Draft, review, signthe clinician reviews and signs each note before it is final
In your EHRthe note is written to the encounter in the system the group already runs
Edit logeach change a clinician makes to a draft is tracked by section

The specialty groups we talk to compare three kinds of scribe. The first is a standalone ambient scribe, and the second is the AI add-on their EHR vendor now offers. The third is a scribe that is part of a wider agent deployment. All three draft a note from the recorded visit. They differ in where the note is written, who builds the specialty templates, and what happens to the orders the visit creates. Flexbone drafts the note, the clinician reviews and signs it, and the note is written to the EHR the group runs. The same team runs fax, referral and prior authorization agents, so the work after the visit can be connected to the note.

How Does an Ambient AI Scribe Work During a Visit?

The work runs in five steps. The patient decides the first, and the clinician decides the last. First, the patient is asked for consent to record, and the answer is logged with the visit. Several states require every party to agree before a conversation is recorded, so the consent wording follows the group's policy and state law.

Second, the conversation is recorded and transcribed. Third, the transcript becomes a draft note in the template the specialty uses, with history, exam, assessment and plan in their own sections. Orders, referrals and follow-up instructions mentioned in the visit are listed separately for the clinician to act on.

Fourth, the clinician reviews the draft, corrects what is wrong, adds what is missing and signs. Fifth, the note is written to the encounter in the EHR. Where the EHR accepts an unsigned draft, the draft is placed in the encounter and signed there. Otherwise the note is written after sign-off. The order is the same either way: draft, review, signature, chart.

Complimentary Flexbone audit

Get an outside read on your ambient documentation workflow

In 30 minutes we map your current volume, the payers and systems involved, where staff time goes, and the highest-ROI calls and follow-ups Flexbone can take off your team first, scoped to the work you actually run.

Book an audit →

Which EHRs Can the Scribe Write the Note Into?

The note goes to the EHR the group already runs, so clinicians do not chart in a second system. Flexbone connects in one of two ways. Where the EHR vendor exposes an API for clinical documents, the note is written through it. Where it does not, a browser automation enters the note in the EHR's web interface, on the same screens a clinician or scribe uses.

Flexbone has built connections that read and write in eClinicalWorks, athenahealth and Tebra for other workflows, such as filing documents and correcting claims. For groups on ModMed or Epic, the connection method is confirmed in the audit before any rollout date is set. There is no EHR migration and no change to where notes are stored.

How Does This Compare with the EHR's Own AI Scribe?

Several EHR vendors now offer their own ambient scribe, and specialty groups ask how Flexbone compares. eClinicalWorks offers Sunoh.ai, and eClinicalWorks says the provider edits the draft and imports it into the Progress Note. The athenahealth Ambient Notes feature lets clinicians choose from ambient models, including Suki AI and iScribe AI. Tebra offers AI Note Assist, which it describes as built into the Tebra EHR. ModMed has announced ModMed Scribe for its EMA EHR, starting with dermatology. Epic launched built-in AI charting in February 2026 that drafts the note and queues orders.

If the only goal is a draft note inside one EHR, the vendor's add-on is a reasonable default. It ships with the EHR, the vendor supports it, and the contract is already in place. Groups choose a separate scribe for other reasons. A specialty template does not fit, several sites run different EHRs, or the group wants the note tied to the work after the visit.

Standalone scribes such as Abridge, DeepScribe, Nabla and Heidi are compared in Suki AI alternatives. Whichever you shortlist, ask each vendor who builds and changes the templates and where the clinician signs. Also ask how edits are tracked and what happens to the audio after signing.

How Are Specialty Templates Built?

Templates are built from the group's own notes, not from a generic primary care format. A forward-deployed engineer sits with the clinicians, reads a sample of signed notes, and maps the sections, phrasing and required fields each provider expects. The template is reviewed with the provider on the first visits before other providers use it.

Specialty detail is where the template matters most. An orthopedic note needs laterality, imaging reviewed, injection details and work status. A cardiology note needs device and test history. A GI note needs procedure history and prep instructions, and a dermatology note needs lesion location, size and morphology. Each of these is set up as a field the draft fills or leaves marked for the clinician.

What Stays with the Clinician?

Clinical judgment and the signature stay with the clinician. The scribe drafts the note; it does not diagnose, choose treatment or sign. Orders and referrals mentioned in the visit appear in the draft as a list for the clinician to place or confirm. The scribe does not place them on its own. The clinician also decides what the note leaves out, and can remove a topic discussed in the room before signing.

Complimentary pilot

Have Flexbone run your ambient documentation workflow on real cases

A complimentary six-week pilot once the BAA is signed, with a weekly review of what the agents handled, which results were written back to your system, and the hours your team got back.

Start a pilot →

How Is Note Accuracy Reviewed and Tracked?

Flexbone does not publish an accuracy percentage for the scribe. A single number hides where the errors are, and the measure that matters is how much your own clinicians change. Each signed note is compared with its draft, and the edits are logged by section: what was added, removed or rewritten.

The forward-deployed team reviews that edit log during rollout and after go-live. A section that clinicians rewrite often points to a template problem, and the template is changed. A drop in edits on that section shows the change worked. The group sees the same log, so the review rests on its own data. Ask any vendor that quotes an accuracy figure how it was measured, in which specialty, and against what reference.

How Is Patient Audio Protected?

Flexbone signs a business associate agreement under HIPAA with each customer before any patient audio is recorded. Data stays in the United States, encrypted with AES-256 at rest and TLS in transit. Model providers run under zero data retention. Audio, transcripts and notes are never used to build another customer's agents.

Each customer's data is tenant isolated: its records, files and logins are kept separate from every other customer's. When you compare scribes, ask each vendor who can access the audio and how long it is kept. Also ask whether it is used to train shared models.

How Do Groups Start with One Provider?

A rollout starts with one provider and one visit type, such as follow-up visits in a single clinic. The forward-deployed team builds the template from that provider's notes and reviews the edits on the first visits with the provider. Nothing is written to the EHR until the group approves the template and the write-back.

When the edit log for the first provider is steady, the next providers are added, then the next visit types. Groups on several EHRs add one system at a time. Across Flexbone deployments, the first agent goes live in about four weeks on average.

How Does a Flexbone Deployment Work?

Every deployment starts from the work as your team does it today, not from a product demo. The steps below are the same whether the first agent answers phones, checks eligibility, or files incoming documents.

  1. Audit first, then automate. We listen to a sample of your calls and read the work queues, then agree with your team on which work to automate first. The audit is complimentary.
  2. Forward-deployed engineers map your SOPs. An engineer sits with the staff who do the work and turns your written and unwritten rules into agent logic.
  3. Agents act on your rules, and exceptions go to a person. Each step follows logic built from your SOPs. A case outside those rules, or one the agent is not confident about, goes to your team with the context attached.
  4. Every action leaves a record. A check in a web application saves a screenshot, a call saves a transcript, and results are written back where your team already looks.
  5. Start with one or two agents on the systems you already run. The agents work in your EHR, practice management system, phone system, and the web applications your staff use, as they are configured today, so nothing is migrated or rebuilt.
  6. The same team stays after go-live. We monitor results, tune the agents as payers and workflows change, and add agents as new work is found.

The engagement model is described in more detail in forward-deployed engineering for healthcare contact centers.