Responsible for more than 80,000 employees, 1.3 million patients, and an estimated annual research spend exceeding $1 billion, Mayo Clinic is an organization where human expertise is core to what makes it special.
It’s been one of the world’s leading healthcare groups for more than a century and now, it’s leading the way for how to augment AI into a human-centric business where trust really is everything.
AI transformation isn’t just a technology decision – it’s about the kind of organization you want to become, and something we explore elsewhere in our article on how agentic you want to be.
The Mayo Clinic’s history is testament to a culture where innovation is the norm, the expectation. One of its most ground-breaking inventions in its history – StateViewer – is the culmination of it all.
Vintage Minnesota Postcard - The Mayo Clinic By Night" by Joe Haupt is licensed under CC BY-SA 2.0
Mayo Clinic’s value lies in a deeply human model of collaborative clinical expertise, but that model is hard to scale through people and buildings alone.
In an interview with Microsoft’s AI Revolution in Medicine podcast, Dr Gianrico Farrugia – Mayo Clinic’s CEO and President between 2019 and 2026 – says the shift started in late 2018, with a clear focus on agentic AI, rather than trying to create the structure and administrative framework to build up to it.
The aim was to provide patients with cures: “If healthcare was perfect, we would wait. Healthcare is not perfect by any means, therefore let’s run and embrace AI.”
AI offered Mayo Clinic a new route to scale and create something revolutionary
Mayo can encode aspects of its clinical reasoning into tools, platforms and partnerships that support clinicians beyond its own walls. It has some 320 algorithms running across the organization, each one under ongoing scrutiny as to whether it’s working well, could be improved, or should be scrapped.
That early work laid the groundwork for StateViewer, which uses AI to examine widely available brain scans to help clinicians identify potential activity patterns linked to nine types of dementia.
Trained and tested on more than 3,600 brain scans, it can help clinicians interpret them almost twice as fast, with up to three times more accuracy than normal. The tool helped researchers identify the type of dementia in 88% of cases.
The human role becomes more important, not less
Even in the earlier days of its pioneering development of AI, the Mayo Clinic contributed to the creation of the Coalition for Health AI (CHAI) and the National Academy of Medicine.
Accountability and guardrails were there from the start.
But in tools like StateViewer, the clinician remains responsible for judgement, context, empathy and patient care. Dr Leland Barnard, an Assistant Professor of Neurology and StateViewer data scientist, says: As we were designing StateViewer, we never lost sight of the fact that behind every data point and brain scan was a person facing a difficult diagnosis and urgent questions.
"Seeing how this tool could assist physicians with real-time, precise insights and guidance highlights the potential of machine learning for clinical medicine."
The real work is organizational
The challenge is not simply building clever models. It is making them usable in real workflows, trusted by clinicians, governed responsibly and adopted under pressure.
The first step is in making it happen in the first place. Our article on agentic AI goes into more detail – it’s all too easy for organizations to become gripped by institutional fear about making the wrong decision, or trying to find the perfect one.
That management of fear extends to persuading colleagues.
Mayo Clinic has a broadly flat hierarchy, with a lot of decisions made by committee, says Dr Farrugia, presenting their own challenges: “You have to work really hard at change, and you cannot change by fiat. You have to change by convincing people.
“I’ve always made the point that the right change agent is a servant leader because that’s how change becomes embedded. But it also means you’ve got to have that personality, the Mayo personality.”
To support that change, Mayo Clinic brought in a tech leader – who was also a clinician – and paired him them with a newly created Chief Medical Officer. “We brought them together so we could get the past and the present and the future working together,” says Dr Farrugia.
The lesson for every organization: where should AI scale decisions?
Before launching more pilots, leaders need to decide where AI should scale decisions, where it should augment people, and where human value must remain central.
Mayo Clinic started with experiments, but went on to create the Mayo Clinic Platform – an entire AI division dedicated to digital health. Before then, it started with confidence in making a decision, and equipping teams with the support, resources and training to achieve their ambition: to use AI to cure patients.
Its leaders were confident without being arrogant; fast without being reckless; thorough without obstructive. Not once did they take trust or human value for granted.
The clinician still makes the call.
If you’re a leader who sees the potential of AI, but need a helping hand, Oxford has been guiding organizations through change for more than 40 years.
Think we could help you with this? Contact us.