Pearson plc

Hello, good morning and many thanks for joining us both live here in person and online for our 2023 Interims presentation. The video you've just seen highlights some of our amazing talent and innovative products that have helped drive the company to deliver excellent H1 results. with 6% revenue growth and 44% profit growth to £250 million, the highest we've seen for at least the last 10 years. Importantly, we've also improved margins, reflecting the benefits from our cost efficiency programme as we deliver on the commitment we made 12 months ago. This progress means we're confident in our group guidance for both full year 23 and the medium term. The products you saw in the video are all part of this growth story. The Pearson Test of English helped our English language division grow 44%. Fathom contributed to workforce skills revenue growth of 9%. Assessment and qualifications led by Pearson VUE grew 7%. And improvements in our higher education products grew adoption retention rates this year. And Pearson Plus delivered 938,000 paid subscriptions, a three-fold increase over last year, with registered users growing to 4.7 million. This growth was supported by our proprietary IP, vast data sets and AI. But more on that later. To touch on some of the highlights in the first half. The demand for English language learning and for the Pearson Test of English remains strong. As you know, we've recently received both migration and study recognition for PTE in Canada, which will ramp up at the end of this year and into 2024. We also continue to gain market share with PTE in India. Furthermore, the team are working hard on the development of a number of initiatives, including Mondly Works, which focuses on English language proficiency for the workplace. In Connections Academy, we're launching a new career pathways program for middle and high school students. This allows students to earn vocational certification and a college credit while they're still in high school. These initiatives will improve the competitive positioning of our virtual schools offering, driving future growth, and also increasing the interconnectivity with Credly and Pearson VUE. Within assessment and qualifications, our Pearson VUE business has shown good growth. with particular strength in both the critical nursing and technology sectors. We've now completed the PDRI acquisition and have started delivering TSA exams where Pearson VUE partners directly with PDRI. In clinical assessments, we've launched the next generation digital platform. We continue to make progress in our offering of on-screen exams within our UK and international qualifications business, and this includes the rollout of GCSE computer science and international GCSEs in English language and literature. Turning to higher ed, I couldn't be more pleased with the progress that we have made over the last six months. We're starting to see results from the investment and focus on improved product and the implementation of new sales teams and processes. We're making good headway with our LMS integrations and the transition to cloud of our MyLab and mastering platforms, which enable us to deliver upgraded, best-in-class features to improve our customer experience. We're also seeing good uptake in our iLab products, where more than 1,900 virtual labs were assigned across more than 600 courses through mastering biology alone in the spring semester. And myLabs IT is driving strong takeaway adoptions for fall with the launch of nine Microsoft Office titles on the platform. Within Workforce Skills, we continue to make good progress with vocational qualifications, including winning the new contract with the Jordanian Ministry of Education to deliver BTECs and winning three new T-level contracts here in the United Kingdom. As I've previously mentioned, Workforce Solutions should be seen as a startup within the broader Pearson landscape. The strength of our portfolio has enabled us to attract new business with enterprise customers growing 12% year over year to a total of over 1,500. As we outlined at prelims, we've built a powerful technology stack that in particular has broken down the core capabilities of Fathom into modular APIs, enabling its use across workforce skills products and beyond into other Pearson products. For example, we're bringing the global scale of English into the Fathom Skills Ontology with a view of mapping English language proficiency requirements to career pathways. We've also added new capabilities, such as modularizing our Talent Lens role assessment application so that it can be personalized to the specific roles or businesses. We spent the last six months testing an MVP of the talent investment platform with large customers. And their feedback indicates that they actually prefer the modular personalized approach versus a one size fits all. And therefore, we've decided to focus our efforts on delivering this way versus on a fully integrated platform. By being agile, flexible and meeting customer specific requirements, we'll be able to take advantage of the unique market opportunity in workforce solutions. And while this decision makes the workforce skills targets more stretching in the shorter term, we remain confident in the market opportunities. Now, I'd like to hand over to Sally, who can give you more details about the company's first half performance, and I'll be back to share further insights and exciting developments, particularly in the generative AI space. Sally?

Thanks, Andy. And hi, everyone. We've delivered another strong performance in the first half, with revenue excluding OPM up 6% and adjusted operating profit growing 44% to £250 million. Assessments and qualifications and English language learning have had a particularly good half, with virtual learning, higher ed and workforce all in line with what we expected. Profit and cash at the half-year point have grown significantly, underlining the achievement of our cost efficiency programme. All of this means we are confident of meeting expectations for the Group for the full year and the medium-term guidance we gave in March. We're also pleased to have closed the disposal of the Poles business at the end of June. Earnings per share increased by 3.1p or 14% to 25.6p due to that profit growth with tax and interest returning to more normalised levels as well as lower issued shares because of last year's share buyback. Our balance sheet remains strong, enabling continued investment in the business as well as increased shareholder returns, with the £300 million share buyback commencing in the third quarter and an increase to our interim dividend by 6% to 7p. Looking at revenue by division, assessments and qualifications revenue grew 7%, driven by an excellent performance in Pearson VUE, particularly in the IT and healthcare segments. Following the disposal of the POLS element of OPM, our virtual learning underlying growth is reported excluding this part of the business. The virtual learning revenue decrease of 15% is driven by the ASU contract element of that business, which will continue to impact growth in the second half of the year. Virtual schools were supported by good retention rates, offset by lower enrolments and lower district partnership renewals. English language learning revenue grew 44%, with an outstanding performance in Pearson Test of English. And workforce skills revenue grew 9%, with good growth in both vocational qualifications and workforce solutions. And higher ed revenue was down 2%, reflecting a continuation of the trends observed in the fall semester of the 22-23 academic year, with a decline in enrolments and a loss of adoptions to non-mainstream publishers. Orbit, the latter, improved slightly. Turning to profit. Group-adjusted operating profit grew 44% to £250 million, driven by operating leverage on revenue growth and the cost-sufficiency programme partially offset by inflation. Half-year profit margin grew from 9% to 13%. At a divisional level, the same group things I've described are apparent in each division, in addition to increased investment in workforce and English, alongside some negative phasing in higher education relating to amortisation and positive phasing in virtual learning relating to contract closures. It's worth remembering that the seasonal nature of the English and HE businesses mean that their margins are routinely lower in the first half compared to the full year. Operating cash flow of £79 million was driven by increased trading profit as well as strong collections and portfolio changes. Our balance sheet remains robust. with net debt at the end of June of £900 million. That increase from £800 million last year is largely due to dividends, tax and the 2022 share buyback more than offsetting strong operating cash flows. In short, our operating and free cash flow performance is stronger than ever. We've continued to invest in the business in H1, both organically and inorganically. Internally, we have continued to shift resource, cost and capital between areas of the business. You'll soon hear about the innovations we have developed in the AI space. All of these have been achieved within the existing spend plan. We've also made strong progress on reshaping the portfolio. We completed the acquisition of PDRI in March, significantly expanding Pearson's services to the US federal government and in the recruitment assessment space, whilst also adding to profits. The disposal of our polls business is also now complete, further focusing Pearson's portfolio towards future growth opportunities. The board has continued to demonstrate its commitment to a progressive and sustainable dividend with a 6% increase in the interim dividend to 7p. And whilst we also remain committed to maintaining a strong balance sheet, given our strong operating cash flows, we can do that, invest in what we need, but also commence our previously announced £300 million share buyback. Moving to the outlet for 2023, we are confident of meeting group expectations with low to mid single digit revenue growth, excluding OPM and strategic review, and group profit margin growing from 12% at the full year last year to mid-teens, benefiting from the achievement of our cost efficiency programme. Importantly, as we've proactively reshaped the group, we are now generating some 45% of the consensus full-year profit expectations in the first half of the year. This represents a fundamental shift from where the business was even last year. We're also reiterating our guidance of mid single-digit group revenue growth over the 22 to 25 period and for margins to rise to the upper end of mid-teens in 2025. In terms of divisional guidance and phasing, there are a few points I should highlight. We are raising our revenue growth expectations for 2023 in English language learning from high single digit to around 20%, following strong growth in the first half, which we expect to normalise in H2. We're investing a portion of that operating leverage on this improvement to support future growth opportunities. In workforce skills, given the considered approach that we are taking to ensure we have the right go-to-market strategies to scale that revenue in workforce solutions, our 2023 and 22 to 25 growth guidance is likely to be more stretching. Continued growth in Pearson Plus subscriptions will lead to a shift in higher education revenue recognition from Q3 to Q4. Contract timing in ANQ will see delivery in earlier quarters, meaning Q4 revenue growth will be lower than average. We're confident in achieving full year guidance. And it's also worth flagging that the termination of the ASU contract will continue to impact growth in the second half of the year for virtual learning. In virtual schools, our expectations for revenue growth remain unchanged, with H2 enrolments impacted by the loss of that large school contract we've previously discussed. I appreciate there are many moving parts given the recent portfolio changes we have, but to put things simply, there'll be no revenues or profits relating to OPM, being ASU plus polls, or the strategic review businesses from the beginning of 2024. So in summary, we've had a really strong first half performance in all financial metrics and are confident of delivering on full year expectations.

And with that, I'll hand back to Andy. Thanks a lot, Sally. Now, I'd like to take a deeper dive into our AI capabilities and highlight some of the things that we're working on. We're still very early into the generative AI journey with new developments happening on an almost daily basis. So it would be wrong to claim victory so early in the game. And we're also very aware of potential risks and turbulence ahead. But to be clear, we see generative AI as being a real long-term positive to the company and not just in higher education, but across our portfolio of businesses for the following reasons. We believe the value of our proprietary IP and datasets will increase over time. We have deep AI experience and expertise across the whole company. We're starting to introduce new AI enabled proprietary IP and datasets will increase over time. We have deep AI experience and expertise across the whole company. We're starting to introduce new AI-enabled products across the business. And we've already achieved... ...but have yet to commit to. What this interest does demonstrate is the real value to be had of owning your own intellectual property. We're also continuing to monitor legal and legislative developments very closely. Now, one of the biggest areas of focus for the team has been around the positive implications of embedding generative AI into our higher education courseware. We've been working with that with one goal in mind, namely how to improve the learning experience for both faculty and students. As I mentioned in our recent update, one of our key advantages is the quality of the data sets we're able to use in generative AI due to the ownership of our IP. Simply put, the purity of any data input equates to greater accuracy of any data output. As you'll shortly see, we've been working on ways to embed generative AI within the learning experience. We think it's very important to utilize this technology in a frictionless and seamless way to benefit and enhance how students learn. Core to everything that we do is maintaining trust and quality, and generative AI is no exception to that rule. We're not interested in utilizing this technology merely to provide students with a shortcut to an answer. When we tested different LLMs with a question from Campbell's Biology, they often didn't get it correct. So we believe delivering product that is reliable, accurate, and trustworthy is paramount. Now, here to give you a quick insight into just some of the product innovation we're launching for Back to School, I'd like to hand over to our new Chief Product Officer, Tony Prentice.

Thanks, Andy. We are working on adding generative AI to our products across the company. But for now, I'm here today to show you the AI features we've been working on in two of our key higher education products, Pearson Plus and MyLab Mastery, both of which will be available for this year's Back to School. The goal of these AI study tools is to help students learn and actually understand tough subjects. We want students to stay and learn inside the Pearson product environment. Let's jump in and start with Pearson Plus. Let's imagine I'm a biology student and I'm using Campbell Biology and Pearson Plus. Just want to note that everything you're about to see is being offered in beta and a small number of titles for back to school. My professor has been teaching genetics. I want more information on what I heard in class today on recessively inherited disorders. I see the icon in the bottom right and I can click there to access the AI study tools. I ask it to summarize and immediately it does for recessively inherited disorders. Okay. Next, I want to learn more about dominant disorders. So I scroll down to that section, and I realize I learn best when I can compare and contrast. So I ask the AI to do that between recessive and a little more practice on this topic. So I can ask for practice, and it gives me questions as well as answers, which is one of the things students ask for the most. So what's really cool for me as a product person that we're taking the Campbell biology content created and refined over the years by subject matter experts and combining it with the conversational nature of large language models to break down the topic in a really personalized way to help students actually learn. Maybe I need some more help to really understand these concepts. So I stay in Pearson Plus, but move over to channels where I see the corresponding section from Campbell's biology. I can see this new feature on a tight schedule. It uses AI to summarize what's in the videos so I can see if it's the information I need. Yep, this looks like it. So I'm going to go in and watch the video. Or the sword. And after that, I can scroll down and get yet another set of practice questions. So that's Pearson Plus and our new embedded AI tools. So now we're in my lab mastering. Imagine I'm a chemistry student and I'm trying to figure out this problem. How much mass is converted to energy? As you can see, this is a pretty complicated problem. I think we covered this, so let me give this a try. This is my first try and I can see I got it wrong. You might remember what it was like to study science, get like a really complicated problem wrong and how frustrating it was not knowing how and where in the equation you were making the error. But now I can ask the chat bot for help. So I click on the button for help. And what you see is a series of smaller bite-sized questions. We're not just giving the student the answer, we're helping the student work through the steps to get to a solution. All right, so let's see. So I answer the first question, and then you can see another problem that goes a little deeper. We're going to ask questions to gently nudge the student through the problem. This is what we're trying to do to help the student actually do the work. And again, this continues for a few more steps. Eventually, I'll have all of the concepts that I need to try and do the final answer again. So now I'm going to go back and put in my final answer. Let's see. Sweet. I got it right. Scientific subjects like this are best learned by completing practice problems, solving equations, and working formulas. Students often get stuck in these complex problems and don't always have the right tools to really work through them. So instead of students having to wait for office hours or just giving them the answers, we're introducing this AI-assisted learning tool to help them learn at their speed. What you just saw is just the beginning of our AI product journey. This is version one. Generative AI is rapidly changing and advancing, and as our own testing and research evolves, you will see our products and features evolve over time to meet the needs of students. I look forward to showing you more as we continue to learn and innovate.

Thanks, Tony. And by the magic of television, you transported yourself from Hoboken to 80 Strand. So Tony's here for Q&A, both for you online and for you in the audience. Welcome again and thanks. More from Tony later. Okay, as you can see, this is just the start of what I believe will herald an exciting future opportunity for Pearson as we thoughtfully utilize the capabilities of generative AI across the whole company. As you can see from the examples in Pearson Plus and in Mastering, the ability to give students a highly personalised experience is the cornerstone of the AI product, helping students learn in ways that best fit their unique needs. It's just one of the ways we can meet consumers' expectations and build a further trusted relationship with them. So in conclusion, I couldn't be happier with our position today and I'm optimistic about our growth prospects going forward. As you've heard throughout today's presentation, we're executing well on our strategy and our financials bear that out. We have a stable, growing and more evenly distributed portfolio of businesses with learning and certification at the heart. We see a growing number of opportunities to leverage our IP and data sets to create compelling and innovative products incorporating generative AI. And all of this means we're well positioned for the emerging future of learning. With that, Sally, Tony, and myself will be happy to take your questions. Joe, over to you.

So we'll go first question from Matt Walker, just in the middle here.

Thanks a lot. Good morning, everyone. I've got three questions, please. The first one is on higher ed. There was a bit of an improvement in Q2. Maybe you can explain what was behind that and how does it impact the full year? Second question is, you're starting the buyback in Q3. Do you have a rough idea of when you're going to start it? and how long it's going to take to complete. And then on ANQ, you had a good result in ANQ. Is the 7% a roughly good figure for the year, or is there any reason it will slow down in the second half?

Clearly three questions for you, Tony.

Three for me? Thanks, Matt. So, yes, Q2, we saw growth. You'll remember that we spoke at the beginning of the year about Pearson Plus and as we see growth in Pearson Plus. We see those revenues come in as the subscription comes in. And so we were always expecting some of the revenue from Q1 to be in Q2. That's what we saw when we looked at the Q1 results and you see that coming through in Q2 as well. And we're confident of hitting that full year guidance with an improvement in higher ed. For the buyback, obviously we announced that at Q1 trading. I've waited to start it because I'm drawn on my RCF at the moment. That's just the normal kind of pattern of cash flows for Pearson during the year and given that we did the share buyback. last year as well. Half One Cash has been really excellent. I'm really pleased with that performance. So that means that rather than, I think I was saying H2 at that point, now we're talking about Q3. So it's imminent. Paperwork's on my desk. We'll be ready to go soon. And then ANQ, really pleased with that growth in the first part of the year, particularly the growth in view that's come through there. I have pointed out in the kind of detail of the guidance that Q4 for ANQ will be a little bit lower. And that's just purely because of where some of the contracts have fallen and rev rec over that. But we're really confident about full year guidance for ANQ as well.

And very briefly on the higher ed piece, as mentioned earlier, I take note of the uptick in retention rates into the 90% plus range. From the 80% plus range, that shows that more faculty, we have great product market fit. With that product, it's often a leading indicator that shows that before we've done generative AI improvements and enhancements, which we think are unique in the market, we've had excellent feedback from faculty. We've been testing this with both faculty and students over the past several months. I think that's a really good indication. And then Tom and the team, as I referenced as well, we've reorganized the sales team and we're seeing real impact from that as well. So, you know, as you know quite well, it's a process as we're going forward. But some of those indicators, I think, are quite positive.

Can we take the next question from Nick?

Just a minute. Yeah, morning. It's Nick Dempsey from Barclays. I've got three left, please. So first of all, can you maybe explain in a bit more detail the exact change of strategy we're talking about in workforce skills? So in other words, what you were hoping to achieve with the skills platform and what you're currently now hoping to achieve and why that is a change to your hopes for revenue growth? That's the first question. Second one, on cash flow, it was pleasing to see in the first half that cash product development spending was pretty much in line with the amortization in the P&L. Is there some timing there? So should we assume a notable gap between the two like we had last year, which was 54 million? Or is this the new normal? And the third question. So in the polls business, I think, Sally, you mentioned a positive effect from contract closures, giving you a bit of a positive bump up in the first half. Can you just explain what drove that and how it affected revenue and profit?

Shall I take the first one and you can take the other two, Sally? Hi, Nick, in short, there isn't a change in strategy and workforce skills. And what we are pointing out is, there is a change in the demands of customers in terms of how they want to access the portfolio of workforce solutions that we're providing. And I wanted to get away from in essence, when we showed the concept car of video at prelims that people thinking of this as a one set product, You shouldn't think of it as one set product. You should think of it as a set of customizable solutions and that vary and are applicable on a company by company basis. One of the things that we're finding is more and more companies become more and more engaged in the upskilling and reskilling process. They themselves are learning a lot more about what capabilities they need and the solutions that they need. vary company by company. So rather than have the impression that this is a one-size-fits-all, I think, as I referenced in my script, you should really look at a suite of the talent investment platform and talent lens, fathom Credly, as a suite of solutions that we're able to offer. And as I say, we had 12% uptick in the number of enterprise clients, including many leading Fortune 100 companies. so we're getting scale we're getting feedback it's about meeting the customer where they need to be and being agile and fluid and i think that is a theme that you're going to see happen more broadly as we move from this role even take higher education you know a textbook was a fixed product and now we move into the the digital world that you see of constant iteration and change So that's the sentiment behind what we mean in terms of the shift to talent investment platform. And it allows us then to really flex and focus on delivering the needs of enterprise customers, whatever they happen to be in each of their own unique circumstances.

So then on the cash flow, well spotted in terms of the first half cash flow. And yes, I think we have talked about that theme of product development, amortization and spend converging more over kind of future periods. And actually for CapEx, that CapEx number coming down compared to where it had been previously. And then for polls and that contract piece, effectively, we've got from a revenue recognition which drops through 100 percent to profit, which is why I'm pointing out for a profit point of view, more from a comp basis in the first part of the year. That's just to do with how the revenue recognition worked on a couple of contracts that we closed out before. the disposal part of the deal. And then you'll see that in up from a common basis reversing in the back part of the year. And then we've got exactly the opposite happening from that high red phasing piece, you know, a negative impact in the first half of the year, which again, will reverse out in the back part of the year. So it's just in the detail when you look at the divisions.

And we go to Adam in the front.

Hi, good morning. It's Adam Berlin from UBS. I've got three questions. First one for Sally. Can you confirm that the H1 performance was above budget? And if so, why aren't you raising guidance for the full year? What are you worried about in H2? My second question is about higher ed. I think a few of you could answer this. How important do you think it is in this market to have the best product rather than the cheapest product? What gives you confidence that all this investment you're making in product development is going to work for adoption share And do you expect to see adoption share gains this fall? And thirdly, a question for Sally. Can you give us free cash flow guidance for 2023, please?

You can take the middle one and I'll take the first one.

Yeah, I have a comment on the first one, but you kick off.

OK, so I won't get into our kind of internal budget, but clearly we're really, really pleased with the first half performance and it's better than maybe people would have expected, which is great to start off the year with a real positive. You know, we're only half one, but, you know, I'm more confident than ever of hitting that full year number. And then in terms of free cash flow, and then I'll let you come back to the middle one. So we've guided from an operating cash flow perspective to 90% plus. Then from a cash perspective and an interest perspective, you'd see similar numbers to what you would have seen before. And then the only other thing to think about is those reorganisation costs that we had at the latter part of last year. I'd guided to... I think it is of that being in cash terms. The rest was just P&L terms. I think we've seen about 40 of that go through in the first half of the year. The rest will go through in the back half of this year and maybe a little bit into next year.

And then quickly back to the guidance question first, Adam. You know, we probably, when you have time, I realise it's early, study the numbers, English language learnings, for example, and Sally referenced it. In the first half, we took some money out of the increase in revenue and reinvested that into providing future growth, both on the ground and preparation for what Canada's coming from. So... we could have delivered higher than the 250 in H1 if all we'd done is focus on a number and hit a number without having to think about also being more flexible and fluid about how we manage the company for consistent, sustained future growth. And you're going to see that pattern continue. That's why we're reiterating how comfortable we are with guidance for H2. But rather than stick a number out there, I'm sure all of you who are far smarter than I will come to your own conclusions. I want to be in a position where we're building this company to deliver sustained growth and remain flexible and agile to react to whatever's happening in the market across our whole portfolio of businesses rather than just being focused on hitting a number because that's really short-term thinking and we are in the business of really trying to grow sustainable growth that's why we gave and said what we said, you shouldn't read any more or less into that. Certainly, we see all things being equal. You know, as we sit here today, we don't see any headwinds or banana skins or whatever in the second half of the year. But we want to remain thoughtful, as we always have about you know, how we manage this company and how we continue to grow, deliver growth, which we have demonstrated, I think, quite a good track record over the last two or three years.

Actually, I'm going to pick that up as well, because you asked about risk and I didn't particularly address that. There's no risk that I'm worried about in the second half of the year. In fact, I'm more confident than I would have been in previous years.

Yeah, I guess, don't overthink it in a way. Quality always wins. Doesn't matter what you're in, quality always wins out. And quality doesn't necessarily mean the most expensive either. You look at the Pearson Plus and the threefold increase in subscriptions, where actually we're offering really high quality and additional features. in addition to the e-techs within the Pearson Plus environment, all at a very affordable price. And so that's one thing I think that we really focus on, trust and quality. The second thing is what we do is really hard. And if you're staying for the demonstration after this, when we're going to get a bit more with Tony and I into our future roadmap and generative AI, You know, we have a great responsibility on delivering a word that I hadn't really come across before I joined the company, efficacy. You know, it's okay having something cheap and not very good if it doesn't work, particularly when it comes to learning and particularly education, any form of learning, but particularly education. That's not good for anyone, and it's not a long-term solution. I think what you're seeing actually is some in higher ed, some of the rebound from, you know, we think lower adoption of faculty of OER and what I called, I think in one of the previous calls, homebrew, where during the pandemic, faculty were kind of making their own stuff up, a bit of this, a bit of that, bit of the other. Guess what? They found it didn't work. wasn't as effective, didn't provide the learning that faculty wanted to deliver to their students. And that's why we're going to remain focused first and foremost on quality and trust and secondarily on price because the two are not necessarily equate to each other. Does that make sense?

Can I add just one thing? Yeah. And just add one thing on that. We've been talking a lot to students, and the number one ask they have is to help get better grades in my class. And so that's what they're asking us for. And so I think, again, and we're going to get into this, it's quite hard, but focusing on how do we help them actually learn, and they know that it's actually quite tricky. If you saw in the video, especially with some of the scientific questions, when you're in the middle of it, you get stuck. It's just really hard and really frustrating. And so that, to me, is what they're telling us where they need help.

Let me go to Tom Singlehurst.

Morning, thank you. Yeah, Tom from Citi. A couple of questions, maybe three. Let's see. On English language, Canada, it's clear that we shouldn't factor anything into our numbers for this year. But can you just give us a sense of, A, the size of that market and then, B, how market share changes? may be expected to ramp based on, let's say, the Australian precedent. I don't know what the share split is between you and ETS and Cambridge in Australia. That would be interesting to know. Second thing, on higher ed, it sounds like you've done your bit on upstream, making sure you've got the adoption in place where you want it. Have you any thoughts at all on how many students are going to turn up in September? That would be interesting. And then finally for Tony on the AI investment, they look like very cool tools. Are they going to be included in Pearson Plus as well as MyLab and Mastering? And then how long will it take you to ramp it up to the full suite of content? Thank you. Do you want to take Canada?

First two, and obviously the last one's for you. We're delighted to have got recognition in Canada and you'll notice it came in two pieces. So the immigration part first and then the really important student piece second. There isn't any market share data, so we have to estimate it, but we think it's around 200 to 250 million pounds, possibly a little bit higher. And, you know, these things do ramp over time, but it's going to be You know, a reason that we'll see good growth in this business and in PTE specifically as we go into next year and quite a few beyond. And, you know, you look at Australia and the period that that ramped over, that feels to me like it's probably a good way to estimate that. Higher ed and enrolment. So, yeah, we're pleased with, to your words, the things that we've done in the first part of the year. So that performance and that sales team that we restructured at the beginning of the year and seeing some proof points in terms of data coming through. We're pleased with what we've done on product from a stability point of view through some of the new things that we're putting in that we'd planned this time last year and some of the stuff that we're putting in from an AI point of view. So lots of reasons to feel good about that from an enrollments point of view. You'll have seen spring normalised a little bit. One thing I've learned is not to call what enrolments are going to be for fall back to school. We'll find out in September, October, I think is the way to think about that. And the AR1s for you.

Yeah, so as we mentioned in the video, what's really interesting is we have this amazing IP, we have this amazing expertise, and then we have some of these great conversational nature of the LLMs. And so what we're doing is trying to take advantage of both. What you saw, we're launching on Pearson Plus, so three titles. I believe it's Tro Chemistry, Biology, and Kotler's Marketing. And the plan is to really put it in as beta because it's not all the way there yet. And so what we're going to do is we're going to put it in, we're going to learn with students. And I think being on the field rather than being on the sidelines is the exact right place to be right now. And so we're only going to do a limited number of titles and we're going to be all over maniacally staying close to the students. And then we're going to iterate. The models are going to iterate and we're going to iterate. And again, we're in the business of helping students learn. And so I think what you've seen is us jumpstarting already ahead with a great solution, but staying really close. We're going to agile kind of like learn together as we go through this.

Perfect. And we'll take the next question from James.

Morning, it's James Tate from Goldman Sachs. Just one question, please. Could you give some more colour on the benefit you saw from the cost savings in H1 compared to the 120 million you've talked about for the full year? Just trying to understand sort of the phasing between H1 and H2 there. Thank you.

Just one. Just one. I think it's one for me. So 120 million for the full year. So half of that 60, we were almost at the 60. So, you know, high 50s in the first part of the year. So, you know, all we have to do in the second half of the year is just execute the same things in the second half of the year. So those cost savings are super secure.

Can we take the next question from Luke?

Thanks. It's Luke Holbrook from Morgan Stanley. Just a question really on the licensing. You said you've been approached by multiple parties. What's preventing you kind of go down that route? And can you elaborate on some of the conversations that you've had in that regard? And just secondly, some interesting, I guess, product launches that you're doing on the AI side. Are you looking to try and commercialize it further through product pricing? Or is this just being wrapped into the existing products fee structure that you have? Thanks.

You want to take the first one?

Yeah, I can take the first one and do the second one. I want to, you know, I don't want to just take the first offer that comes along because, as I said in my script, I think the value of this data set actually increased. And once you think about the data sets we have around Fathom, the data sets we have around Credly, everyone thinks about the data sets that's just within our high-res courseware. Because what we have is really valuable. And it's easier, I think, for us to comprehend with a book. So I use that as an example. You know, these large language models are being trained on ingesting words and facts and things like that. Now, if one were to take a light novel, you know, nothing wrong with Mills and Boone, but a Mills and Boone type, you know, novel, and ingest that, that is going to deliver a certain amount of output. You get a lot more than words, right? with our product. Because as you even saw in that one example, the complexity around the construction of how we develop those textbooks, which has been developed over decades, is very, very valuable. So we're not in the business of just selling content for the sake of selling content. We want to be very thoughtful and specific as to what we get out of this versus what a third-party LLM gets out of it. There are some interesting opportunities we have that present more of a walled garden approach. There are some broader opportunities, but the space itself is moving at a highly, highly fast pace. And so being first and just announcing a deal for the sake of announcing a deal is probably in hindsight, you may look back and go, that probably wasn't the smartest thing to have done at the time. And so, you know, it's not for a lack of interest. It's not for, you know, lack of thought on our side. It's just that we want to be prudent in terms of, because once you put this stuff in, it's hard to get out, as you know. And then there's, you know, not this isn't dictating us, but it is still early days in the litigation and legislation areas as well. And that's moving at a pace. So being first is not necessarily the best long-term strategy with this. Utilizing what we have in products like what you've seen and constantly iterating is the way we think to go forward. But it opens up a lot of, to reiterate myself, we have a lot of really, really valuable data sets. Anything you want to add, Tony?

Yeah, so I've been here for about three months. And one of the things I've been really pleasantly surprised by is, first off, how much existing machine learning, data science, AI already existed at the company. I think you saw that in the videos already starting with the Pearson Test of English. so i think there's a lot that we always had and have in our dna then as we look forward i think our goal is to focus on helping students learn and so we'll what we'll do is we'll look at how we can bring these together but don't forget there's commercial um lms there's building our own there's some of these smaller models so there's a whole space here that's evolving and so what we'll do is we'll continue to be evolving and we'll make sure our customers have that of the best of both worlds And then we're always really looking at value propositions. And so we'll always be looking at how do we solve those problems and what's the right value prop that we bring at it. So those will all be what we're looking at. And again, that's why on the AI thing specifically, we're only going for a few titles so we can test and learn and be really agile as we go through this.

And on the pricing question, of course, very simply put, yes, we could extend that through. There is actually a pricing increase in Pearson Plus at this back to school. And for those who are not taking MLM, I'm getting my... acronyms mixed up, there's a pricing plan going to be introduced for channels as well. We've seen really good traction for channels in over 170 countries around the world, and you saw how we're integrating generative AI into that feature. So let's, as Tony says, let's learn, let's iterate, and clearly we then have pricing opportunities.

And as well as pricing, of course, those added features and the extra value there is for the student will also start eating into the secondary market and people come into the product because you're getting something that you can't get. You just can't get it from a textbook. When you saw it on the screen, it's super cool. You can't get it out from a textbook.

Yeah, that notion of not leaving your environment is really important. You don't have to click out to click into somewhere else. It's all embedded. It's all within the one environment. That's really, really important. Any others?

Go question from Ben, just

Hi there, thanks for the presentation. Two questions, please. Firstly, just to clarify on the English language guidance, you previously guided to high single digit growth with improved margins. And now you're highlighting that you're investing a portion of operating leverage. So should we be assuming a similar level of operating profits despite the high growth? And then secondly, on AI, Could you describe how AI additions place Pearson's offering against peers? And I'd be really interested to hear how you think Pearson's offering stacks up against peers today and what the vision would be over the next couple of years. Thanks.

So on, I'll take a yellow. So the guidance effectively from a profit perspective would be the same. So we would have had high single digits from a sales point of view. Now it'll be circa 20%. You'll still see an improvement in margins. And then that investment, we've just taken some of that sales margin. Extra. So the difference between five single digits and 20. And rather than seeing that drop all the way through to sales, we've invested some of that in two things. Marketing. We're seeing markets like China open up. We want to make the most of that. And also in some places we're getting to the capital. To capacity, nice problem to have. So the kind of view test centre network, opening some new test centres across India to make the most of the extra volume that we've got, particularly as we start to think about Canada coming online as well. You know, it'd be a real shame to hit up against capacity. We want to make the most of all the great work that we're doing.

It really plays into Adam's question earlier around the broader guidance. You know, investing now for the future to maintain sustained growth, I think is really, really, really important for this company.

Do you want to kick off with the... Sorry, I missed... Was there a name of a specific competitor that you were saying? I missed the last part of your question. Just in general. Okay. So I think if you think loosely, two big areas, there's the over-the-top commercial LLMs, BARD, ChatGPT, kind of think like that in one space, and then you have closer in ed tech. I think from the broader stuff, what we've seen is more... let's say, great, amazing LLMs that can answer questions, but not necessarily targeted towards helping students learn, which is where we're focused. So that's kind of how we see that. I don't think we see anybody in the ed tech space yet, but I think we have to assume that there'll be stuff coming. So I think what you're seeing is actually quite differentiated for now. And I think the whole point of us being in market is to try to stay ahead. So that's how I would think about the two sides of the equation. And I think that's how we're going at it and making sure we're there now.

And as we said in the prepared remarks, we're not in the business of just giving students answers, because that's very short term. And actually, the research students realize that. The more and more they're getting to realize that actually, it's the learning piece of this that's critical and very important. And that's why we're focusing very much on that. 4,009 students, I believe, is the exact number that we've talked to and been iterating with. Why it wasn't 4,010 is beyond me. But getting feedback and getting constant iteration is really important. And for those of you who are able to stay after this presentation in person, we're going to get a bit into the future. We're starting in higher ed, just as we did with Pearson Plus in higher ed, because we think it's of interest. But you can imagine as being able to, and you saw some examples, you know, if you think about Pearson as being in the learning and certification space, delivering a certificate every 1.6 seconds and having this mechanism, whether from Pearson Test of English through Pearson VUE to Credly, And the content that we're able to create, applying generative AI across that suite horizontally is very, very interesting for us as a company. And that's unique. No one, we have no peers who can do that end-to-end solution. That is unique. That set of assets is quite unique to us. And you then start to utilize generative AI on an individual, as you as a learner, to help improve you. as you go through your career. We will get more into this after because there's some stuff that we'll say offline. Any questions, Jo, from online?

Yes, there are the two. So let's start from one of them. Analyst has asked, how should we think of slower growth in workforce? If I can answer that question, then I'll come on to the next one.

Oh, sorry, they're one by one, are they? So I don't think we said slower growth. I think we've said that, and I'm not changing the guidance that's out there, which is double digit this year and more than 20% over the medium term period. I'm just pointing out that as we work through making sure we're really delivering to customers what excites them, that it's a little bit more stretching. So I didn't say slow growth, though.

Thanks, Ellie. And three questions from Alpha Value. So the first is, are the generative AI tools developed on open source LLMs? Second question, do you run them on your own servers? And the third question is, do you have your own team for machine learning?

Do I take those? Yeah. Yeah. So I'm going to take them backwards. We do have our own team, although to jump to the first question, we're using, for the ones that you saw both on MyLab and on Pearson Plus, we're actually using ChatGBT behind the scenes. We're using it through Azure. No, we're not running it locally on our servers, although we are also testing almost every single open source model out there against the commercial ones. And what we'll continue to do is look at performance versus cost as this all evolves. But specifically, we're using the commercial ones. I don't know if this is what was behind the question, We take data privacy, data security incredibly, incredibly, incredibly important. And so we're using Azure in part to help us. And one of the most important things that both Azure and GBT will do with us is when students put their data in, that won't be used to train future models. And that's an absolute criteria for us as we're going through this. And I don't know if that was hidden underneath the question, but that's an absolute kind of principle for us.

And with that, Andy, I'll hand back to you.

Thank you. Well, thank you all for those of you who have joined us online. Appreciate your interest in the company. And thanks again for those of you who joined us in person. And we will look forward to sharing more about Gen AI with some of you here very, very shortly. And see you in October.