90: HealthTech Hour with Steve Roest - Episode 9090: HealthTech Hour with Steve Roest - Episode 90
UK Health Radio Podcast
Steve Roost talks with AI specialist Robin Matthews about how agentic AI is reshaping software development in health technology while keeping clinical safety central. The conversation touches on regulation, cybersecurity, and how tech roles and mindsets are adapting to rapid change.
49:41•17 Nov 2026
Agentic AI, HealthTech, and Changing Roles: Steve Roost Chats with Robin Matthews
Episode Overview
- Agentic AI can dramatically increase software delivery speed, clearing long-standing bugs and technical debt while keeping humans in control.
- Large language models now perform a high proportion of tasks comparable to senior engineers, turning natural-language problems into functioning code.
- In regulated health environments, AI is best applied first to non-clinical parts of the product, with a risk-managed approach to adoption.
- Many concerns about AI in health stem from confusion between using AI for software development and using it for clinical decision-making.
- Governments and AI developers need collaborative approaches to handle new security risks, especially from powerful open-source models.
“My role is changing and I've got a whole new set of stuff to learn.”
Curious about how others navigate their sobriety journey? While this instalment focuses on health technology rather than alcohol directly, it offers plenty of food for thought on how fast-changing tools can affect any regulated field, including addiction care. HealthTech Hour, hosted by Steve Roost, brings on Robin Matthews, a "battle-scarred technology leader" and Agentic AI pioneer, to unpack what modern AI actually does in healthcare product development.
Robin explains large language models in straightforward terms, saying they’ve jumped from simple word prediction to systems that can perform “over 80% of the equivalents of a really experienced senior engineer” when turning human-language problems into code.
You’ll hear how PocDoc, the company where Steve is CEO and Robin is Chief Technology and Product Officer, uses agentic AI as a kind of software factory: agents spot bugs, propose fixes, and help rewrite big chunks of code at speed, while humans still review and sign off the results. Robin notes that this has cleared long-standing technical issues “pretty much eradicated… overnight,” but stresses that human oversight and clinical safety remain non-negotiable. They also tackle scepticism.
Many health leaders jump straight to fears about AI making clinical decisions or mishandling patient data. Robin calmly separates that from the current reality: the focus today is on AI helping to build safer, more robust tools, not replacing clinicians. He also flags emerging risks, like powerful open-source models that could supercharge cyber-attacks, and calls for governments and frontier AI developers to work together rather than rely on static rules.
Underneath the tech talk sits a simple theme: roles are changing, and the people who thrive will be those who treat AI as a new set of tools, not a personal threat. As Robin puts it, it’s about “every day just taking a step forward and learning something new.” If health, regulation, and future-facing tech matter to you, how are you preparing for this shift?

Do you want to link to this podcast?
Get the buttons here!
More From This Show
The latest episodes from the same podcast.
