I’m equally an AI optimist and an AI realist. I’ve spent the past few years watching healthcare pour resources into AI and waiting for it to fundamentally change how healthcare operations run. I believe it can, but I don’t believe it has yet.
Headless is the first AI architecture with the potential to do exactly that, by taking on the coordination work that has always depended on people.
Healthcare has seen this kind of evolution before. Digital front door technologies existed for years before COVID-19, but adoption remained incremental. The pandemic accelerated adoption because it forced the industry to think differently about engaging patients.
I believe AI has reached a similar point.
Health systems have spent the past few years applying AI to individual tasks while leaving the underlying operating model largely intact. Headless has the potential to change that.
Today’s AI Has Delivered Value, but Not Yet the Results Healthcare Needs
Healthcare organizations have embraced AI. McKinsey reports that more than 70% are already pursuing or implementing generative AI capabilities. Those investments are already delivering measurable value. Yet McKinsey’s broader research suggests very few organizations have fully integrated AI into the way work gets done.
That’s because today’s AI is already improving isolated tasks. AI can draft a response, summarize a note, answer a question, or complete a specific task. Clinical teams are getting great ROI from ambient listening and AI scribes, for example, and many health systems have adopted AI call routing systems.
Some health systems have adopted AI at a remarkable scale. Tampa General, for example, has publicly discussed using 61 AI systems. Across healthcare, however, those capabilities haven’t fundamentally changed how operational work gets coordinated. People still work across multiple systems to get work done.
A referral coordinator looking to get a patient scheduled with a specialist might review AI-generated visit notes in the EHR and look for an AI-routed fax in their inbox, but they’re still doing the bulk of the work to actually schedule the visit. AI is helping them complete tasks, but it hasn’t changed how they work.
Headless (and AI working in harnesses) change that model by accomplishing entire chains of tasks across systems, reducing the need for people to move between systems to coordinate work.
AI can assemble information across systems, build a plan, execute routine work, and escalate to people only when human judgment is required.
Instead of the referral coordinator gathering the context and executing next steps themself for each patient, they can now scale their work. They can tell headless AI that they want to improve preventive care for diabetes patients. Then, they can check back over several days as AI reaches out to patients who haven’t been seen in a while, newly diagnosed patients, and high-risk patients to set up consultations with specialists.
For the first time, keeping work moving doesn’t have to depend on people.
Why Healthcare Needs What Headless Promises
Healthcare has more to gain from autonomous (headless, or harnessed) AI than almost any other industry. It’s high stakes, but highly manual. Health systems with the latest technology are still manually processing hundreds of thousands of faxes annually, and every fax can delay care if it’s not acted on immediately.
Few industries match healthcare’s level of operational complexity with its dependence on disconnected systems and people to coordinate across them. But ongoing labor shortages, tight margins, and a healthcare system that is increasingly complex to run make the old approach unsustainable for health systems. Headless comes at the right time to upend the status quo.
With AI as the headless coordination layer across healthcare systems, human expertise can stay focused on the most valuable parts of care delivery instead of on tracking information down system by system.
How Headless Changes the Way Health Systems Organize Work
As organizations adopt this operating model, leaders will need to answer different questions.
Which workflows should AI own? When should work escalate to staff? How should performance be measured? What governance is needed to ensure AI acts safely, consistently, and transparently across systems?
One way to think about headless is through what some AI systems call “Plan Mode.” Instead of asking AI to complete one task at a time, users define the objective. AI builds a plan, users review it, and then it carries out the work.
This gives leaders more choices when designing workflows, because they don’t need to assume people will manually connect every step. Leaders can intentionally design coordination into the workflow itself, deciding where technology should act and where people should apply judgment. For example, when expanding a service line, leaders can begin by deciding which parts of the work AI should own end to end, and where staff should step in.
Health systems that can now rapidly scale their staff will need to spend less time on routine tasks, and budget more resources for reviewing AI-built plans and ensuring headless AI follows the right processes without drift.
Headless AI Could Transform Healthcare Processes
Headless AI’s ability to scale processes could significantly cut labor costs, but the benefits could extend far beyond productivity. Patient access, referral completion, prior authorization, care coordination, and other processes across the health system could all be faster and better without depending on someone manually stitching things together.
I’ve spent the last several years leaning into my “AI realist” side, but headless AI has me truly optimistic about the next phase of AI adoption in healthcare. After decades of just digitizing information, healthcare can now move to taking action on it.
Headless is the first architecture that gives healthcare a realistic path toward that future.
