Embracing Innovation and Making AI Work for Healthcare

Welcome to our Staff Spotlight where we introduce our great team who work tirelessly to support ACHS Members and the broader healthcare community.

In this edition, we meet Ujas Patel, ACHS’s Artificial Intelligence (AI) Lead, to explore how he is helping shape a practical, responsible approach to AI at ACHS.

Can you tell us about your role as Artificial Intelligence (AI) Lead at ACHS? 

My role is to guide and implement ACHS's AI strategy with a roadmap that is practical, achievable, and safe. In practice that means three things: finding where AI creates genuine value for our Members, finding where it makes us more efficient internally, and being clear about where it doesn't belong. 

A lot of my focus over the past few months has been on governance. That means establishing how ACHS assesses the risk of an AI system before it's used, how we classify and protect the information that goes into these tools, and who is accountable for the outcome. Governance done well doesn't slow people down, it helps them move faster by making clear which tools are safe, what information they can use, and where the boundaries are.

Capability building is a key part of the role. Strategy only works if people can act on it, so I spend much of my time helping teams understand what these tools can do, where their limits are, and how they fit into existing workflows. I also stay close to the delivery itself because being hands-on keeps the strategy honest.

What inspired you to pursue a career in AI, and what impact do you hope to make through your work?

I studied neuroscience alongside computer science, and I was drawn to neural networks, largely because of their links to the brain. I've spent a long time since then working at that intersection, and it's striking now to see those same ideas play out at a scale nobody could have imagined back then.

What excites me most about the current generation of technology isn't the capability itself, it's who gets access to it. These tools have become an enabling layer that people can use without being technical. You can now understand a system, interrogate it and build with it in plain language rather than code. That makes capability far more equitable than it was.

The impact I want to make is about enablement: helping people who aren't technical, particularly leaders and the organisations they run, understand what these tools genuinely do and where they fit against the problems they're trying to solve. For ACHS specifically, that means using AI to serve our Members better, whether through accreditation services, training or the way we support them day to day, so they can keep monitoring and improving quality and safety in their own services. 
 
What has working in AI taught you about how people respond to change?

The deciding factor is rarely the technology itself. It's whether the tool fits the way people already work and whether it genuinely solves a problem. If it sits on top of an existing workflow as one more thing to learn, nobody in their right mind will adopt it, and they shouldn't. That's especially true when the current process works fine. The bar for replacing something that works is high, and it should be.

I've stopped framing things as "here's an interesting way to do this" and started asking what I can take off someone's plate. That framing changes the conversation entirely. The other thing I've learned is that the stated concern is often not the real one. People say they're worried about AI accuracy when they're actually asking who is accountable when the output is wrong. That's a fair question and it deserves a direct answer, which is why a person reviews the output in everything we've built.
 
What's one conversation about AI that you think leaders should be having more often?

What happens when AI gets it wrong. There's a lot of discussion about governance and speed, but far less about failure modes, and that's the conversation that determines whether an organisation deploys AI safely or finds out the hard way.

It's worth being clear-eyed about what these systems are. They're pattern recognition at a remarkable scale, and they can reason impressively, but they can also be confidently wrong. That isn't a reason to avoid them, it's a reason to design for it. The useful version of the conversation is specific: if this produces a wrong output, who notices, how quickly, who is accountable for the decision it informed, and is the error recoverable? 

Governance is still catching up to the technology, and I think healthcare leaders sometimes feel defeated by that gap. The answer is to be proactive rather than wait for certainty. Put a risk framework around it, define the controls, decide in advance what you'll do when something fails. Healthcare has spent decades building exactly this discipline through clinical governance, incident management and open disclosure. We already know how to do this. The task is extending it properly, not inventing it from scratch.
 
Looking ahead, where do you see the greatest opportunity for AI to create positive change in healthcare?

This is a hard question to answer narrowly, because AI is an enabling technology rather than a single application. The closest comparison is computing itself. Computers have transformed every industry and every role.  AI is the next layer of that, with the difference that it can reason and interact with us in far more ways than typing and clicking. Now we can talk to computers, show them our world, and instruct them to take actions in natural language rather than structured code.

Within healthcare, I am watching diagnosis and drug discovery particularly closely. In screening and imaging, with proper clinical oversight, systems that can analyse scans at speed mean patients are diagnosed and treated sooner, and that's a genuine improvement in outcome rather than just throughput. I’m also interested in new diagnostic approaches and AI-driven drug discovery, particularly for rare and complex diseases.

The opportunity for the healthcare sector that I find most compelling is giving clinicians and quality teams their time back. A significant share of healthcare work is documentation and reporting, and this is an area where AI has proven genuinely capable, drafting summaries, surfacing gaps, pulling information together. The test I apply is whether a use case makes us faster or genuinely better. Both are worth having, but in healthcare services only the second changes patient outcomes, and that's where the effort should go. It also raises the questions that matter most in this space: who owns the output, and who is accountable for it.

Thank you, Ujas, for sharing your insights and experience with us. We look forward to bringing you further Staff Spotlight articles in the future and showcasing the incredible talent that we foster at ACHS.