The Invisible Infrastructure: Corti’s Andreas Cleve on Healthcare’s AI Revolution

Corti now supports around 250,000 patient interactions every day across multiple countries. From your perspective, what does this scale reveal about the readiness of healthcare systems to adopt AI in frontline care?

Supporting 250,000 patient interactions every day tells us something important: the appetite for AI in healthcare is real. But there’s still a long way to go. A recent YouGov study found that more than 80% of European clinicians want AI to ease their workload, yet only about a quarter actually trust the tools available today. Globally, the picture is even tougher: MIT estimates that 95% of AI projects never make it past pilot.

Breaking out of that cycle requires infrastructure built specifically for healthcare: systems that understand medical language, workflows, compliance, and context from day one. That’s exactly what we do at Corti. It’s the gateway to scaling quickly, earning trust, and moving from pilots to real impact. And we’ve seen that once clinicians use AI that truly works for them, trust follows fast. With billions of people worldwide still lacking access to care, scaling safely at this pace isn’t anywhere near the finish line – but it is a significant step toward closing that gap.

General-purpose AI models are making headlines globally, yet you argue they cannot meet healthcare’s hyper-specialised needs. Can you explain why “clinical grade” AI infrastructure is fundamentally different?

General-purpose AI is extraordinary at many tasks, but healthcare is not a general-purpose domain. It’s highly specialised, deeply regulated, and the stakes are life and death.

For AI to scale in this environment, it has to master thousands of clinical niches – the language, reasoning, and workflows of each specialty. A consumer model can be ‘good enough’ for drafting an email or testing a new haircut: fun but often flawed. A clinical model must be precise, auditable, and continuously aligned with medical standards. That’s why we build clinical-grade infrastructure: systems trained for healthcare contexts, designed for compliance from day one, and robust enough to stand behind in a hospital. It’s not all about power – it’s about the right power, applied in the right context.

Corti is active in both Europe and the US, where regulatory frameworks and healthcare models differ greatly. How do you adapt your technology to work across such varied systems?

In healthcare, you don’t get to choose between speed and safety – you need both. Europe emphasizes stewardship, compliance, and patient protection; the U.S. prioritizes speed, liability protection, and ROI. Corti is built to operate in both worlds. Our infrastructure has GDPR and HIPAA compliance baked in, with audit trails, sovereign cloud, and clinical alignment ready from day one. That means our partners can scale with confidence across borders, without rebuilding for every new market. Flexibility is what makes trust scalable – whether you’re proving yourself hospital by hospital in the U.S. or navigating national regulators in Europe.

Frontline clinicians are often overburdened and under immense pressure. How do you ensure Corti’s tools integrate seamlessly into their workflows rather than adding complexity?

The first rule is: don’t add to the burden. A YouGov study early this year showed more than 80% of European clinicians want AI to ease their workload, yet many are still spending up to three hours a week correcting outputs they don’t trust. That’s not the future anyone was promised from AI.

To fix it, we take on the hard work and complexity at the infrastructure layer, so it doesn’t spill further down – whether that’s wasted cycles for the builder, or frustration for the clinician at the point of care. Corti is designed so that the developers and providers building on us deliver tools where trust, accuracy, and clinical relevance are built in from day one – instead of leaving clinicians to fix mistakes after the fact.

We think of it as the factory floor of healthcare AI: the place where compliance, interoperability, and auditability are industrialised at scale. When those foundations are handled, developers and providers can focus on creating experiences that fit seamlessly into clinical workflows. That’s how adoption happens – when AI gives clinicians time back rather than taking more away.

Corti has raised over $100 million from high-profile investors like Atomico and Prosus Ventures. How are you prioritising the use of this capital, and what growth areas are you focusing on next?

We’re doubling down on what we do best: infrastructure. Our focus is on making Corti the most reliable place in the world to build healthcare AI – strengthening the API, expanding our healthcare-specific models, and continuing to strip away the administrative work that nobody came into medicine to do.

We’re also starting to see some very exciting developments on the horizon – more agentic breakthroughs, more automation of complex workflows – and our job is to make sure those capabilities are safe, compliant, and usable so that developers can scale them in real clinical settings. We’re not chasing every shiny opportunity. We’re industrializing the foundations so others can innovate confidently on top. That’s where the real leverage comes from.

With AI adoption in healthcare comes understandable concerns about trust, safety, and transparency. How does Corti build confidence among patients and clinicians that its AI can be relied upon in life-critical situations?

In healthcare, even a 1% error rate can cost lives. That’s why trust has to be engineered into the system, not bolted on afterwards. Corti has always been built on research and scientific validation. Our models are trained on millions of hours of clinical dialogue, validated in peer-reviewed studies, and continuously stress-tested in live deployments. Every interaction is supported by audit trails, real-time quality checks, sovereign cloud options, and alignment with medical standards.

For hospitals, that means they don’t just get an AI system – they get the evidence and safeguards to prove its reliability. For clinicians, it translates into confidence at the bedside. And for patients, it means AI that is held to the same bar of accountability as any other medical technology. Trust isn’t an aspiration – it’s something we measure, publish, and improve every day.

Looking ahead, what role do you envision healthcare-specific AI infrastructure like Corti playing in reshaping the global patient experience over the next 5 to 10 years?Success for us will be to be invisible. The most impressive technologies – like electricity or the internet – disappear into the background because they’re so reliable you don’t have to think about them. Healthcare-specific AI infrastructure should be the same: the invisible scaffolding of global care.

In five years, patients won’t be asking whether AI is in the room – they’ll just experience shorter waits, faster diagnoses, and more time with their doctor. In ten years, the most advanced systems will be those where AI quietly handles the complexity: compliance, documentation, reasoning support, interoperability.

That frees human caregivers to focus on empathy and judgment – the reasons they entered medicine in the first place. And as this infrastructure extends into underserved areas, the impact is even bigger: making safe, trusted care accessible to billions who don’t have it today.

The post The Invisible Infrastructure: Corti’s Andreas Cleve on Healthcare’s AI Revolution appeared first on European Business & Finance Magazine.