athenaOne®: Building the AI-native clinical architecture

Building the AI-native architecture into the EHR

Most electronic health record (EHR) vendors have added AI. But an important question for organizations faced with the difficult choice of evaluating these platforms is whether AI is integrated, how seamlessly that AI is integrated, what that integration makes possible.

Individual practices and health systems are starting to see value in AI adoption for clinical use cases. The question, for many, has shifted from whether to introduce AI to how to introduce it without creating yet another disconnected system clinicians have to learn and trust.

Being a truly AI-native clinical platform means AI shares the same data, workflow, permissions, and clinical context as the EHR itself. It’s not a separate application integrated after the fact. When AI operates inside a unified platform, drawing from the same longitudinal data and network intelligence that powers scheduling, documentation, billing, and patient engagement, it can do something point solutions cannot: carry context continuously across the entire clinical encounter.

Every standalone AI tool a health system adopts creates a new data boundary, a new governance process, a new workflow interruption, and a new integration dependency. CMIOs evaluating AI strategy recognize this problem immediately. The promise of AI in healthcare has moved beyond automation at an individual capability level to an intelligence partner that compounds across the visit, across the care team, and across the patient relationship.

That is the premise athenaOne is built on. Not AI as a feature. AI as the operating layer of the clinical experience.

Continuous intelligence across the visit

The clearest way to see why architecture matters is to follow a single patient encounter from preparation through close and ask, at each stage, what becomes possible when the AI has full context rather than a fragment of it. In theory, AI-native should help change the clinical encounter at every stage:

• Before the visit → AI synthesizes patient context so clinicians arrive prepared
• During the visit → AI documents in real time and surfaces clinical intelligence
• After the visit → AI accelerates documentation close and care coordination
• Across the practice → AI supports quality management and operational insight

The clinical AI capabilities in athenaOne are meant to achieve this goal, giving clinicians an intelligent partner throughout the entirety of the patient journey.

Before the visit

Knowing a patient well before they walk in the room changes the quality of the visit. For new and complex patients or those seen infrequently, that preparation has historically meant navigating years of notes and still arriving at the encounter with an incomplete picture.

The Patient Overview functionality in athenaOne addresses this challenge directly. It delivers AI-generated, problem-oriented summaries that distill a patient's active problems, recent treatments, and relevant clinical history into a structured pre-visit briefing. Notably, AI draws from the patient's longitudinal record inside athenaOne, not a single note or a summarized fragment.

Chart Assistant with Sage™ extends that preparation into a conversational interface. Clinicians can ask natural language questions to an AI assistant directly in the encounter workflow — recent lab trends, medication changes, ED visits since the last appointment — and get answers from the chart without navigating away. Because Chart Assistant operates inside the platform rather than alongside it, the context built during preparation carries forward into the encounter. That helps reduce retyping and context loss across workflows. The preparation becomes the foundation for what follows.

During the visit

At its best, a clinical encounter is a conversation, one where the clinician can show empathy and give their full attention to the patient rather than worrying about typing notes on the keyboard. That presence matters clinically. It changes what gets noticed, what gets asked, and what gets caught. Documentation requirements haven't eliminated that ideal, but for many providers they've complicated it. athenaAmbient™ and the AI-native clinical encounter are designed to restore it.

athenaAmbient is athenaOne's native ambient digital scribe that listens to the patient-provider conversation and generates clinical documentation in real time. This proprietary ambient tool works on desktop and iOS, with no third-party login and no workflow interruption. Because it runs inside athenaOne, the documentation it produces is immediately connected to the patient record, the billing workflow, and the quality infrastructure.

Additionally, athenaAmbient is being bolstered to enhance its capabilities for diagnosis suggestions and suggested medication orders, surfacing clinical intelligence from the visit transcript itself. The Clinically Inferred Diagnosis Gaps capability complements this by using AI to identify CMS HCC diagnosis gaps from comprehensive chart data, which can help clinicians close gaps that might otherwise surface only in retrospective review.

Finally, there’s explicit optionality. Clinicians can choose to toggle the AI-native encounter on and off, depending on their preference. Existing or new customers also have the option of opting for Ambient Notes at their preference, which itself has multiple documentation models and multilingual transcription capabilities.

What do these pre-visit and encounter capabilities ladder up to, ultimately? Early adopters are seeing strong return on investment. One athenahealth customer using AI-native athenaOne across the clinical workflow reduced chart prep and documentation time by nearly six minutes per visit, with results reflecting topline customer data.¹ Meanwhile, clinicians using AI-native athenaOne have increased same-day chart completion to more than 80%.²

Across workflows, these capabilities help contribute to one of the most meaningful aspects of medicine: empathy for the patient. As Dr. Lynn Joffe of DTC Family Health PLLC* stated, “I’m totally in the moment with that patient.”

After the visit

Because ambient scribes like athenaAmbient and Ambient Notes help capture the encounter as it happens, documentation is largely complete when the patient leaves the room. Providers review, refine, and sign rather than reconstruct. That frees up more time for meaningful medicinal attention and care.

More complete documentation can help create a more accurate record of the care that was delivered. When the clinical narrative is thorough, the coding that follows it can be, too. Diagnoses are supported. Complexity is captured. The result is a billing picture that more faithfully reflects the work of the visit, which matters both for appropriate reimbursement and for the integrity of the patient record over time.

That completeness compounds in value-based care environments, where accurate, longitudinal documentation is foundational to performance. Risk stratification, quality measure closure, and care gap identification all depend on a record that reflects the full clinical picture. When AI helps capture that picture in real time, the downstream benefits extend well beyond the individual encounter: cleaner claims, a reduced chance of denials, and a more complete view of patient health that supports the kind of proactive, coordinated care value-based models are designed to reward.

Why integrated AI outperforms point solutions

Standalone AI tools may have solved an immediate problem for many organizations, especially before deeply integrated options were available. But as health systems continue to transition from experimentation to enterprise deployment, a different challenge has emerged: stitching together multiple AI experiences that don't share context, don't share governance, and don't compound on each other's outputs. That experience doesn’t reduce complexity; it can add to existing frustrations.

The capabilities in athenaOne are individually valuable. But their compounding effect comes from operating on shared context inside an integrated platform. It’s the totality of how the capabilities work together that help make this a differentiated experience. Preparation informs documentation. Documentation informs coding. Coding informs quality. Quality informs operations. That chain of continuity is what AI-native architecture makes possible.

athenaConnect™, athenahealth's network connectivity infrastructure, is a meaningful part of why this also works. It enables cross-setting clinical data sharing between practices, health systems, and care settings. The AI models powering athenaOne draw from a breadth of longitudinal network context that standalone AI applications operating outside a connected ambulatory platform cannot typically access.

athenaOne can service real-time updates and deploy them at scale, with an Advanced Intelligence Layer that also leverages different AI models, engines, and data services to engage with the data in targeted ways depending on the desired output. The network intelligence is a structural advantage that becomes more valuable as the platform is informed by additional longitudinal context across a growing network, not a static dataset.

Beyond the encounter 

The AI-native encounter addresses the individual visit. athenaOne's broader clinical AI suite aims to help practices reduce operational pitfalls and scale accordingly.

Data Explorer and Quality Management can help give practices a unified view of performance across all quality measures, with an AI assistant that can identify patients with the highest needs and surface potential actions to take. Care teams move from insight to intervention without manual report-running or list triage.

For patients returning from a hospital stay or care event, Care Events can help deliver concise, AI-generated summaries of those interactions, highlighting what's most relevant for the treating clinician. Rather than parsing through dense discharge documentation, the clinician gets a clear picture of what happened. The right information surfaced in context, at the moment it's needed, inside the same workflow where the next encounter is already being prepared.

AI-native Clinical Inbox and Work Management will bring the same platform intelligence to inbox management. Rolling out later this year, the Inbox Assistant will enable clinicians to navigate the clinical inbox conversationally, surfacing answers without manual searching. Inbox Triage with AI will help remove routine administrative work from the queue, so clinical attention goes where it's needed.

For organizations thinking about AI strategy at scale, these capabilities matter not just as productivity tools but as governance infrastructure. This architecture provides a consistent, auditable, platform-native approach to AI deployment that doesn't require managing a portfolio of disconnected vendor relationships.

What this means for organizations evaluating EHRs 

The question for CMIOs and health system leaders isn't whether to adopt AI. It's whether the AI they adopt will compound in value over time or plateau at the boundaries of each individual tool. Platform architecture determines that answer more than any single feature does.

Organizations debating AI deployment at scale may want to evaluate whether the architecture behind those features will become more valuable as adoption grows, and whether the intelligence powering those features is informed by the kind of network-scale longitudinal data that only a connected ambulatory platform can provide. That's the distinction between AI that automates individual tasks and AI that translates across the care experience.

athenaOne's clinical AI suite is expanding throughout 2026, and rapidly evolving. That includes innovating and testing product features directly alongside customers and champions at different organizations. Because AI only works when it works for your practice.

For organizations ready to move beyond the feature comparison and evaluate what AI-native architecture makes possible, the conversation starts here. Request a demo to see athenaAmbient, Chart Assistant, the AI-native encounter, and the full clinical AI suite in action.