Good morning, everybody. Thanks for joining us. I'm in New York, and I'm joined by Mary Basterfield, by Scott Spirit and Jean-Benoit Berty, who are all in London. So, we have a presentation on the first half. There are seven sections to it, it’s on our website. The first, Mary will give us a summary of the results. Scott will run us through market momentum, client analysis and artificial intelligence. And then I'll come back with a summary and outlook, and we'll have Q&A after that. There's an appendix also with certain other information that supports the presentation. So, with that, Mary, would you take us through the results, please?

Thank you, Martin. Good morning, and thank you for joining us today. I'd like to start with the financial highlights. Reported net revenue for the half year was down 16% to GBP376 million or 14% on a like-for-like basis. Global macroeconomic uncertainty, together with continued high interest rates, resulted in ongoing caution, especially among our large tech clients. Net revenue was also significantly impacted by expected lower transformation activity from one of our larger Technology Services clients. Given this lower net revenue in Technology Services, operational EBITDA was GBP30 million, down 8% like-for-like and 18% reported. We continue to exercise tight cost discipline, including headcount and discretionary costs. The number of months has reduced by 1,000 or 12% compared to this time last year and is now around 7,550. Operational EBITDA margin was 8%, in line with the first half last year, with significant margin improvements in Content and Data & Digital Media, offset by technology services. Adjusted operating profit was GBP25 million and adjusted earnings per share were 1.2 pence. Net debt at the end of June was GBP183 million and leverage was 2.2 times. Moving to the income statement. Revenue decreased 18% on a reported basis to GBP423 million and was 16% lower like-for-like in a challenging environment. Reported net revenue of GBP376 million decreased 16% or 14% like-for-like. Operating expenses of GBP343 million were down 16% on a reported basis and 14% lower like-for-like. This is the result of our action to reduce costs last year and during the first half. Operational EBITDA for the six months to June was GBP30 million, down 18% on a reported basis or 8% like-for-like, with like-for-like operational EBITDA margin improving by 50 basis points. I have given you a breakdown of adjusting items in the table on the left-hand side. Total adjusting items were GBP29 million, down from GBP37 million due to less cost for share-based payments and lower combination costs. Finally, the net finance expense was lower due to a positive impact from foreign exchange compared to a negative impact in the first half last year. This offsets a higher Euribor rate on the term loan, which provides us with secure long-term financing. Looking next at our three different practice areas

Content, Data & Digital Media and Technology Services. My comments here are all on a like-for-like basis. Net revenue in Content, our largest practice, declined by 9% to GBP234 million. It was a challenging first half with some larger tech clients remaining cautious and facing budget constraints. Data and Digital Media net revenue decreased 8% to GBP96 million, again reflecting market conditions with lower activity in the Media Activation business line, partly offset by growth in Media Agency of record. Technology Services net revenue of GBP46 million decreased as anticipated and was down 37% against a strong comparator. This was due to lower transformation activity from one key client in North America and longer sales cycles for new business. From a regional perspective, the Americas were impacted by this technology services client as well as lower spend from some tech clients. So, net revenue was down 15% despite strong growth in Latin America. The Americas remains our biggest region at 78% of the mix. EMEA declined 8% with lower activity in the Netherlands and the Middle East. And our smallest region, Asia Pacific, was down 9% with slower demand and less new business in India, offsetting robust growth in China. Moving now to operational EBITDA by practice, again, my comments are all on a like-for-like basis. We improved operational EBITDA in Content from GBP7 million last year to GBP16 million this year by taking action to reduce costs. This is also reflected in Content's operational EBITDA margin, which grew 450 basis points to 6.9%. We continue to exercise tight cost control as the market is expected to remain challenging, and we will reduce costs further in the second half to support EBITDA delivery. Data & Digital Media also benefited from actions taken in 2023 with operational EBITDA of GBP18 million. We improved margin by 330 basis points to 18.5%. Technology Services operational EBITDA decreased significantly to GBP6 million as a result of the reduction in net revenue and central costs were 27% lower at GBP10 million for the half due to continued cost control. Moving to the next slide. You can see that we continue to maintain a healthy balance sheet with sufficient liquidity and long-dated maturities. Our EUR375 million term loan matures in August 2028 and our EUR100 million revolving credit facility, which remains undrawn, matures in August 2026. We have now completed all material payments relating to previous year's M&A and currently have comfortable headroom against the key covenant. Moving to cash flow on the next slide. CapEx of GBP4 million is mainly investment in IT infrastructure. Interest paid reflects increased payments on the term loan, driven by the higher Euribor rate. We incurred GBP4 million of restructuring costs, mainly related to rationalization, and spent around GBP2 million on our ERP program. We continue to focus on working capital and receivables. There was a working capital inflow in the half year of GBP4 million. Net, this resulted in a free cash inflow of GBP3 million compared to an outflow in the first half last year. This takes net debt to GBP183 million. Turning now to guidance for the full year. We maintain our full-year profit guidance. We still expect overall operational EBITDA to be broadly similar to 2023 on a like-for-like basis as we continue to focus on our cost base. Given the outlook for Technology Services and wider market uncertainty, we expect overall like-for-like net revenue to be down on last year and to a greater extent than assumed at our last trading update in May. At a practice level, we expect an improvement in Content EBITDA and margin driven by cost reductions and Data & Digital Media to deliver a top-line performance similar to 2023 on a like-for-like basis with some margin improvement. The outlook for Technology Services remains challenging, and we anticipate both lower revenue and EBITDA following the reduction in activity with one key client and longer sales cycles for new business. We continue to maintain a disciplined and active approach to cost management as we right size the business for current market conditions, together with a focus on utilization, billing and pricing. We expect the year to be heavily weighted to the second half, supported by further cost savings. We anticipate a net finance cash charge of about GBP28 million and an effective tax rate between 30% and 32%. Finally, we maintain our full year net debt guidance of GBP150 million to GBP190 million. Thank you very much. With that, I will hand over to Scott.

Thanks, Mary. Good morning, everyone, and thank you for joining us this morning. If you look at our two main addressable markets, the digital media market remains quite healthy with a strong end to 2023 and a good start to 2024. Growth is expected to moderate over the course of 2024, but the major platforms will still grow double digits. Obviously, that contrasts with our own performance so far, but our revenues are aligned with their marketing spends rather than their revenue growth. And we're still seeing caution from our clients, particularly technology clients. For example, if you look at Meta's Q2 results, their revenues were up 24%, but their sales and marketing expenses were down 16%, and we see similar patterns at the other large tech companies. For tech services or digital transformation, this is a more difficult market, which began to slow in 2023 and has turned slightly negative in 2024. We actually had a strong 2023. So, the negative market trends have caught up with us in 2024, and it's also driven by the reduced spend with one large financial services client that Mary referred to. Most of our client categories were stable as a percentage of revenues in Q2 with technology still by far the main category, representing 44% of our revenue. We saw a decline in financial services, and that was largely driven by the decline in spend with that specific client in the tech services practice. We saw improvements in fashion and auto, which saw us gain traction in those two categories and the recent GM win should boost growth in auto going forward. The continued softness we're seeing in technology client spend and the specific large financial services client in tech services, which has reduced spend, have had a negative effect on the average revenue size of our top 20, 10, and 50 clients. Whilst we continue to see headwinds with some of our larger clients, this is primarily driven by reduced spending rather than any client losses. On the positive side, our progress in new business has started to help drive an increase in clients in the GBP1 million to GBP5 million category. Now, Wes, I'm afraid, can't be with us today. He's in client meetings. So, I'm going to step in and cover Artificial Intelligence update for you. We've been giving a recurring update on our AI progress and positioning. As you know, we were Adweek's inaugural AI Agency of the Year. And we have since won a series of AI business intelligence awards for our consulting practice, our proprietary technology solution, Monks.Flow, more on that later, and Wes' leadership in the field. Since we last met, we are also one of the few companies to win awards for AI creative work at Cannes, some of which you can see here. We were at Cannes again to show updates to our clients, but also to illustrate real proof and progress in our AI technology stack. We also announced our partnership with Adobe, with whom we have a strategic partnership around their Firefly product, which is driving our model practice, our team which trains and tunes bespoke models for our clients. Of course, you'll hear lots of people in our industry talk about AI, their progress, their partnerships, et cetera, but the reality is none of this matters if it's not going to change the economics of digital advertising for our clients. And the work we are doing with platforms like Google and Meta clearly shows that every global marketing organization needs 10 times to 100 times more content than they're currently getting from their content supply chain, and that's to get the full value of their media spend. Until now, this has simply not been economically viable with purely human models. We are changing the economics for some of the world's leading brands. You hopefully saw the announcement from GM who have appointed us as their foundational agency. That means we're working with them to bring simplicity, consistency, scale, and cost savings to how their current content supply chain operates across all of their brands. We're doing that through the lens of technology with partners like Adobe, but also through our own platform, Monks.Flow, which has been set up for them as GM Flow. Our direct client, Molly Peck, communicated this through their announcement, saying we will bring a modern approach to real-time efficient content development and described our partnership as a significant change to the way that we are doing business. We see this as a new model. It orchestrates technology and talent. It moves away from traditional time and materials commercial models. It's about technology making us faster, giving us the ability to create more, and it's about putting the very best talent on our clients' business, but still fundamentally lowering their overall headcount, which has a positive cost impact for them. We're doing that with the use of Monks.Flow. We officially launched this at CES at the beginning of the year, and we have talked about it before. We are fully in market both internally and for some of our key clients. We have more than 10 applications and a mature road map. But what's more interesting than this, it's not just point solutions, it allows us to put infinite workflows together for our clients through the lens of Pathways. It was key to the GM win. And whilst we can't name names, we're also excited that one of the world's largest tech companies has chosen Monks.Flow as one of their AI transformation platforms. The transformation happens through the lens of what we call the big six and we've gone back to a first principles approach for digital marketing. We think there are six steps needed to create great digital marketing. But the question is, can we collapse the time spent, the manual labor needed and quite frankly, the cost for our clients? We focus on insights and strategy, which is a key part of our data play. We are using our tool sets to generate deeper insights more quickly and get better understanding of marketing and product strategy, giving clients the ability to go to market faster and deliver more profitable growth. In Content, we go from high-end creative to scale adoption, transcreation and translation. There's a lot of speed, scale and spend challenges that we're solving there with our technology stack. And then in media, it's really more about automation and algorithms. How do we deliver that additional wave of content into platforms like Google, Meta, Amazon, TikTok and others and get the highest amount of performance for our clients? I think this is best explained by an example. So, let's look at what we call the impossible ad. This is a media innovation project we did together with Google, and it was presented by Google in Cannes at the Pallet, and it was celebrated as one of the hero media generative AI cases. And it really shows what using AI in the end-to-end digital advertising process can enable an entail, and we have a short video. [Video Presentation] So, these are the results, and this is not us marking our own homework. These figures have been vetted and communicated by Google. This way of working, supercharging creative and Media took down manual labor by 50%. It upped the quality of the work, which you can see through the higher click-through rates and engagement figures. But interestingly, since we were feeding the media algorithm with much more content, it also dropped the cost of sale by almost a third. Now, a bit more detail about this Monks.Flow stack. If you look at data, we have three products in market, Insights Engine, Persona and Clarity. Insights Engine helps our clients understand close to real time any trends or threats that are meaningful to their business and brands, and it helps them ideate more quickly. Persona and Clarity, I'll go into a little more detail on. Persona is a voice of the customer solution. We've built a series of data partnerships that allow us to build digital twins of any customer segment and allow us to have real-time conversations with those customers, which can replace both qualitative and quantitative surveys. We've proven that out statistically. This is not a vanilla LLM skin, although it does work with any of the major frontier models. In fact, if you have these conversations directly with one of the LLMs, then a lot of the time, that model will be confidently wrong. Having these conversations with Persona gives you statistical relevance. We do that through our partnerships with large data providers like Claritas and GWI or more niche providers like Motive Action. And we can extend these third-party partnerships with first-party data, which, of course, is part of our data practice, which has been helping clients in this area for well over a decade. When we say statistical relevance, we've been working with the MMA, a global marketing organization focused on the science of marketing by running tests for a series of clients and pushing existing survey data through Persona. And you can see that we almost perfectly predict the trend line of the answers, and we get 85% to 92% accuracy, which means in insights and strategy, we can collapse work for clients that usually takes from three weeks to three months to just three prompts. We have a similar insights and strategy solution in media called Clarity. Clarity is our close to real-time marketing mix modeling tool, currently up and running for one of the largest beauty businesses in the world across their portfolio of brands. It allows you to do close to real-time marketing mix modeling, a much more granular understanding of the impact on your business depending on your choices, the economy and what your competitors are doing. Again, all of these are actively being used by both clients and internal teams. In the creation and adaptation space, we'll spend a little less time as it's the most understandable part of generative AI. It's about visual output. Most importantly, we've built one of the strongest creative teams in that space. You can see that from our awards in Cannes. We launched the first AI director roster that we announced earlier this year. And we've been hiring some highly regarded talent that's doing some amazing work for our clients in photography and film. On the slightly less sexy side of the spectrum, we're solving a lot of typical adaptation challenges for clients that sit in translation and transcreation. For example, our translation tool has been widely tested against other solutions and consistently ranks as number one. It's taken down traditional translation work by up to 80% and now almost 90% of translation work is now done by compute rather than humans. The last part of our end-to-end stack is around media, and a lot of this is around automation. A good example, and again, one that was shown at Cannes this time by Meta as their only generative AI case, is one where we showed what it looks like to generate a multitude of content from a content supply chain, feed that into the algorithmic media platforms. In Meta's case, that's Advantage +. To create that much content in close to real-time, we have Asset Planner. It ingests media plans and other variables and can compress around three weeks of work into three hours. With Meta, we use that to test a hypothesis with our client Forever 21. The hypothesis here was, let's try and bake off a traditional digital campaign with Advantage+, a traditional campaign that has many smart people thinking about media strategy, planning, and buying, and it usually has around 10 assets per campaign. We bake that off against an Advantage+ campaign with no human-based media planning or segmentation strategy with around 270 creative assets, all created through Monks.Flow. Spoiler alert, no massive surprise, but scaled content with algorithms heavily outperforms the traditional methods, and we'll end with a video on that particular case. [Video Presentation] And the entire Monks.Flow team for all the work and leadership that they're showing there. And with that, I'll hand you back to Martin for the summary and outlook.

Yes. Thanks, Scott. Thanks, Mary. And we'll have a Q&A session after this summary. So, 11 points to make on the final slide. Firstly, half one profitability was as anticipated at GBP30.1 million EBITDA. The net revenue decrease was minus 13.5% like-for-like, and that reflected the continued global macro uncertainty and higher interest rates. Our client caution has persisted, particularly amongst our technology clients, which account for almost half of our revenues, along with the expected significantly lower transformation activity from one of our larger technology services clients. The improvement in operational EBITDA margin in content was up 450 basis points and also in DDM was up 330 basis points due primarily to cost actions, which were taken in 2023 and continued in the first half of 2024. We maintain a disciplined and active approach to managing that cost base with an increasing focus on driving efficiency across the company as well as with utilization and pricing. The number of months in the company is now 7,500, which is down around 12% versus the first half of 2023. Net debt of $183 million is after $10 million of combination payments settled in the first quarter. And we've now completed all our material M&A payments for prior combinations. And our targeted range for net debt for this year remains between $150 million and $190 million. We rebranded the company to just months. to reflect more of what we do overall. And we're streamlining our current capabilities into two practices, Marketing Services, and Technology Services. We've achieved B Corp status, and we continue to make progress in three areas of ESG, net zero impact workspaces, sustainable work and diversity, equity, and inclusion. We are capitalizing on our prominent AI positioning, as you just heard from Scott, which has been key to recent new business wins, such as General Motors. And we maintain our profit target for the year of broadly similar operational EBITDA to 2023 on a like-for-like basis. We target net revenue to be down year-on-year and to a greater extent than that assumed in May, given the outlook for technology services and continued global macroeconomic uncertainty. Finally, we remain confident in our strategy, business model, and talent together with scale client relationships, which position us very well for growth in the longer term with an emphasis on deploying our free cash flow for shareowner returns. Now, all significant combination payments are well behind us. So, with that, operator, we can turn to Q&A.

[Operator Instructions] Our first question comes from the line of Tom Singlehurst from Citi. Please go ahead.

Yes, good morning. Tom here from Citi. Thank you for presentation and thank you for taking the questions. I mean, obviously, you've been very clear about the sort of specific drag on Technology Services. I'm hesitant to go for a sort of growth before bad stuff. But is there any way that you can quantify that impact from that one client and whether on an underlying, underlying basis, there's been any sort of stabilization in tech services outside of that impact? And then also, when is specifically anniversary that one individual client sort of stepping back? I presume it's from the fourth quarter, but I would love to check that. So, that was the first [Multiple Speakers]

You mean the fourth quarter of last year, Tom?

Yes, exactly. I presume that's when that individual client started stepping back. I just wanted to confirm.

Yes, that's correct. That's correct. I mean in terms of impact on Technology Services, I think as Scott mentioned, in terms of the addressable market, that's been tougher this year, as you can see from competitive analysis, than last year. I think we continued our momentum in tech services through last year -- through the end of last year, but you'd be right in assuming the impact really started from the beginning of this year and started to fall away in Q4 of last year. It will cycle out, if that's the right way of describing it through this year. The prospects, I mean, Scott, do you want to comment on the prospects for the addressable market in digital and technology services?

Yes. I think that market is still quite tough. As I said in the presentation, when you look at it competitively, our direct competitors there have gone from strong growth in 2022 to relatively tough year in 2023 and then actually negative in 2024. So, I think the long-term prospects are still good and clearly people need digital transformation to continue, but I think it's still quite muted from a growth perspective, and that's certainly the feedback we're getting from competitors.

Very clear. The second question I had was on the FTEs. I think, Martin, you mentioned at the end, there was a 12% reduction in FTEs of year-on-year, which is, I suppose, roughly consistent with the revenue trajectory. Given Monks.Flow and all the efficiencies that potentially come with that, I'm presuming the pivot to, the step down in headcount is a function of just the reduced opportunity to generate revenue, but I wonder whether there's incremental opportunity to reduce headcount because of the benefits of automation. Is that something you can quantify at this stage?

Well, I think the first thing we have to do, Tom, is to get our net revenue and staff costs into line. I mean, JB, do you want to comment on that because you've been spending some time on that?

Yes, of course. So, obviously, as Martin mentioned, we are moving from three pillars to two practices, marketing and tech services. With this process, we are continuing to simplify the organization. Along with it, we are continuing to integrate and streamline our processes and tools that will allow our workforce to become a lot more effective. So, we are expecting to continue reducing our personnel costs as a percentage of net revenue as we pursue this integration. We know that today is still too high, but with the simplification of our organization and focusing on the operational effectiveness of our business through integrated tools and processes, we believe that we are going to become a lot more efficient moving forward.

I mean just on the headcount issue, Tom, as far as AI is concerned, it's varied. I don't think it came through on what we said about AI today, but we've said it previously. The headcount impact on the industry, if you take the five areas where there has been impact so far and there will be visualization and copywriting, there will be reductions because we can be far more efficient. I mean, every AI presentation we make, we center on the efficiencies that clients can drive, which we price at 30% in terms of the production and creation of advertising. So in visualization and copywriting reduction, in personalization at scale, probably an increase. We're moving to a model of -- an output model rather than a time and staff model or time and materials model. That doesn't mean that we've ditched time materials. We continue to do that, but assignments will now be retainer FTEs and output pricing on assets sold and price per asset. So That, I think there will be more opportunities. Media planning and buying, not so much for us because we don't have integrated networks, but you heard from Scott on what we're doing to link media planning and buying to AI efficiency and MonksFlow. I think for networks, there will be a reduction. General efficiency, probably it's mixed. There will be some increases and some decreases and knowledge transfer, probably some opportunities. So it's a mixed picture, but in the two areas where we're seeing the greatest traction at the minute, which is the first two, visualization, copywriting, there will be reductions generally in the industry of number of people. With personalization at scale, if that continues to be implemented at scale like we're seeing with GM, probably significant opportunities for increased employment. But we've got two tasks. One is to get the balance, right? Which JB just talked about. The second is, to absorb the pluses and minuses that we're looking at in the industry through AI.

Final question, if it's okay. New business wins, I mean, the GM win is significant. You've talked about this new mandate from a tech client. When should we expect them to kick in? Is it 1st of January?

Well, our GM relationship started at the beginning of July. I would say, gently and then it will increase as we go through towards the end of the year in Q3 and Q4. It will kick in obviously fully next year. I think next year, we'll see some impact from that. On AI, I mean, do you want to comment, Scott, on what we're seeing from clients generally on AI?

Yes. I think there is a lot of traction, I mean, GM is a great example of that. AI was fundamental to that win, but you can't necessarily describe it as a pure AI win, right? I mean, we're doing their content supply chain essentially, which they already had, it's now powered by AI and that's really what we're seeing in a lot of the conversations we're having with clients now. Monks.Flow is at the forefront of that conversation, talking about how they can modernize, how they can achieve cost savings, how they can create infinitely more content to fill their media budgets and get the results that you saw from the case studies we presented. So there's a lot of traction there. But, it's really -- it’s not a separate new thing, AI. It's really powering our clients' existing advertising and making it much more effective and efficient.

Perfect. Thank you very much.

The next question comes from the line of Dina Abu-Rahmeh from Morgan Stanley. Please go ahead.

Hey, good morning. Thanks for taking my question. Maybe just two on my end, both on guidance. So the first is, the lower revenue guide for the full year and an unchanged EBITDA implies more margin expansion. What makes you comfortable that this can be achieved? And then secondly, on the guidance for revenue, it's down year-on-year. Is that mainly because the tech services had a reduced client spend? And then what sort of ballpark growth are you expecting now? Would it be around down double-digit percentage? Thank you.

Mary, do you want to try and answer those?

Yes, sure. So, good morning. To take the first one, when we're talking about a lower revenue guide and what we've been seeing in terms of the market and in terms of the business pipeline. So obviously, we've said we expect our revenue -- net revenue for the year to be down year-on-year and to a lower extent than our last update in May. So that's driven by two things. Both related to -- So that's driven by two things. So, firstly, we are seeing longer sales cycles for Technology Services clients and we're seeing continued caution in our tech clients. Now, we're managing the business very carefully. As we said in the release, we will be taking further action on costs to ensure we can protect profitability for the full year. And, therefore, we will expect to see the margin expand in the second half, as you've commented on. And then in terms of the second question, which was specifically about Technology Services and the impact of that significant client. So as we said in answer to Tom's question, we saw that client begin to slow down in Q4 last year. We did see a further reduction in Q1, and then it has dropped as anticipated in Q2 and will leave us with a lower number, obviously, for the full year. Does that cover your questions?

Yes. That was good. Thank you so much.

Okay. Thank you.

[Operator Instructions] And our next question comes from the line of Joe Spooner from HSBC. Please go ahead.

Good morning. Can I just ask about the cost savings? Can you give a sense of how material those cost savings that you're now planning for the second half may be? And then just secondly, on AI, you kind of started to allude to this, but can you give a sense of what kind of price points that AI work is coming in at relative to kind of traditional work? Are you having to pass along a significant amount of those cost savings on to clients? I guess the question is kind of really thinking about as you start to look to rebuild revenues going forward, do you have to bring in or attract much more work because that's going to be more AI-focused than in the past? Thanks.

Yes. Maybe, Mary, can you talk about the cost savings in the second half? And then Scott, maybe I will comment on the AI price points.

Yes, sure. Hi, Joe. So, when we think about cost savings in the second half, obviously, as we said, at the end of June, our headcount is around 12% down year-on-year. We have taken further actions since the end of the half year, and we will be continuing to actively manage the cost base. We're not planning to comment specifically in terms of magnitude. However, I think it's fair to say we expect that the personnel costs will be lower in the second half than both the first half and the second half of 2023 and that we expect the number of Monks to be lower at the end of the year than what we see today.

Yes, I would just add to that, Joe, if you look at -- the key metric is staff costs to net revenue. And there is an improvement in the second half because there is some seasonality in the revenue. So, what you see at the end of the first half, which you can get from our numbers, will be lower. The staff costs to net revenue figure will be lower in the second half. So, it will improve as we go through the year. But there's still work to be done, and that's the guiding principle for us. On AI, Scott, do you want to comment on price points and activity?

Yes, sure. So, I think, firstly, as I said before, I mean, AI is not a separate thing that we sell or that clients buy. It’s really something -- The Monks.Flow platform is something that's powering most of our conversation, both with existing and new clients. From a price point perspective, obviously, as Martin said and as I discussed in the presentation, one of the key attractions of AI is that there are opportunities for clients to get much more effectiveness from their budget. Now, that could be if they want to do the same amount of work, then they could do that for less, and you can see some of the results that we showed. I think what a lot of clients are doing is keeping their budgets at the kind of levels that they're used to, but expecting to get a lot more output, particularly content for their money. I think from our perspective, from a margin impact perspective, I think there are actually good opportunities for margin using AI. Obviously, we don't get to keep all of that. Clients will expect some of those savings, and that's how you prove the effectiveness of the platform. They'll expect some of those savings to flow through to them. But there are margin opportunities. I think there are also interesting opportunities around the commercial model. So, as we talked about in the presentation, this is a new model. So, it's moving away from the traditional purely time and materials-based model to more of a hybrid model that still has elements of that and you still have retained core teams. You do still need people, but it also moves us to a more output-based model. There are opportunities for more SaaS-style models where there are license fees applied to some of the technology we're building. So, I think that's the opportunity for us to build out new commercial models that potentially, I think, could be more profitable.

Yes. Just to add to that, as I said before, if you look at the two areas that are primarily affected, to directly answer your question, direct visualization and copywriting, time periods get compressed, number of hours gets compressed. And I think a number of people involved also reduced. If you look at personalization at scale, the pricing model changes in part to output-based and the scale is a quantum difference. So, there's more opportunity and more headcount, more efficiency as a result. So, I think the net answer to your question is it's mixed, Joe. It depends on what you're talking about and where you're talking about. And it depends on who you're talking about. I mean, for the holding company groups, it's going to impact the number of people in their operations. And you see that already as they try to rationalize and reorganize their offers, which you're starting to see quite prominently displayed. So, I think it's a mixed answer. But we, in marketing AI do focus very significantly on the efficiency and as well as effectiveness.

Thank you.

We have no further questions in the queue. So, I will now hand the call back over to Sir Martin for some closing remarks.

Thank you. Thanks, everybody. Thanks for joining us this morning. We have another call shortly with the U.S. community. But any further questions, please let Mary, Scott, JB, and myself know. Thanks for joining us this morning.