Thank you for standing by and welcome to the Appen Limited Fiscal Year 2023 Results Release Conference Call. All participants are in a listen-only mode. There will be a presentation followed by a question-and-answer session. If you wish to ask a question, you will need to press the star key followed by the number 1 on your telephone keypad. I would now like to hand the conference over to Mr. Ryan Kolln, CEO and Managing Director. Please go ahead. Thank you very much, Andrea, and good morning, everyone. Welcome to Appen's FY 2023 results presentation. Today I'm here with Justin Miles, our CFO, and Rosalie Duff, our Head of Investor Relations. Turning to the agenda on slide three, there are four sections to our presentation today. First, I'll provide an introduction to our FY 2023 performance. Following my intro, Justin will share further detail on our financial performance. Then I'll present our strategy and priorities before discussing our FY 2024 outlook. Turning to slide five, where I'll highlight three key elements of our performance in FY 2023. First, revenue was materially down in FY 2023. In response, we adjusted our cost base to reset the business. Second, although full-year revenue was down, we did see some positive trends in Q4. Third, we had good traction in generative AI-related project wins, particularly in the second half of the year. Now I'll go into more detail on each of these. On to slide six. At a group level, our FY 2023 revenue was down 30% year-on-year. This reduction was mainly driven by a large customer and was influenced by the tech spending slowdown and uncertainty related to generative AI. This is clearly a very disappointing result. In response, we reduced our annualized OpEx by $60 million. Because of the timing, we did not capture the full benefit of the cost reductions in FY 2023, which impacted our profitability. In December, we realized the benefit of our actions as we exited the year Cash EBITDA positive. Turning now to page seven, where we have a quarterly revenue breakdown. In Global Services, we saw typical Q4 seasonality. We acknowledge that for G lobal Services, Q4 was materially down year-on-year. However, it's a positive signal that revenue declines had stabilized after a challenging year. China delivered a record quarter in Q4 of $11.1 million, signaling a strong return to growth for China. This is mostly due to the return to normal following COVID. Finally, New Markets excluding China also experienced growth in Q4 when compared to Q3. These results are encouraging, particularly as we look for revenue stabilization across the business. On page eight, I'd like to share some highlights on our generative AI progress. FY 2023 was an exploratory year for generative AI. At the start of the year, many customers were testing different approaches to building LLMs, and this resulted in a lot of pilot projects for Appen. As the year progressed, our customers scaled their data needs. Because of this, existing customers increased their spend on data, and we added new customers who were looking for more established data vendors. As a result, we saw a 410% half-on-half revenue increase from LLM-related work. We now have several projects where annualized run-rate revenue is well above $1 million. What's very encouraging is that we're now working with roughly 80% of the leading LLM model builders globally. We expect many LLM companies to further scale their data operations in FY 2024, particularly into international markets. One of Appen's strengths is scaling AI data internationally, and therefore we're very bullish on LLM-related growth in FY 2024. Moving to slide nine, I'd like to address the disappointing decision from Google to end their contract with Appen. While Google's decision did not impact our FY 2023 performance, it's an unfortunate outcome that I'd like to address. We responded swiftly to the news by reducing our cost base. The key focus areas of cost reductions were the direct and indirect costs related to Google. As previously announced, we identified $13.5 million in cost out. The work that we did for Google was challenging, particularly related to the onboarding and management of the crowd workers to the Google platform. A significant number of internal resources were dedicated to this account, hence why we have been able to identify and reduce costs quickly. So while the news was very disappointing, we have taken the opportunity to streamline our operations, which will benefit other areas of the business. I'll now ask Justin to go further into our FY 2023 performance. But before I hand it over, I'd like to take a minute to congratulate Justin on his appointment as CFO. Justin has been with Appen for over eight years and has done a tremendous job throughout that time, specifically during the last eight months as interim CFO. I've worked very closely with Justin throughout my time at Appen and couldn't be happier that he is our CFO going forward. I'll now hand over to Justin. Thank you, Ryan. Good morning, everyone. A reminder that we report in US dollars and that all comparisons are to the full year ended 31 December 2022 unless stated otherwise. Starting with the financial summary on slide 11. Total revenue decreased 30% to $273 million. This is mainly due to a lower contribution from our global customers, reflecting lower volumes as our customers optimize their spend and reduce costs in response to the challenging external environment. The lower contribution from our global customers also flows through to the New Markets business, where we saw a decrease in Global Product revenue. Our gross margin percentage, which is revenue less crowd expenses, was down 130 basis points to 36.3%. The decrease was largely due to a shift in the customer and project mix during the year. Underlying EBITDA and associated margins were significantly impacted by lower-than-expected revenue and gross margin, as well as a proportionally higher cost base coming out of FY 2022. Also, the first full-year benefit from our cost out programs executed during FY 2023 will be in FY 2024. Primarily due to the decrease in EBITDA, we reported an underlying net loss of $52.8 million. We recorded a statutory net loss after tax of $118.1 million, which includes a non-cash impairment charge of $69.2 million pre-tax. The impairment relates to Global Services. Turning to revenue on slide 12. At the group level, revenue was down 30%. Most of this decline can be attributed to Global Services as well as the impact from Global Products. Global Services revenue declined 36%, impacted by reduced volumes as customers looked to optimize and reduce their costs. Most of the revenue decline was due to a large customer. However, we saw encouraging growth in LLM-related work across Global Services in the second half of FY 2023. In New Markets, revenue declined 8%, primarily impacted by lower contribution from Global Products. Excluding Global Product, New Markets revenue grew 2.2%. And pleasingly, 2H 2023 was 19% higher than 1H 2023, where we saw strong momentum in China. Additional commentary on revenue is provided in later slides. Over to slide 13. Group underlying EBITDA before the impact of FX was a loss of $20.4 million, impacted by lower gross margin and a proportionally higher cost base coming out of FY 2022. Our operating expenses decreased 11.3% compared to FY 2022. The decrease predominantly reflects the cost out programs completed during FY 2023. However, as expected, the first full-year benefit of these cost out programs will be realized in FY 2024. Global Services division reported EBITDA of $17.5 million, down 68% on the prior corresponding periods. The primary driver was the impact of reduced customer spend on revenue and gross margin that I mentioned previously. New Markets reported an EBITDA loss of $32.7 million compared to an EBITDA loss of $36.5 million in FY 2022. The improvement is due to a higher margin project mix in FY 2023 compared to FY 2022 and some of the benefits of the cost out programs implemented during FY 2023. Slide 14 shows monthly group revenue, underlying EBITDA, and underlying cash EBITDA both before FX. The return to profitability in December, a milestone that was flagged during the recent equity raise, reflects the stabilization of revenue in Q4 and the benefit of cost out programs completed during FY 2023. Turning to slide 15. This slide shows global revenue by quarter, where we saw stabilization in spend in the second half of FY 2023. This was primarily driven by stabilization in spend from a large customer. We benefited from some seasonality in Q4, but not to the extent as previous years. We also saw encouraging growth in LLM-related work in the second half of FY 2023. Over to slide 16. Our China business returned to growth. This follows a protracted impact of COVID, which has now subsided. Pleasingly, the China business had a quarterly record revenue in Q4 with revenue of $11.1 million. China demonstrated significant success winning many LLM projects and is now supporting many leading LLM builders. Total, the China business, which includes Japan and Korea, recorded 82 new customers in FY 2023, including eight customers with multiple LLM deals. Turning to slide 17. This slide details revenue by quarter for the balance of New Markets, being Enterprise, Government, and Quadrant. Pleasingly, there were 89 new customer wins across this group during FY 2023, with the average deal size up 5% for Enterprise to 147,000 compared to FY 2022. There is solid traction in LLM-related work, with work ramping up in the second half of FY 2023. This includes a $4 million+ revenue run-rate project with a leading generative AI model builder. This is an example of a project being delivered through our ADAP platform. We saw growth both across Quadrant and Government, albeit from a small base. For the balance sheet on slide 18. Cash balance at 31 December 2023 was $32.1 million and included the net proceeds from equity raise during the year. The cash balance at the end of January was $34.2 million. The decrease in net assets to $92.8 million was due to trading performance and the $69.2 million pre-tax impairment charge that I mentioned earlier. This is non-cash and relates to Global Services. Non-current assets include $38 million intangible assets relating to Appen platforms, including ADAP. Current liabilities were $17.4 million lower and reflected lower cost of sales and operating expenses. Current liabilities include a $3.8 million earnout liability relating to the Quadrant acquisition. This was settled in January 2024 by the issue of ordinary shares. Non-current liabilities have decreased due to the Quadrant earnout liability becoming current and a decrease in deferred tax liabilities. Turning to the cash flow summary on slide 19. As just mentioned, the cash balance at the end of the period was $32.1 million and included the net proceeds from equity raise during the year. Cash outflow from operations of $23 million reflects the trading performance during the period. Cash has primarily been used to fund operations, while the turnaround is in progress, some CapEx, and one-off costs associated with cost reduction programs. That concludes the financial performance slides. I'll now hand back to Ryan. Thanks, Justin. I'll now talk to our strategy and provide a 2024 outlook statement. Moving to slide 21. Appen has been supporting the development of AI since we were founded in 1996. The industry has evolved a lot since then, and now with the introduction of generative AI, the potential for AI is at an all-time high. As an example, Bloomberg and IDC predict that the generative AI market will reach $1.3 trillion by 2023 at a 42% CAGR. Time will tell whether this plays out, but we are very bullish on the impact of generative AI, and our strategy is strongly focused on capturing value from the market. Moving to slide 22, where I want to provide some perspective as to where Appen fits within the generative AI ecosystem. This slide provides at a very high level the generative AI landscape. I'll step through each of the layers. Infrastructure is a critical layer due to the significant compute power required to train large language models. NVIDIA has seen tremendous success on the back of the demand for compute, as an example. The next layer is the foundation model builders, with the most well-known being OpenAI. These are the companies that are building the large language models. As mentioned, many of these companies are Appen customers. To support the build of the models, there are model tool vendors. These vendors are used by the data science teams in the creation and optimization of the models. The final bucket is the data tools and services layer. High-quality data is required for generative AI, including human-created data. There is growing demand for tools and services for the creation of high-quality data sets, and this is where Appen participates today. When you put all these together, it enables enterprise and industry applications. Moving on to slide 23, where we'll talk about the impact of generative AI on Appen's TAM. On the left-hand side, we see our traditional TAM. This is the data services we provide to companies that are building deep learning models. Analysts estimate that this market will reach $12-$17 billion by 2030. This alone is clearly a large number. Generative AI will expand our TAM in two main areas. The first are the companies that are building the models. This group includes both very large tech companies and a slew of startups. We estimate that we are already working with roughly 80% of the companies, mostly providing data that is used to improve their models. Second opportunity to increase our TAM are from enterprises who are adopting generative AI models into their products and services. Combined, we estimate that generative AI will expand our TAM by $4-$8 billion by 2030. I'll now describe our strategy that addresses both the deep learning and generative AI opportunity. Moving to slide 24, where I'll introduce our growth strategy. There are five pillars. These are aligned to the core customer segments I spoke to on the prior slide. The first element of our strategy is to replatform our core internal system for crowd and project management. Second is the greater focus on using automation in how we create data sets for our customers. The third is a SaaS platform that supports enterprises who are customizing LLMs for internal use. This is one we're very excited about. The fourth is a modernized approach to sales and marketing. And finally is our laser focus on managing costs. I'll now go deeper into each of these. Onto slide 25. We are replatforming our crowd and project management system, which we call Appen Connect. Appen Connect is a proprietary platform, which has resulted in a lot of engineering investment to support. Because it's been built using legacy architecture, we have not been able to innovate as fast as we would like. There are three core benefits of the new platform. First, we are embedding AI as a core capability, including how we manage our crowd. This will add a lot of automation to our internal operations. Secondly, we are shifting from building everything ourselves to utilizing best-of-breed technologies. This will enable much faster innovation and provide access to leading capabilities. Finally, the new technology will deliver a modern experience for our crowd, including how we match them to opportunities at Appen. There's a lot of operational benefit, and it will also reduce our engineering investment, including a lower cloud spend. We are already underway in this project and will be fully migrated by early H2. This is a big unlock for productivity in the business and something we're all excited about. Moving on to slide 26. We're always looking for ways to improve our delivery processes, particularly how we use AI to support the creation of data sets for our customers. We have seen recent success in automating parts of our data workflows, including using LLMs to perform tasks that would have traditionally been performed by our crowd workers. This will continue to be a major focus in 2024. I'll note that not all projects are amenable to automation, and currently, this only impacts a fairly small part of our work. Secondly, we are identifying ways to fully automate data set creation with LLMs. As an example, we are using our crowd to fine-tune an LLM that is then used to create and quality-check data sets automatically. Again, it's early days, but we will continue to push for smarter ways to deliver data at scale. Finally, we're evolving our delivery process with a specific focus on multimodal data. Multimodal is where the LLMs combine language, image, and video. An example of this is the recent Sora model from OpenAI. We have a unique advantage here because we are able to collect, annotate, and evaluate data across speech, image, and video, and many more. This is looming as an emerging demand for generative AI. There are clear financial benefits to automating parts or all of our data set creation, mostly on our margins. And as I discussed, we have multiple projects underway with automation embedded, and we'll continually look for more automation opportunities going forward. Shifting to slide 27, where I'd like to share some details on the SaaS opportunity for LLM. Over the past 6-9 months, we have spoken with many enterprise customers that are looking to solve internal use cases with LLMs. Many people we have spoken to have a backlog of over 1,000 internal use cases, so the demand for generative AI is exceptionally high. There are three things we consistently hear. First, almost all enterprises are looking to adopt existing LLMs rather than build their own. This means that they will need to customize a model to their specific use cases, which requires high-quality internal data. Secondly, enterprises need to invest in getting the data required to customize an LLM. A major part of getting the data is to have internal domain experts working on the data. Closing the loop between the internal domain experts and the data science team is key to building accurate AI. Finally, because this approach is new, there is limited tooling to support the collaboration between AI teams and internal experts. We see this as a big opportunity for Appen to fill. Moving to slide 28, where I'll explain how we are addressing the opportunity. The good news for Appen is that the annotation platform we acquired from Figure Eight is highly flexible and can be configured to support LLM customization. So we already have the bones of a great solution for enterprises. We're building on top of this in three specific areas. First, we are supporting on-premise deployment. Data privacy is very important for enterprises, so having on-prem is an important requirement. Second, we're building custom workflows to support the data needs of enterprises. I won't go into details here, but there's a complex set of data sets needed to customize and test generative AI. We support most of these and are building templates in our software to simplify workflows for the most common use cases. Finally, we are building an insights layer so customers can trace performance and track data lineage. This is an important step to support internal data compliance requirements. Because it's for internal use, the business model is predominantly SaaS and comes with all the financial benefits of a recurring software model. We're currently in alpha phase and are close to signing one of the world's largest telcos as a lead customer. It's an exciting evolution for Appen. It's early days, but we are getting great signals from customers and a confidence of getting product-market fit. I look forward to updating you on our progress here in the future. Now onto slide 29. The fourth pillar of our strategy is to modernize our sales and marketing. Firstly, we now have a strong technical go-to-market team in place led by Andrew Ettinger. Most of our customers are data scientists or technical product managers, so it's important that our customer-facing teams have similar backgrounds. Secondly, we have launched our new website and brand. We're going to get a lot more focus on our expert capabilities in AI versus some of the higher-order messaging from the past. Finally, we have dedicated account managers in place for all of our largest accounts. We expect all of this to lead to more new logos and improved account expansion. Now onto slide 30. The final pillar of our strategy is about financial controls. Justin and I are laser-focused on our costs, ensuring that we are as proactive as possible to adjust our costs to business performance. We did a lot of work to get the Cash EBITDA break-even by the end of 2023 and are keeping much of that rigor as we go into 2024. Onto slide 23, oh, sorry, onto slide 31, where we have the go-forward management team. I'm super happy with the leadership team. It's a great mix of tenured executives, new additions from adjacent markets, and the promotion of some new leaders with deep operational expertise. This is the team for Appen going forward. I'm not planning to make any changes. So that concludes our strategy section. I'll now shift gears into the FY24 outlook. Before I get to the outlook, on slide 33, I'll share some relevant market observations. As discussed, FY23 was a transitional year for the AI market. Many companies experimented with and evaluated the potential of generative AI. The experimental period, along with a broader focus on cost optimization among tech customers, resulted in challenging market conditions, particularly in the first half of FY23. In the second half, we saw experiments turn into scaled operations, and we expect this to continue into FY24. And as shared earlier, generative AI is expanding our TAM, and we see the next two years a transition period for the market. This year, we are providing an update on our January trading performance on slide 34. Note that these numbers are for the group, and the January numbers are based on unaudited management accounts. In January, we recorded revenue of $22.7 million, underlying EBITDA of -$0.1 million, and underlying cash EBITDA of -$0.2 million. So while it's not quite break-even, it's certainly headed in the right direction. Now turning to the outlook on slide 35. As we noted in the presentation, the revenue declined from a large customer stabilized, and based on our conversations with them, we expect this trend to continue into FY24. We spoke about new product development. Please note that costs associated with the development of new products will be contained in the current product and engineering spend. We put in place tight financial controls in FY23, and these will continue into FY24. In FY24, we will get the full benefit of the $60 million cost reduction program we implemented in FY23. As part of the Google announcement, we stated that we are implementing a further $13.5 million in cost out. We remain on track to deliver these savings in FY24. Finally, ongoing cash positivity is a high focus. Our ability to achieve cash EBITDA profitability in FY24 will be dependent on revenue growth from our non-global customers, the timing of which remains uncertain. So this concludes the presentation. It's an exciting time to be in AI, albeit during a challenging period for Appen. Despite the challenges, we are very bullish on generative AI. And because of our deep expertise in capabilities and AI data, we are well-positioned to capture growth. However, profitable growth is a top focus for Appen. This means that we must invest wisely and execute flawlessly. I'm confident that we have the team and the expertise to do both well. Thank you. I'll now hand back to Andrea, who will open the call for questions. Thank you. If you wish to ask a question, please press star, then one on your telephone and wait for your name to be announced. If you wish to cancel your request, please press star, then two. If you are on a speakerphone, please pick up the handset to ask your question. Our first question will come from Josh Kannourakis of Barrenjoey. Please go ahead. Hi, guys. Can you hear me okay? Yep, we can, Josh. Cool. Just Just first one, just wanted to clarify on the revenue and into the January update that you provided, just how much of did you still have flow on work from Google, or was that sort of the end of it at that point? And just remind us of the wind-down period for contracts, please? So it does include Google work. The wind-down is March 19. So yeah, it does include Google. Okay. No, that's helpful. And then I guess just in terms of the generative AI base and the revenue opportunity there, you gave some good context around some of the more material contracts and sort of run rates that you mentioned there. But could you give us just a little bit more context around the sort of total revenue pool today, but also, I guess, if things go to plan, how you think that could how that can grow as a sort of percentage of the business over this next 12-month period and then sort of medium-term? Yeah. So the reason we focus a little bit more on run rate is because we are seeing growth. So it was very quiet in the first half of the year. We saw growth in the second half, and that's continuing to ramp. So that's why internally, we look a little bit more on the run rate. What that could look like going into next year, it's a little bit tricky to say at the moment, but we are seeing and we're hearing the demand for international expansion in the LLMs. A lot of them have been U.S.-centric to start with. And we've seen this play out in other AI types of models before where it starts U.S. and then expands into international markets. But at this stage, it's difficult to predict what that revenue will look like. Got it. Understand. And then just in terms of, I guess, R&D or investment in this FY2024 period, you've obviously talked to cash EBITDA profitability, which I guess includes the capitalization and the like as well of R&D. But could you give us a bit of a feel for if we look at the different sort of cost line items, how you sort of see, I guess, where some of that additional sort of $13.5 million of cost is coming from, and then just how we should be thinking about the approximate spend on the different line items in terms of sales, in terms of R&D, etc.? Yeah. I'll talk a bit to the cost out. Look, we focused, firstly, on the direct and indirect costs related to Google. So there was an internal team who were focused on Google as an account. But there was also a lot of people supporting Google. It's a pretty complex project, particularly with the onboarding and the identification of the crowd workers. So there's a lot of simplification that comes with the Google work from our operations all the way through to our product and engineering investment. Hi, Josh. And just on the platform development side, so we obviously haven't given specific guidance on dollars there or percentage of revenue. But what is fair to assume is, and there's a slide in the appendix in the investor presentation, is that the spend for 2024 will be lower than 2023. I think that's a fair assumption. And that's based on the phasing of the cost out during 2023 was more towards the end of the year versus the start of the year. With our platform development, I think what's important to note is it's been more of a transition to lower-cost resources versus a number of people necessarily. So we are still investing in that space, but it will be contained within the existing product and engineering spend on the load. Okay. That's really helpful. And then just final one from me, just more so on the competitive dynamics and what you guys are seeing. I mean, firstly, just with regard to the Google contract, do you know where some of that work has gone to competitively, and have you had any sort of further feedback on that process? And then maybe just secondly, in terms of a lot of your work around the LLM space, just maybe to give us a feeling for how you feel you guys are positioned from a competitive dynamic and who some of the key players are that you're seeing in the market at the moment. Thank you. Yeah. So the rationale for the Google decision remains unclear. In some of the projects where we're doing handovers, it's to the typical competitors that we see at Google. But in other areas, we don't have the visibility of the handover. So it's a bit of a mixed one on where the work is going to, to be honest. In terms of generative AI, we're seeing a lot of strength that we have versus competitors in the LLM space, particularly as the demand for high-quality data ramps up. I kind of called out in the call that for the LLM model builders, quality data is really high importance to the point where quality data for generative AI is more important than it was for deep learning-based approaches. What we're finding is that some of our competitors can do quality at smaller scale, but when they try to ramp to larger scale, that's where there's some challenges. It's an area where we have a lot of track record and a lot of expertise. So we're pretty comfortable with our position and confident vis-à-vis our competitors at the moment. But there's always more work to do, and we're laser-focused on delivering high quality for our customers. Our point of view, if we deliver quality data that's higher quality than our competitors, then we will win the work. So we're more focused on getting that high-quality data to our customers. All right. Thanks very much, Ryan. Thanks, Josh. Appreciate your time. Thanks, Josh. The next question comes from Wei Sim of Jefferies. Please go ahead. Hi, Ryan. Hi, Justin. Hi, Rosalie. I've got three questions. The first one is just in regards to our non-global customers. Are you able to give us a bit of color as to what kind of pipeline and visibility we have on these potential customers? Thanks. Yeah. So we don't disclose the pipeline, as you'd imagine. But we have some visibility, and this expands both into the expansion of existing accounts and new customers in the pipeline. So we continue to focus on growing our non-global customers. Okay. And just in terms of, I guess, the visibility, is it something that over the next 3-6 months, or what kind of lead time would you have in terms of that visibility, just to give us a bit of sense? So look, it depends a little bit. For existing customers that are working with us on longstanding projects, we've got visibility now. There are customers that are looking to do new projects that we've got very good visibility on. But we also, particularly in the LLM space, often it starts with a small pilot, and then it ramps very quickly. The extent of that ramp can depend on a lot of factors. So we are in with a lot of customers at early levels that the demand can ramp quite quickly. So it's a little bit variable, I'd say. Okay. Got it. The second question is just you mentioned before Figure Eight, Ryan. I was just wondering if you could talk to some of the benefits of that platform that we would be looking to leverage into, I guess, future endeavors? Yeah. Sure. So I'll just start by noting that a lot of the LLM work that we're doing today is being delivered through the Figure Eight platform, which we call ADAP internally now. So it's proven super beneficial for the work that we're doing in the LLM space. What we're going to be doing is customizing the Figure Eight platform for internal use by enterprises. So for enterprises who are adopting LLMs for their internal use cases, they need to access their internal network of experts to get data to build the use cases. So as an example, if a bank is building a model that is going to assess credit risk on individuals, you don't want the LLM doing that by itself. You want to make sure that you're getting feedback and data from your credit experts within the bank. Now, connecting data science with the credit risk team is a perfect use case for the ADAP platform. So you can think about it like the credit risk would be analogous to a crowd worker in our current model. So what we're going to be doing is what we've got ready today is an on-premise deployment of ADAP where our data science teams data science teams of our customers can set up jobs and allocate work to the internal experts, and they get all the benefits of a robust platform with data traceability and auditability. So they've got a full track record of where the data is being created, by who, and at what time. Understood. Okay. Cool. No, that's really good. And then just the last question is Appen Compass. I was wondering if that's something that is still in development or if there's any updates that you might be able to provide on that. Yeah. So Compass, the strategy behind Compass has evolved into this internal platform that I just spoke about. So we'll be building some of the features that we showed at Compass into this platform. Okay. Perfect. That's all from me. Thank you so much. Thanks, Wei. There are no further questions at this time. That does conclude our conference for today. Thank you for participating, and you may now disconnect.