Head-to-Head Architecture Comparison
When clinicians evaluate documentation systems, the real bottleneck is rarely note formatting. It is the distance between a finished draft and the chart field where it must live, along with the billing evidence that draft carries. Nabla and Merry AI both produce competent ambient notes for primary care. The differences appear in three places: cost structure, where text lands inside athenaOne, and how longitudinal complexity is written down.
| Feature | Merry AI | Nabla |
|---|---|---|
| List pricing | $54/mo (Pro Annual) | $119/mo (Pro, monthly list) |
| Annual cost per clinician | $648 flat | $1,428 |
| athenaOne write path (DOM injection speed) | Chrome DOM injection into the active encounter field, sub-second in internal testing | One-click 'Export to Athena' after the visit closes |
| Group note-splitting | Per-participant notes from one diarized session | Single-encounter note; no published split |
| CPT G2211 capture | Visit count, titration plan, and named clinicians prompted | General complexity in A/P; no dedicated G2211 module published |
| Attestation logging | Timestamped clinician attestation block per note | Not published as a note-level block |
| Data retention | Audio discarded after finalization; the note lives in the EHR | Audio not stored by default; confirm terms in BAA |
Nabla's athenaOne export is legitimate, and it deserves a fair reading. Its integration pushes HPI, physical exam, assessment and plan, and social history into the chart once the visit ends. Merry AI takes a different route. The Chrome extension writes into the DOM input selector of the encounter you are actively documenting. It runs inside your existing authenticated session and enters the text as though you had typed it yourself.
Injecting Into the athenaOne Assessment Field
The practical difference is timing. Export happens after the visit closes; injection happens while the chart is open and the patient is still in the room. Merry AI locates the assessment and plan field and traverses any embedded iframe to reach it. It then populates the text without bypassing athenaOne's own save and lock behavior. The clinician reads, edits, and signs in the native chart. No marketplace API fees are involved, and no separate portal tab competes for attention.
The Pricing Math, Line by Line
Nabla's Pro tier lists at $119 per clinician per month at the time of review, or $1,428 per year. Larger groups typically negotiate enterprise terms, so treat that figure as the published reference point for small practices. Merry AI Pro Annual is $54 per month, or $648 per year. The difference is $780 per clinician annually, a 54.6% reduction. A six-physician practice keeps $4,680 each year.
For context across the category, Freed lists at $99 per month. Nuance DAX Copilot enterprise contracts are commonly quoted near $8,000 per clinician per year. Merry AI charges a flat rate with no per-encounter metering. That matters when a clinician sees 22 patients on Monday and 9 on Friday. Predictable spend is also easier to defend in a practice budget meeting than variable usage tiers.
G2211 Evidence in Primary Care
G2211 is an add-on code describing the complexity of office or outpatient E/M visits (99202–99205, 99211–99215). It applies when the visit serves as the continuing focal point for a patient's care, or when it is part of ongoing care for a single serious or complex condition. Coding definitions and E/M structure are maintained at https://www.ama-assn.org/practice-management/cpt. Primary care is the canonical setting for this code. It is also the setting most exposed to vague documentation.
Replacing Generic Chronic Language
'Chronic conditions, stable, continue current' is the sentence auditors see most often, and it proves almost nothing. Merry AI structures the longitudinal paragraph around three verifiable elements:
- The number of visits in the prior 12 months.
- The active titration plan, with doses and the next step.
- The named coordinating clinicians.
A typical output reads: 'Seen 5 times in 12 months; lisinopril increased 10 to 20 mg, BMP in 2 weeks; co-managing with Dr. A. Patel, cardiology, LVEF 38%.' These values come from the chart and from clinician confirmation. The extension prompts for them and never invents them.
The revenue arithmetic is modest but cumulative. At a national average near $16 per claim, roughly 975 supported G2211 visits a year, about four per clinic day, yield $15,600 or more per physician.
Modifier 25 Guardrails
Since January 1, 2025, CMS pays G2211 alongside a Modifier 25 E/M only when the same-day service is a Part B preventive service, an immunization administration, or an annual wellness visit. Merry AI checks the encounter composition and flags G2211 suggestions that fall outside those combinations. This avoids encouraging a claim that is likely to be denied.
Audit Posture, Retention, and Next Steps
Every Merry AI note carries an attestation block recording the timestamp, clinician identity, and sections edited before signature. The Clinical Logic Bridge maps each G2211 element to the exact sentence that supports it. A reviewer can trace the claim without rereading the full note.
On retention, precision beats slogans. Audio is processed transiently and discarded after note finalization. The signed note lives only in athenaOne. Nabla publishes a similar no-audio-storage default, so confirm current terms in either vendor's BAA. For California practices, patient-facing text generated this way still requires AB 3030 disclosure.
If you want to verify this, bring two real encounters and an athenaOne login. Schedule a 15-Minute Workflow Audit and we will measure both workflows against your own charts.
