EHR vendors are racing to build AI tools that ease clinician workloads, but the people who actually use those tools are often left out of the development process. athenahealth is trying a different approach with its EHR AI CoLab, a monthly forum where 20 clinicians and EHR power users work directly with the company's engineers, data scientists, and product designers to shape AI features before they reach the broader market.
The CoLab, launched in early 2026, gives clinicians a structured channel to flag workflow problems, share end-user feedback, and test AI capabilities in alpha and beta phases. Thousands of clinicians across the athenahealth network participate in testing the tools the group helps shape.
"We can co-develop with our users [through the CoLab], and more so than ever before, we can partner with our users, we can get their feedback and then instrument and implement that feedback directly in the product in real time," said Chad Dodd, athenahealth's vice president of product development.
Clinician feedback shapes AI documentation tools
The CoLab's input has already changed several athenahealth products. For athenaAmbient, the company's ambient clinical documentation tool, members pushed for better diagnosis creation, fewer duplicates, and user-configurable settings.
"Initially, there really weren't any user preferences or settings that could be established. And now we do have some of those options, but all of this is continuing to evolve, too," said Liz Bauske, clinical information systems manager for Barrington Orthopedic Specialists and a CoLab member.
The group also influenced Sage, an AI-powered copilot built into athenaOne. CoLab members requested the ability to save frequently used questions as text macros, so clinicians don't have to retype the same queries every time they review a patient chart. Bauske said that feature, which she can manage at an administrative level, has been well received by her staff.
Ideas also flow in the other direction. Clinicians told athenahealth that reviewing the sheer volume of data in patient charts was a major pain point. That feedback led to a problem-based summaries feature in Sage that uses AI to surface the most relevant patient information. The feature is currently in alpha testing.
Why clinicians belong in the AI development loop
Dodd sees the CoLab as a way to merge two distinct skill sets. Clinicians understand patient care needs but not how to build AI to meet them. Engineers know the technology but not the clinical realities of daily practice.
"When you bring the two together on a call, you give birth to this new customization, personalization element," Dodd said.
That personalization matters for adoption. Bauske said Sage initially saw low uptake at her orthopedic practice, but usage rose once athenahealth added the saved-questions feature.
Involving clinicians in development also builds trust in AI, which remains a hurdle among medical staff. Dodd noted that some clinicians want to understand how the technology works before they'll use it. Helping build and test the product gives them that visibility.
Bauske emphasized that governance and guardrails remain a priority as AI tools spread through healthcare organizations.
"Something we really have to continue to be cognizant of is having governance over the use of AI technologies as an organization, both with our patients and our staff, and making sure that we have guardrails to protect both our staff and our patients," she said. "So, we will continue to work with athenahealth to ensure that those guardrails are available to us and present and really visible for our providers if they need to be."
athenahealth plans to track usage metrics, including time spent on chart review and documentation, same-day encounter close rates, and physician satisfaction, to gauge whether the CoLab's work is reducing administrative burden.
The broader lesson for healthcare IT teams is that AI adoption depends on who gets a seat at the design table. Clinicians are the ones who will live with these tools daily, and their input can determine whether a feature gets used or ignored. For IT and development professionals, the CoLab model offers a template for closing the gap between technical capability and real-world usability - one that applies well beyond healthcare. For more on how AI is being applied in clinical settings, see AI for Healthcare coverage, or explore AI for IT & Development resources for building collaborative AI workflows.
Why this matters for IT and development professionals
The CoLab demonstrates a practical answer to a problem most development teams face: users reject tools that don't fit how they actually work. athenahealth's approach - embedding a small group of expert users in the development cycle and scaling their feedback through broader beta testing - shortened the loop between user complaint and product change. The saved-questions feature in Sage is a direct example. That kind of iterative, user-driven development is a model any IT team building AI tools for internal or external users can apply. The takeaway: build the feedback channel into the process from day one, not after the product ships.
Your membership also unlocks:
