Skip to main content

What is AI adoption?

The AI Glossary

In our classes, we often pause to explain the ideas behind AI and the language used to describe it. These posts grew out of those conversations. They’re for anyone who would like to understand AI better, with no technical background assumed.

AI adoption is the process through which a tool becomes part of actual, sustained work. Buying a license, making a model available or completing a demonstration is not the same as adoption. People must be able to use the service in the circumstances where the work occurs.

Adoption can be measured in several ways. Availability records who could use a tool. Activation records who has tried it. Continued use shows whether it becomes part of a routine. None of these measures alone establishes benefit. Frequent use can coexist with high correction effort or poor outcomes.

People adopt tools for practical reasons. Does it help with a task they recognize? Is it easy to access at the right moment? Does it fit the existing information flow? Can they correct a mistake without starting again? These questions often matter as much as small differences in model performance.

Non-use can also be informative. A tool may be unsuitable for a specialty, inaccessible to some staff or difficult to use with a particular language. Comparing only enthusiastic early users can overstate its value for the wider organization. The evaluation should include who is not using it and why.

Adoption should not be treated as a one-way journey toward maximum use. Sometimes a narrower scope works better. Sometimes a manual or simpler software process remains preferable. A mature implementation allows that decision without treating it as failure to embrace innovation.

For an ambient documentation system, a team might examine which clinicians continue using it, how much editing is required and whether performance differs by encounter type. It should not assume that high use proves reduced burnout or improved patient care. Those outcomes need their own evidence.

What I look for is appropriate use: the right tool for a defined task, with a workable response when it is uncertain or unavailable. Adoption is valuable when it supports a better service, not when it merely raises the number of AI interactions reported to leadership.

Sources: ASTP/ONC — Health IT Playbook: Electronic Health Records; Lukac and colleagues — Ambient AI Scribes in Clinical Practice: A Randomized Trial (2025)

To learn more, take a look at the AI in Health Care Certificate of Specialization from the Advanced Learning Academy at Harvard T.H. Chan School of Public Health. You can take an individual course or work toward the full certificate.


AI is a fast-moving field. This explanation reflects our class discussions and related material at the time of writing. Terminology, capabilities and tools may change.

For suggested additions or corrections, please email me at tpanch@hsph.harvard.edu.

Dr. Trishan Panch
Program Director, AI in Health Care
Harvard T.H. Chan School of Public Health — Advanced Learning Academy

Lead with AI in Health Care

The AI Pragmatist is a monthly newsletter from ALA faculty, with new explainers & essays from the Advanced Learning Academy