The case against starting with ambient documentation

Ambient documentation captures physician-patient conversations and generates draft clinical notes using AI. The technology works. The ROI case is compelling. The implementation risk is higher than most health systems anticipate.

Here is the problem with starting here.

The workflow complexity problem

A clinical note is not a transcript. It is a structured clinical document with specific requirements that vary by payer, by visit type, by diagnosis code, and by regulatory context. The ambient documentation tools that perform best do so in organizations that have done the hard work of standardizing their note templates, their documentation workflows, and their quality standards before deploying AI.

Most health systems have not done this work. Deploying ambient documentation into an environment with highly variable documentation practices produces highly variable AI outputs, which requires significant manual editing, which erodes the time savings that justify the investment.

What to do instead

Start with AI readiness. Before you deploy any AI documentation tool, understand your baseline: what your notes currently look like, where they deviate from standards, and what template standardization is possible.

Then pilot deliberately. Start with a specialty or a care setting where documentation is most standardized and where the workflow is most consistent. Measure rigorously. Expand based on evidence.

The organizations that achieve the highest ROI from ambient documentation are not the ones who moved fastest. They are the ones who moved most deliberately.