AI Scribe for Concierge & DPC Practices

Specialty Documentation Architecture for Membership Medicine

Clinical documentation in Concierge and DPC practices operates on a fundamentally different axis than fee-for-service medicine. The record is not a series of discrete, billing-driven encounters but a continuous longitudinal relationship in which the member's own priorities, verified clinical facts, and clinician judgment must remain traceable and separable. Merry AI is engineered around a single anchor truth: it converts member-provided priorities and verified clinical information into a clinician-reviewed longitudinal membership summary and care-goal record, never a delegated clinical decision.

The architecture must separate three streams that membership practices too often blend into a single unstructured PDF. Membership administration—plan tier, recurring fee, effective date, renewal status—lives apart from clinical care, which holds problems, medications, preventive-care status, and care goals. Legal documentation—the membership agreement version, informed consent, Notice of Privacy Practices, and telehealth consent—forms the third stream. Merry AI respects these boundaries so billing data is never mistaken for clinical fact when a member exercises their right of access.

Provenance labeling sits at the center of the entire model. Every material statement Merry AI drafts carries a status tag: Confirmed, Member-reported, Clinician-assessed, Imported from external source, Pending verification, or AI-generated draft. The system will not silently convert uncertain information into fact, and it will never attribute a goal to a member the member did not state. This discipline is what makes the generated summary defensible when it later becomes part of the designated record set.

The Clinical Logic Matrix

Membership documentation must bridge three simultaneous obligations: the diagnostic framework, the clinical data required to satisfy it, and the billing evidence that defends longitudinal complexity. The matrix below maps how Merry AI structures each layer for a continuity-focused practice.

This matrix operationalizes the guidance found in peer-reviewed and federal sources. The continuity logic underpinning G2211 is drawn directly from CMS documentation, and the broader documentation-burden framework aligns with evidence catalogued in the National Library of Medicine's PubMed Central repository, where E/M simplification research demonstrates that structured, clinically relevant capture reduces clinician burnout without sacrificing audit defensibility.

Preserving the Member's Voice as Structured Data

Member-authored goals are the most valuable and most easily corrupted element of a concierge record. A member may say, 'I want to improve my sleep enough to avoid afternoon fatigue.' Merry AI captures that statement verbatim, tags it member-authored, and only then converts it into a structured care-plan object with a category, baseline, desired outcome, review date, and clinician-approved interventions. The member's language and the clinician's interventions remain permanently distinguishable.

Longitudinal goal tracking is where the anchor capability earns its keep. Goals persist across every encounter rather than appearing once at onboarding. Merry AI surfaces status prompts—'Goal progressing,' 'No recent update,' 'Goal requires clinician review,' 'Potentially overdue follow-up'—that support the clinician's judgment without making autonomous clinical decisions. This is assistive documentation, not delegated care.

Draft-State Generation and Statement-Level Sign-Off

AI-generated documentation always enters a draft state. A summary does not become part of the permanent medical record merely because it was generated. The clinician reviews source references, accepts, edits, or rejects each statement individually, corrects medication and allergy data, confirms or modifies goals, and records the date and author of approval. The original AI draft and every clinician change are preserved in an immutable audit trail.

Dual-View Summary Generation

Membership summaries serve two audiences, so Merry AI produces two synchronized views from the same approved source data. The clinical longitudinal summary retains full detail and provenance for care coordinators and clinicians. The member-facing summary uses understandable language, excludes unnecessary sensitive information, and forms an appropriate after-visit document—each respecting the minimum-necessary principle central to HIPAA.

Zero-Friction Integration with Closed-Garden EMRs

DPC and concierge platforms are frequently closed browser environments—Hint, Elation, Atlas MD, and similar membership EMRs—that lack open FHIR APIs. Rather than forcing a vendor integration queue, Merry AI's Chrome Extension performs DOM injection, writing the finalized structured note directly into the active EHR field inside the clinician's own browser window. There is no IT ticket, no API configuration, and no synchronization lag.

Interoperability remains standards-aware where the platform allows it. When FHIR endpoints or CCD export exist, Merry AI normalizes medication and allergy data, imports lab and imaging results, and tracks referral status with documented provenance. Every imported field is labeled so the clinician verifies external data before it is treated as confirmed.

Governance, Compliance, and the Anchor Truth

A formal AI governance policy governs every permitted and prohibited use. Merry AI may draft membership summaries, organize member-defined goals, identify missing documentation, and flag overdue follow-up for human review. It is expressly prohibited from autonomous diagnosis, autonomous prescribing, emergency triage without qualified oversight, unreviewed changes to the legal medical record, or inferring sensitive conditions no authorized source documented.

HIPAA and state DPC obligations frame the deployment. The practice designates Privacy and Security Officers, completes a risk analysis, executes BAAs with vendors including Merry AI, and maintains right-of-access, amendment, and secure-deletion workflows. Because DPC law varies by state, agreement elements—scope, exclusions, the statement that the arrangement is not insurance, cancellation terms, and Medicare opt-out where applicable—must be validated against local requirements. Ready to see this architecture in your practice? Schedule a 15-Minute Specialty Workflow Audit and Access Specialty Prompts at templates.scribing.io.

The central compliance principle is unwavering: assistive documentation, not delegated clinical judgment. Merry AI creates a personalized summary from verified data, preserves the member's own goal language, distinguishes it from clinician recommendations, sources every material statement, labels output as draft until signed, records the approving author and timestamp, and updates the longitudinal record without erasing prior versions. That framing aligns the tool precisely with the operational and legal reality of membership medicine.