ImExHS Limited (IME) Earnings Call Transcript & Summary

Good morning, everyone, and thank you for joining us. My name is Dr. German Arango, CEO of IMEXHS. Joining me today is Fabio Carrillo, our new Chief Financial Officer. Today, we will review our H1 FY '26 results, the progress we have made with our AI platform and the areas where we still have work to do. Then we will open the line for questions. The headline is straightforward. 12 months ago, our AI agents were a plan. Today, they are running in production in real hospitals on real patient volumes. This is the most important change in the business and much of today's presentation connects back to it. On Slide 2, we are one company with 2 complementary businesses. The first is Software, cloud-based medical imaging software that radiologists use every day to manage view and report studies. Since April, this has also included our own AI agents embedded directly into the workflow. We sell both directly and through our partner network across Latin America. The second business is Radiology Services through RIMAB. Here, we operate Radiology Services ourselves, outsourcing teleradiology clinics and mobile units, mainly in Colombia. The strategic rationale for having both businesses is clear. The AI agents built by our software team can run first inside our own radiology operation. We test them against real volumes, real workflows and real cost before taking them to customers. Very few companies in our industry have that advantage. Slide 3. That connection also gives us something that is very difficult to replicate, data. Our agents are trained on workflow data from our own operations and client base. The scale is significant, close to 2 billion new images in the last year across more than 560 sites, nearly 3 million people use our patient portal, up 60%. And we now work with 47 partners, up from 25. We now have 8 proprietary AI agents. Importantly, we are not simply licensing somebody else's algorithms. We build these agents ourselves using data generated within our own ecosystem. Turning to Slide 4. A quick word on our footprint. We operate across Latin America and hold the regulatory approvals required in our markets, including FDA clearance. In the United States, an ISO 27001 certification for Aquila+ and our development process. In health care, certification is not just paperwork. It's a real barrier to entry. It takes time, investment and experience to achieve, and it makes our position harder to replicate. On Slide 5, the half can be summarized in 3 points. First, we moved from launch to production. The agent platform went live in April. And by the end of June, it was already running inside the public hospital network. We launched with 5 agents. We now have 8. Second, earnings improved materially. Underlying EBITDA was well ahead of the same period last year. Radiology performed ahead of plan. Software was broadly in line and revenue grew in both reported and constant currency terms. Third, our commercial execution was uneven. We delivered a strong large public sector win, but the flow of smaller, more repeatable deals was not as consistent as it needs to be. So the commercial transformation is progressing, but it is not complete. The actions already taken to address this will be covered later in the presentation. Turning to Slide 6 on the financial results. Sales revenue was $16 million, up 17% and up 13% on a constant currency basis. So the underlying business grew. This was not simply a currency effect. ARR was $36.8 million, up 12% as reported. On a constant currency basis, ARR was essentially flat year-on-year. Underlying EBITDA was $1.7 million (sic) [ $1.27 million ] around 4x the level of the same half last year. This is the operating leverage we have been talking about now starting to appear in the numbers. Cash was $2 million, down from $3.3 million at December. Importantly, that reduction reflects working capital timing, not trading losses. We remain profitable at the EBITDA level. Our largest customer delayed a payment during the half and settled it the following month. That illustrates the environment we are managing in Colombia, which I will discuss in more detail later. Debt remains very low. Slide 7 shows where that performance came from. Radiology was the main driver. Revenue was $11.2 million, up 24%. ARR increased 20% and EBITDA reached $1.1 million compared with $0.3 million a year ago. That is the turnaround showing up in the numbers. Software delivered $4.8 million of revenue and $11.8 million of ARR. Software EBITDA was $0.8 million, down from $1.3 million. So software earnings moved backwards in the half. There were 2 main reasons: currency movements on contracts priced in U.S. dollars and investment ahead of revenue as we build the agent platform. Corporate costs decreased from $1.3 million to $0.6 million. So the group earnings improvement came from radiology and from cost discipline. The next step and the one that matters most for the value of IMEXHS is getting software growth to follow. Next slide, #8. This chart shows ARR over the last 5 years. The dark bars show reported ARR. The green bars remove the currency effect, so we can compare the underlying trend like-for-like. On a reported basis, ARR increased from $32.8 million to $36.8 million. On a constant currency basis, it was essentially flat year-on-year. Currency is part of that picture with more than 70% of our software contracts now priced in U.S. dollars. The key point is that our recurring base remained stable, while the platform we needed to drive the next stage of growth only went live in April. The product is now different. And the next slide shows revenue that is already contracted, but not yet included in these numbers. Slide 9. Signed revenue and billing revenue are not the same thing. At the end of June, we had $4.4 million of ARR already contracted, but not yet billing. That represents 12% of total ARR. 6 months earlier, that number was $1.7 million and 12 months earlier, it was $2.1 million. So it has more than doubled year-on-year and increased significantly over the last 6 months. This is contracted revenue already in implementation. It is not pipeline and it is not forecast. It is roughly split between radiology and software. Alongside the flat constant currency ARR, it is important to consider that a meaningful part of the revenue that should convert into future ARR has already been signed. Turning to Slide 10. On the platform, this is what we mean when we say our agents work across the full workflow. They can triage urgent studies, detect findings, structure reports, check quality and coordinate the process end-to-end. The important point is that these are not isolated tools. They operate together as one system inside one platform. Before I go deeper into the product, this is the strategy the whole company is now aligned around 5 priorities: an AI-first product strategy, one consolidated cloud-native platform in Aquila+, protecting and expanding our existing customers, profitable growth, not growth at any cost and simplifying how we operate. The fourth priority is particularly important. We are managing this business so that growth and profitability come together, not one at the expense of the other. That discipline is one of the reasons EBITDA improved this half, even while software growth remained modest. Turning to Slide 12. 5 agents at launch in April, 8 by the end of June and already operating in One live public hospital network. The following slides show how they operate. But before, let me tell you why this matters commercially, and it's for a clear reason. Imaging demand continues to grow, but the number of radiologists does not grow at the same pace. Health systems around the world are dealing with that gap. And this is the part that is often misunderstood. Much of the cost delay and operational error is not in the image interpretation itself. It is in the -- in everything around it, scheduling, triage, study distribution, monitoring and follow-up. The industry has generally responded by adding AI at individual points in that process, one algorithm for a scan or one assistant for a report. That can help, but the underlying workflow remains largely unchanged. We took a different approach. We built agents directly into the workflow so they can manage and resolve many of those operational tasks within the process itself. As far as we are aware, no competitor currently addresses the workflow in radiology this comprehensively. On Slide 14, these are the first 4 agents. I won't go through every one of them, but 2 are particularly important. The scheduling agent reads medical orders and insurance authorizations, applies each institution's rules and books the appointment automatically. In testing, it has reduced call center staffing requirements by up to 80% and cut patient waiting times by roughly 50%. Development is complete, and its first production deployment is planned for this half. The distribution and monitoring agents are already live across our client base. Monitoring tracks every study in the queue and flags cases at risk of missing their deadline. Distribution can then rebalance the workload across radiologists to address the problem. That combination is important commercially, and I will explain why in 2 slides. On Slide 15, these are the other 4. The reading copilot is exclusive to Aquila+. It improves detailed reports and brings patient and study context directly into the reading workflow. Hanging protocols are complete and embedded in our new Web Viewer. The line at the bottom gives the development status. All 8 are our own intellectual property. We build them and we own them. On Slide 16, so why can this model compound over time? There are 4 reasons in roughly the order they can impact the P&L. First, cross-sell. We already have 569 installations that are contracted, integrated and training. Every new agent give us an additional product to offer into that installed base without having to acquire a new customer. That is highly efficient revenue. Second, it changes the basis on which we compete. When a tender is evaluated on operational outcomes rather than only software specifications, the discussion moves beyond the line-by-line price comparison. Third, switching cost increase as each agent becomes more embedded in how a customer operates, replacing the platform becomes much more than a data migration. It becomes an operational change. Fourth, unit economics improve. Less manual effort is required per study, both for our customers and inside our own radiology businesses. That allows volume to grow without cost increasing at the same time. And there is one commercial mechanism I would ask you to remember. Monitoring agent makes an operational problem visible to the customer. Then distribution agent helps solve that problem. In that sense, one agent can create the demand for another. Turning to Slide 17 on the software business. Software ARR was $11.8 million on revenue of $4.8 million. On new businesses, our largest win was the Zacatecas public tender in Mexico, adding $348,000 (sic) [ $384,000 ] of new ARR. We also won Pulso Salud in Peru, CESAC in Colombia and Hospital La Misericordia also in Colombia as well as new multisite wins in Venezuela, El Salvador and Peru. Our installed base is now 569 live sites, 20 more than a year ago. And every new contract now deploys on Aquila+. We have also rebuilt the sales leadership and team and introduced a stage gate qualification and lead scoring. These are not cosmetic changes. They are designed to address the commercial inconsistency I mentioned earlier. On Slide 18, the core platform also continued to improve while we were building the agents. There are 3 things I would highlight. Aquila+ is commercially live and every new contract deploys on it. We hold ISO 27001 certification for both the platform and our development process, which is particularly important when selling to hospitals and governments. And implementation is now 60% faster than with our previous version. That last point is important commercially, not just technically. Faster implementation means we can start billing sooner, which directly affects the contracted revenue I showed you earlier. Turning to Slide 19. The partner program remains our main route to market. And this half, it demonstrated something important. We now have 47 partners across 13 countries in Latin America, up from 25 a year ago. Around 75% of our new software ARR in the half came through partners. Zacatecas is an important proof point. It was an enterprise scale public sector contract delivered through the channel. One fair question about the partner model was whether it would -- it could support large transactions as well as mid-market volume. Zacatecas demonstrates that it can. Two new partners joined in Q2 and both generated new recurring revenue in the first active month. That is a faster ramp than we have achieved previously. We also completed a full anti-money laundering and counter financing review across the active partner network and every active partner passed. Slide 20, on commercial execution. This is probably a slide I most want you to understand. We wrote $1.1 million of new software ARR in the half. That is below where it needs to be, and I am not going to explain it away. The issue is the pattern. We proved that we can win a large transformational contract, but underneath that, the flow of smaller deals was inconsistent. We cannot depend on occasional large wins. We need a steady and repeatable mid-market engine. That remains our main commercial challenge. The important point is that the changes are already in place. New sales leadership, a rebuilt team, stage gate qualification and lead scoring, tighter forecasting and a cleaner model for repeatable deals. Marketing generates the opportunity and partner converts it. Pulso Salud in Peru is a good example of that model. Our priority for H2 is simple, turn proven large deal capability into repeatable performance across multiple territories. Slide 21 on radiology services. This business is now a meaningful contributor. Revenue was $11.2 million. ARR was $25 million and EBITDA was $1.1 million. We operate across 38 sites with 157 radiologists and perform around 90,000 procedures each month. The turnaround has 2 stages. Stage 1 was margin, repricing cost discipline and automation. That stage has been delivered ahead of plan despite a sector environment that became more difficult, not less. Stage 2 is working capital, and that work is underway now. In practical terms, the priority is collections. I will explain the sector context on the next slide. Two new contracts, Sanitas and Colsanitas, are expected to start in Q3 and add approximately $2 million of ARR between them. And one final point, our own AI agents run inside these operations, particularly where cost and operational complexity concentrate. So RIMAB is both a customer of our technology and a proving ground for it. Slide 22. This is the context behind the cash position. In Colombia, government payments to health care insurers have been slow, and that pressure flows down to providers like us. That continued through the half and through the election period, which concluded with the June runoff. With a new administration coming in, we are seeing some early improvement in the sector sentiment, but I want to be careful with the wording. Better sentiment does not yet mean better payment. This pressure has built over several years, and it will take time to unwind. So we are managing the business conservatively. We have tightened credit controls, maintained conservative pricing assumptions, increased our focus on working capital and selectively exit customers in financial difficulty. Our largest customer by revenue delay a payment during the half and settled it the following month. Our controls are designed precisely to manage that type of volatility. Slide 23, briefly on the leadership team. Fabio Carrillo joined us as Chief Financial Officer on August 1. He is a Chartered Accountant with 20 years of experience in public practice, supported by a strengthened finance team in Colombia. I also want to thank Reena Minhas, who has resigned to take up a new opportunity for the financial governance and reporting discipline she has helped build at IMEXHS since 2020. Earlier in the half, Daniel Laverde joined as Chief Operating Officer; and Mario Huyo as Sales Manager. Both bring experience in scaling technology businesses in Latin America. That is the team responsible for delivering H2. Slide 24. So what are our priorities for H2? In software, convert large deal capability into repeatable revenue across territories, start building the contracted book, put the scheduling agent into its first production deployment, release the prioritization agent, develop Phase 2 of the copilot and deepen the partner network in Mexico and Peru using Zacatecas as a reference case. In radiology, collections and working capital come first. We also need to protect the margin we have built, start the Sanitas and Colsanitas contracts and continue improving service quality. And for the group, deliver the guidance. That is a short and specific list by design. I would ask you to judge us against those priorities at the full year result. Slide 25. Finally, our FY '26 guidance. We expect revenue between $31.4 million and $33.7 million, representing growth of 8% to 16%. We expect underlying EBITDA between $2.4 million and $2.7 million, up 48% to 66% on the prior corresponding period. At the lower end, that implies an H2 broadly similar to H1. At the upper end, it implies a stronger H2. On earnings, the guidance assumes that we sustain the improvement already delivered rather than build in a significant further acceleration. We have set the range deliberately. We still have contracted revenue to convert and the collections environment in Colombia remains outside our control. We would rather provide guidance that we believe we can deliver. Let me finish with 4 points. Our AI platform is real. It is live and it is ours. Earnings have improved materially. Our recurring revenue base is stable with $4.4 million of contracted ARR still to convert. And our software sales engine is not yet where it needs to be. We recognize that, and we have already made the changes to address it. Thank you. Fabio and I are happy to take your questions. So Fabio, do we have any questions so far?

At the moment, we don't have anything on the question chat, but we probably should open it for anyone else that has any questions that hasn't put in the chat.

Let's wait one more minute to see if anybody has a question. Okay. We have no questions. So we will close the session today. Thank you, everybody, for joining us this morning.