How AI model choice benefits clinical workflows

AI model optionality helps empower clinicians

The expanding use of AI in healthcare can offer clinicians a smart partner to lean on and help them make informed decisions to provide better care. But as with any tool, one size doesn’t fit all—especially when it comes to generative AI tools integrated in electronic health records (EHRs).

At athenahealth, we’ve made it a priority to offer clinicians the ability to choose from leading AI models within the AI-native athenaOne® electronic health record (EHR) specifically the utilization of Ambient Notes. This choice isn’t just a tech perk; it’s an attempt to enhance personalization in how clinicians document care, engage with patients, and ultimately deliver quality healthcare. Here’s why customizable model selection matters.

How do AI models function?

In the context of health information technologies, AI models are equipped with learning algorithms that are developed and trained in unique ways to assist clinicians. Each model undergoes specific training processes using distinct datasets, clinical scenarios, and coding techniques, which help shape how it understands and generates information.

The differences in model architecture and training lead to variations in the natural language processing (NLP) tools embedded in the EHR. As a result, the AI-generated outputs—such as clinical notes, summaries, or recommendations—may differ in style, detail, and focus depending on which model is applied.  

Different AI models for different documentation styles

Clinicians are as diverse as the patients they serve. Their documentation and writing styles may vary based on level of experience or practice workflows. Some physicians might favor succinct, pithy notes that distill key information quickly—ideal for fast-paced settings like emergency medicine. Others may prefer more detailed, thorough documentation to capture the nuances of complex patient cases or varied specialties.

Primary care providers could opt for a model that provides concise summaries and notes effectively capturing patient histories. Their AI model may be trained on current care concerns, developing care plans for pain management or chronic diseases, and brief reminders regarding preventative care. Orthopedic surgeons may favor detailed procedural notes, exact operative steps, and comprehensive pre/post-op documentation—all captured with an AI that understands surgical terminology and flow.

Physicians are likely not the only professionals who need to document patient encounters and help close patient charts. Practice administrators and care coordinators working in the EHR can also harness the power of generative AI in different workflows. They might lean toward notes that emphasize clarity and actionable items within a patient’s care pathway. After all, practice management extends beyond just note-taking—it includes patient outreach and communications, follow-ups, and other documentation touchpoints that can influence the revenue cycle and cash flow.

Giving clinicians the final say

At athenahealth, we want to empower clinicians with an AI-native EHR that’s capable of providing the most up-to-date data available, and generative AI tools that can readily access that data while being adaptive and catering to different practice needs.

That’s why we developed a solution within athenaOne that gives clinicians choices from multiple AI models in Ambient Notes, designed to flex with their preferences, specialty needs, and workflow rhythms. This customization is not just about text length or tone; it influences how the documentation gets structured and what style nuances are emphasized. Clinicians can also work with the models to help train them on the level of detail. And, with consistent use and clinician input, the models can grow more refined and accurate over time.

That level of customization can help lend itself to flexibility–and experimentation. Some clinicians may not have a hard and fast documentation style and thus may choose to work with multiple models to find the right fit. Offering multiple Ambient Notes models allows them to do just that.

At athenahealth, we want to empower clinicians with an AI-native EHR capable of providing the most up-to-date data available, and generative AI tools that can readily access that data while being adaptive and catering to different practice needs.

How the models may look in athenahealth’s AI tools

The athenaOne AI toolset includes native AI that helps to update, structure, and summarize data in patient charts to ambient tools that can help with same-day documentation and suggesting clinically inferred diagnoses. Model stylization can look different across clinical workflows, with a level of customization for each:

• Ambient documentation tools: For clinicians who prefer natural, conversation-style dictation, some AI models can transcribe and organize spoken data with context sensitivity. Other ambient tools may generate more formal, structured notes, particularly depending on how a clinician chooses to describe an encounter and leverage voice-to-text dictation tools. Moreover, ambient tools synthesizing historical patient data and encounter history may provide longer or shorter recommendations for clinically inferred diagnoses, depending on model type.
• Patient Outreach Tools: The AI models working behind patient communication tools adapt to whether the audience prefers brief reminders or more explanatory messages, enhancing patient engagement without extra administrative burden.
• Chart Assistant & Intelligent Summaries: Two tools currently in alpha testing, Chart Assistant and Intelligent Summaries, leverage generative AI to search across all available patient data, including documents, charts, and clinical events. These AI-powered tools provide clinicians with summaries of key elements in patient history and enable quick access to relevant insights–which can help enhance clinician preparedness before patient encounters.

Clinician control: AI on or off, always

Just as important as model choice and customization is the control clinicians have over AI usage. We respect that AI, while powerful, is a tool—one that should always serve to complement, and not replace, clinicians. That’s why Ambient Notes can be turned off at any time, ensuring clinicians remain in the driver’s seat regarding their documentation.

This level of control reinforces our commitment to a user-centric experience, avoiding the “one-size-fits-all” imposition of AI on clinicians’ workflows.

Wrapping up: the power of choice

For clinicians, the choice of AI model isn’t just technical—it’s about embracing their individuality in how they document and communicate within the EHR. Tools within AI-native athenaOne, like Ambient Notes, recognize and celebrate this diversity by offering a variety of customizable model options, enabling better, more efficient, and more satisfying clinical documentation and patient engagement.

This powerful differentiation enhances athenahealth’s promise: an intuitive, flexible platform built for real-world clinicians and the varied demands of modern care delivery.

If you’re interested in exploring how AI-native athenaOne can support your practice’s unique documentation needs, reach out to your athenahealth representative or visit our Resource Center to learn more.