The Microsoft AI diagnosis tool works the way a doctor does. Instead of swallowing a patient’s full file at once, it asks questions, orders tests, reads the results and updates its thinking step by step. In a study published on Thursday in the journal Science, it proved more than twice as accurate as general physicians at solving complex medical cases.
Microsoft researchers say the system identifies and coordinates the questions that should be asked of a patient, as well as the exams and tests that should be ordered. An “orchestrator” organises how the AI tackles a problem by running a simulated panel of clinical experts that collaborate to solve diagnostic cases.
How the Microsoft AI diagnosis tool differs from existing systems
Existing medical AI models typically take an all-at-once approach. They examine all of a patient’s symptoms, lab results and imaging studies together, then generate a diagnosis.
With the orchestrator coordinating the steps, the Microsoft model first decides what to ask about the patient’s history and what is significant in the physical exam. Based on what it learns, it decides which tests to order, considers the results and orders further testing as appropriate. All the while, it keeps updating its working hypotheses and plans, the research leaders said in an interview.
Tasked with diagnosing more than 300 highly unusual and confusing medical dilemmas, the model was more than twice as accurate at “working up” the patient and determining the diagnosis as general physicians who were using textbooks and the internet.
The tool also outperformed human physicians when presented with scenarios drawn from a major database of emergency room admissions.
Why the testing method matters
Assessments comparing older AI systems to human doctors usually rely on multiple-choice questions, the researchers said. That format flatters machines.
“AI has become very good at getting a 100% pass rate on those,” said Dr Dominic King, the corporate vice president of the health unit at Microsoft AI, whose team is developing the new technology. “But of course, that’s not real-world. No patient ever came to see me with a nice summary.”
The point is that real diagnosis is messy. A doctor starts with a complaint, asks follow-up questions, examines the patient and only then orders tests, revising the list of suspects at each stage. The study suggests that making an AI imitate that sequence, rather than skipping straight to the answer, produces far better results.
What comes next
The system remains a research project, not a product in clinics. Questions of regulation, liability and how doctors would actually use such a tool in practice are still open. But the result lands at a moment when health systems everywhere are under pressure, and diagnostic errors remain a major cause of harm.
For readers following the AI beat, the study also illustrates a wider shift: the frontier is moving from models that answer questions to models that plan investigations. That same idea is already shaping cybersecurity tools and the devices people carry every day.
Pakistan’s own debate over technology access shows how fast new tools reach the public once they leave the lab. Whether AI diagnosis follows the same path will depend on clinical trials and regulators, not just benchmark scores.
The announcement also lands as Pakistan’s tech sector posts strong numbers, with PTCL Group reporting a Rs47 billion profit for the third quarter of 2026.
Frequently asked questions
What is Microsoft’s sequential diagnosis AI?
It is a research system that diagnoses patients step by step, like a doctor, using an “orchestrator” that runs a simulated panel of clinical experts to decide which questions to ask and which tests to order.
How accurate was it in the study?
On more than 300 highly unusual and confusing medical cases, it was more than twice as accurate as general physicians using textbooks and the internet, according to the study in Science.
Is the tool available to patients?
No. It is a research project reported by Microsoft’s health AI team, not a product approved for clinical use.
