AI lets clinicians “chat” with electronic medical records

February 2026

A recent article published by Stanford Medicine describes a new artificial intelligence tool that allows clinicians to interact directly with electronic medical records (EMRs) using conversational prompts.

The tool, known as ChatEHR, enables doctors and other healthcare professionals to ask questions about a patient’s medical record and receive structured, summarised responses drawn directly from the chart.

In much the same way people interact with large language models such as GPT-4, Stanford clinicians in a pilot program can now “chat” with patient records to retrieve relevant clinical information.

According to the Stanford report, the system is integrated directly into the EMR platform and is designed to improve efficiency without replacing clinical judgment.

Faster search, summary and information gathering

Clinicians using ChatEHR can ask questions such as:

• Does this patient have documented allergies?

• What were the latest lipid profile results?

• Has the patient undergone colon cancer screening?

• Summarise this patient’s cardiovascular history.

The AI retrieves relevant data directly from the medical record and presents it in a concise format.

Nigam Shah, MBBS, PhD, chief data science officer at Stanford Health Care, noted in the article that AI is only useful if it is embedded within clinicians’ workflow and grounded in medical context. The system, he explained, pulls securely from relevant patient data within the electronic health record.

Importantly, ChatEHR is not intended to provide medical advice. It functions as an information-gathering and summarisation tool. Clinical decisions remain entirely with the treating physician.

Why this development is noteworthy

Electronic medical records contain vast amounts of information — laboratory results, imaging reports, operative notes, clinic letters, medication lists and longitudinal trends.

For many clinicians, navigating these records can require significant time, particularly when managing complex cases or urgent presentations.

The Stanford article highlights how AI may help:

• Reduce time spent searching across multiple record sections
  
  

• Provide rapid summaries during emergency or transfer cases
  
  

• Extract relevant trends from historical data
  
  

• Support administrative review tasks
  
  

Jonathan Chen, MD, PhD, a physician quoted in the Stanford report, noted that when a patient presents acutely, understanding the entire clinical history quickly can be challenging. AI-assisted summarisation may help clinicians gain that overview more efficiently.

Beyond summaries: automation features

The Stanford team is also developing “automations” — evaluative tools based on a patient’s record that can assist with administrative determinations, such as transfer eligibility or care planning pathways.

These features are still being evaluated, but they demonstrate how AI may extend beyond simple data retrieval into workflow optimisation.