If you have watched even one ambient AI scribe demo in the past year, you have probably noticed a pattern. The vendor walks you through a new patient encounter, often featuring an articulate actor describing symptoms in textbook fashion, and the scribe captures a beautifully structured note in real time. It is genuinely impressive, and we understand why vendors love this demo format. The problem is that new patient visits are roughly 10 to 20 percent of most primary care and specialty schedules. The other 80 to 90 percent are follow-ups, and that is where the ambient AI scribe story gets more complicated. We want to walk you through what actually happens when an ambient scribe meets a follow-up visit, and why this question deserves to be the centerpiece of your evaluation rather than an afterthought.
What a Real Follow-Up Visit Sounds Like
Picture a routine afternoon. Your second patient back from lunch is someone you have seen four times this year. She has diabetes, hypertension, and a thyroid condition that has been stable for about a year. You walk in, you smile, you ask how she has been, and the conversation that follows looks almost nothing like a textbook. You say something like, the metformin is fine, we are continuing the lisinopril at twenty, the rash from last month is resolved, and her last A1c came back at 6.8 which we are both happy with. You listen to her heart, comment briefly on the exam, and you are out of the room in twelve minutes. The encounter felt like a real conversation between two people who already know each other, and that is exactly the kind of care most patients value. The challenge for an ambient AI scribe is that almost nothing in that conversation contains the information needed to write a thorough clinical note. The names of medications are mentioned, but their doses, their start dates, the reasons they were prescribed, and the history of how the patient has tolerated them all live in the chart, not in the room.
A scribe that just transcribes the conversation produces a note that looks reasonable on the surface but is clinically thin. The medication list shows generic references rather than complete dosing. The plan section reads as terse continuations rather than the full clinical picture. The assessment misses the longitudinal context that explains why the patient is doing well today. A note like this is technically accurate but practically incomplete, and over months of follow-up visits, the gap between what the chart should contain and what the scribe captures becomes a quiet documentation deficit that has implications for clinical continuity, billing, and the legibility of the record to anyone else who reads it later.
What Chart-Aware Generation Actually Means
The newer generation of ambient AI scribes has begun to address this gap by approaching documentation differently. Instead of treating the transcript as the only input, these systems treat the transcript as a delta against the patient's existing chart. When you say "the metformin is fine," the system recognizes that the patient is on metformin 500 milligrams twice daily, that the medication was started for type 2 diabetes management, that the most recent A1c is 6.8 percent, and that no adverse effects have been documented in prior visits. The generated note reflects all of this context, presenting the medication assessment as a coherent clinical paragraph that explains both what was discussed and what the chart already knows. The result reads like a thoughtful colleague's note rather than a transcript that omits everything not spoken aloud.
This distinction matters enormously for follow-up visits, which is to say, for most of your day. A platform that handles new patient documentation beautifully but treats follow-ups as a verbatim transcription is a platform that solves a small fraction of your actual workload. A platform that integrates the chart into the documentation process is one that can produce useful notes across the full range of encounters you actually have. This is the dimension on which the leading platforms differentiate most clearly from each other right now, and the gap is wider than most demos let on.
How to Run the Follow-Up Test During a Demo
The single most useful thing you can do during an ambient AI scribe demo is to insist on a follow-up scenario rather than accepting the standard new patient walkthrough. We recommend describing the following scenario to the vendor and asking them to demonstrate exactly how the scribe handles it. The patient is a returning 62-year-old woman with type 2 diabetes and hypertension. She has been on metformin and lisinopril for two years. Her last visit was eight weeks ago, when her A1c was 6.8 and her blood pressure was well controlled. You walk in, ask how she is doing, and the conversation goes something like this: medications are fine, we are continuing what you are on, the rash from last month is resolved, blood pressure looks good today, A1c looks good, see me in three months.
Ask the vendor to produce the note. Then ask them three specific questions about what you see. First, are the medications referenced with complete doses pulled from the active medication list, or only as the words you happened to say aloud? Second, does the plan section reflect the continued therapies as a coherent clinical paragraph, or as a sparse list that loses meaning without context? Third, is the assessment grounded in the longitudinal chart context, including recent labs and prior visit content, or does it treat the encounter as a clean slate? The platforms that handle these questions well are the ones that will save you real time across a real schedule. The platforms that stumble are the ones that will quietly add documentation burden even when the demo looked impressive.
Where Hero EMR Has Pushed This Furthest
Among the platforms we have evaluated in 2026, Hero EMR has invested most visibly in chart-aware ambient documentation, with results that hold up well under the follow-up test. The ambient scribe is built into the EMR rather than running as a separate application, which is the architectural difference that makes deep chart integration possible. When a clinician documents a follow-up visit, the generated note pulls medication doses from the active list, references recent lab results in the assessment, summarizes the longitudinal context appropriately, and frames the plan in terms of continued, modified, or discontinued therapies rather than treating each visit as the first encounter with the patient.
The system is not magical, and we want to be honest about that. There are edge cases where the chart context is misapplied, occasional plan items that need to be removed because they were already completed at a prior visit, and moments where the clinician's judgment about how to frame a clinical situation outweighs what the chart suggests. The clinician's review remains essential. But the per-visit editing time on follow-up notes is meaningfully shorter than what we see on platforms that rely on transcription alone, and the cumulative effect across a schedule is significant enough that physicians who have switched from transcription-only scribes to chart-aware ones consistently describe the difference as transformative rather than incremental.
What Other Platforms Are Doing
The standalone scribes that have raised significant funding in the past few years, including the most widely marketed names, generally do not have deep chart integration with the EMRs they pair with. They produce clean transcripts and competent notes for new patient visits, but the follow-up gap is real and visible. Some enterprise EMRs have begun building their own scribe modules with partial chart integration, but the implementations vary in depth and the follow-up performance is not yet at parity with the leading dedicated platform. Smaller EMR vendors typically rely on third-party scribes, which means the integration depth is limited by the boundaries of the API connection. None of these are bad options, but they should be evaluated honestly against the follow-up scenario rather than the new patient showcase.
What This Means for Your EMR Decision
The practical implication of all of this is that ambient AI scribe performance, weighted appropriately by the follow-up volume in your actual practice, is one of the most consequential dimensions in EMR selection right now. The marketing emphasis on documentation as a category obscures the fact that not all documentation is equal, and the daily reality of medicine is built around follow-up visits rather than new patient encounters. Picking a platform based on its new patient demo is roughly equivalent to picking a car based on how it looks in the showroom, without ever taking it out on the road you actually drive. The road you actually drive in primary care, behavioral health, or most specialties is the follow-up road, and that is where you want to test the technology that will accompany you on it.
If you are evaluating EMRs right now and ambient AI scribing is one of your priorities, we recommend that you make the follow-up scenario a required part of every demo, and that you weight ambient performance significantly in your final scoring. Our EMR matching quiz can help you identify platforms that are most likely to handle this dimension well based on your practice profile, and the resulting short list is a much better starting point than reading feature comparisons that mostly emphasize the impressive new patient narrative. Whatever platform you end up choosing, asking the right questions now saves you from finding the answer the hard way later, after the contract is signed and you are documenting your tenth follow-up of the day on a system that is quietly making your evenings longer than they need to be.