, the policy itself has not changed. We're rebasing to a level that aligns with our strategy. But from this new base, we intend to maintain a progressive dividend policy going year on year. To summarize, let's look at where we stand today. We are two months into our new financial year, and we're already seeing solid trading momentum. The strategy is not just a plan. It is in motion, and it's bearing fruit. RWS has evolved into an AI solutions company, uniquely positioned to capitalize on the AI revolution. We're building the critical cultural intelligence layer for enterprise AI, and we're pivoting our business model to deliver improved, sustained quality of earnings. We're confident in the path ahead. Thank you, and we're now ready for some questions. Morning all. James Bayliss from Berenberg. Two questions from me, if I may. Ben, I was hoping you could share with us some of the internal and indeed external feedback you've had with regard to the new strategy you've taken RWS on. You gave us initial thoughts at the time of the launch in June from the internal side, but how have customers been responding to the new go-to-market strategy, and how have internal kind of conversations gone since we last caught up, and then my second question, just with regard to capital allocation and the fourth pillar with a return to acquisitions in due course, should we be thinking about those being focused on the AI side of the business, or would there be instances where we might see more professional services type acquisitions to complement your offering? Thanks. Thank you. So in terms of feedback, internally, I think the message has been quite phenomenal. And it's led by two things. One is the influx of new talents, new teams that really make this vision real. We've had our global sales kickoff last month, where the energy was really palpable, where people feel confident of the road ahead because they know that the market for this cultural intelligence layer is just immense. It's ours to grab. We are working a lot more closely with clients. We have a weekly product development update where Christina's team is really sharing all the progress of the new products. They're co-designed with design partners. We're working hand in hand with some trusted clients. The partnership that we've talked about with Cohere, for example, there's been more test partners than we could ever imagine. People want to be part of this innovation. So that really validates the road ahead. In terms of capital allocation, I really don't have any news to share today. We are looking at our assets, deciding which ones are core, which one may be not so much core. And there are plenty of opportunities. We're being very picky. Now that we have this clear vision, that Christina has a clear map of what we want to build, I think it's much easier to detect the company that will help us go into that direction. Morning, James Bayliss from Deutsche Numis. A couple of questions from me, please. Firstly, on TrainAI, I was just wondering if you could talk about the longer-run growth runway for that business. I think we can see that that side of the business has grown very strongly over the last couple of years. I sense a perception from investors that there's a question mark about whether there's a sort of the longer-run sort of growth opportunity for that business as large language models become sort of increasingly fluent and knowledgeable. And then second question, sort of slightly related to that, was within the Generate division, the division grew 22% organically in FY 2025, but you're targeting mid-teens growth rate in 2026. What is the driving factor there behind that sort of slightly slower expected rate of growth within Generate in FY 2026? I'll be frank. On TrainAI, when I first joined, I had the same doubt, saying, "This is a fad, and once the LLM are trained, what else is there to do?" This is not what we're seeing at all. We've moved from a generation of helping LLM understand data, understand context, so image annotation, things quite simple, to putting the best possible expert in the world. Imagine your thesis professor that is teaching the LLM specifically. We're reaching a point where LLMs are extremely intelligent, smart, but the influx of data is continuing to pour in, and so when we talk to those large LLM builders about this contextual AI, this is definitely something that is extremely hard to understand. You probably have seen some of the chatter recently about the ceiling of intelligence that those LLMs are reaching. We have a way as humans to understand context, to have a hierarchy of relations that's really hard to replicate in LLMs, and if we want really to reach this general intelligence, there's still a lot more help to bring in. Beyond that, what we see is also vertical-specific need. And beyond the core LLM builder, we have more and more demand by all kinds of clients that need to adapt AI to their own company, to their own way of working, to their own process. And this is a worry that's completely disappeared, at least for me. There's so much growth to achieve in the short to medium term that we want to focus on that. When it comes to Generate, this year has been quite extraordinary on TrainAI. It's been an exploding business. There has been a lot of competition. We've seen the new Llama model from Meta, the new Gemini, the new ChatGPT-5. And I think we want to be a bit cautious. The level of CapEx from large tech companies has been extraordinary. I don't want necessarily to bank that it's going to be exactly the same for the year forward. We think that they are going to be a bit more precise on what exactly they want to achieve and not just use brute force going forward. So that's why we think our expertise and special expertise and global presence is very relevant, not just brute force, but we want to be a bit cautious on the future. Do you want to take the mic? Oh, good idea. That one? Hello? Yep. So, starting with, in RWS's 2025 Translation Technology Insight Report, recently published, it was mentioned that market participants were seeing early signs of end clients coming back to more traditional service, having used LLMs. Can you elaborate on that, please? Yes. Or, Christina, do you want to take that? I mean, I think the slide that Christina has shown around the mix of AI technology to achieve localization between natural neural machine translation and LLM is interesting. LLMs have a much wider context window than the Neural Machine Translation that we used to have. So they end up having a localization that's more fluid, but it's less precise. And so you can put the same text twice in an LLM and get two translations. And for enterprise clients, this is a problem. So that's why I think there's been a lot of hesitation. What we're doing with Cohere is really trying to get the best of the Neural Machine Translation and the management of workflows that is really our heritage, but with the fluidity of the LLM. I don't know if you want to add anything. You answered it quite well, but yeah, no, it's time to fine-tuning and verification, but also if so LLMs aren't deterministic. So if you need a 100% guarantee that you get the same answer every time, the LLM currently isn't the right answer. Are you seeing that return to, sorry, go back to the mic. Are you seeing the return to traditional service more in certain verticals over others, or? I would say no because clients need to translate more. They don't have extensive budgets. And so we need to find more efficient solutions. So really, what we're focusing our work on is building the right guardrails for this new technology to be efficient for some clients. But as I said, the traditional translation has a natural flow. The work that we do in life science, we need to have, for example, by regulation, two certified doctors doing the translation. This is a regulation. And we live in a world where regulations tend to increase, not decrease. Thanks. Thank you. Dan Ridsdale from Edison. Could you just elaborate a bit on your partnership with Copilot, just in terms of how that's working economically and practically? So this is something that we've worked on with Microsoft, sorry, for quite some time. The system at the moment is quite simple. It might get bigger and more complex in the future. But today, if you are a customer of Trados, so you have the direct relationship with RWS and direct billing relationship with RWS, Trados is available and compatible within Copilot. So your company, all of the employees will access the localization tool from their Copilot in a secure way. So there's no shadow IT with people using whatever systems they want. And it uses your company tone of voice. So we all have our own lingo. There are built-in technologies, the dictionaries or glossary relative to companies. The beauty of it is also it's self-reinforcing. So the more you use it, the more the system learns also about making better translation. And in terms of sort of uptake, do you have any data around that? It's very early days. I mean, it's been launched, I think it was the 19th of November, the Ignite conference. So it's very, very early days. Yeah. Thank you. Sorry. What's the target balance sheet structure for the company, and what considerations have been given to buybacks? To what? Buybacks, repurchase. So I think I've explained the capital allocation policy. We've done buybacks in the past, which haven't proven very successful. But I think the company reserves the right to consider buybacks in the future. We have, in particular, a lot of work going at the moment on how to manage our long-term employee incentive plan. And that might be actually an opportunity for buybacks. From online, could you expand on the prospects for linguistic validation? Yes. So at a high level, linguistic validation is a key element in the process of developing a new drug. There's been a lot of flows in the development of drugs because of regulation, because of change of some policies. The reality is going forward, we see more drugs being developed, helped by AI again. But there's more and more projects of drugs being developed. So we see the market solid, probably growing. I've explained that in our half-year result. We've had some change, unfortunately, of management at RWS, which is creating a little bit of a softness. But this is an area where we are investing. We've welcomed three new salespeople, which are doing great. And so that's a segment where I personally spend a lot of time with the team because it's a very solid segment for traditional workflows. Thank you. Another one online. Candy, according to slide 11, the impact of FX on the Adjusted PBT was GBP 20 million. The question is, could you please clarify where the impact from FX is accounted for on the bridge on slide 13? So cash generative operations specifically, which bars does it fall in? Thank you for the question. I think you would find the majority of FX sat in the Adjusted PBT bar with a very small impact in net working capital. Sorry. Another one just come through online as well. Can the modern RWS exceed previous peak group profitability? Yes. I mean, that's definitely our ambition. So I think we're, I hope you've heard the message today that we are looking to return to profitable growth, expansion of margins. We've used the word gradual because we want to build on solid foundation as we modernize RWS, but this is both a market and an organization that's poised to have a great future. I think that's all the questions online as well. I think that's it. Thank you very much for joining us today, and we'll see you in six months.