Head-to-Head Architecture Comparison
When cardiologists evaluate documentation systems, the bottleneck is rarely prose quality. The real bottleneck is the last inch: moving an ejection fraction, a functional class, and a medication history from a draft into the signed note without retyping. Suki AI is a capable ambient assistant built for enterprise EHRs such as Epic, Oracle Health, athenahealth, and MEDITECH, with developer SDKs and a cardiology specialty profile. Merry AI is a browser extension that works on the chart you already have open. The table below shows where those design choices diverge.
| Feature | Merry AI | Suki AI |
|---|---|---|
| Real pricing | $54/mo Pro Annual ($648/yr), flat | Unpublished; commonly reported $299 to $399 per provider per month under contract |
| DOM injection speed | Under 1 second into the open Practice Fusion note | No documented Practice Fusion write-back; copy-paste required |
| HF discrete capture | LVEF % with date, NYHA class, and GDMT doses enforced | Cardiology vocabulary; no enforced HF block documented |
| Group note-splitting | Diarized per-patient notes for cardiac rehab education groups | Single-encounter model; not publicly documented |
| CPT G2211 capture | Flagged with a modifier 25 conflict check | Not publicly documented |
| Attestation logging | Clinical Logic Bridge block with source mapping | Relies on the EHR audit trail |
| Data retention | Audio deleted after note finalization; text per practice policy | Set by enterprise contract and BAA |
These figures reflect public information as of October 2026. Suki does not publish list pricing, so the range shown is what practices commonly report for per-provider contracts. Confirm with the vendor before modeling. Where a Suki capability is listed as not publicly documented, we mean exactly that, not that it cannot be built.
Practice Fusion Chart-Tab Injection
Merry attaches to the active Practice Fusion chart tab. It scopes itself by URL and container markers and listens for navigation between the Summary, Medications, and Encounter subtabs. When the encounter note editor loads, including editors rendered inside iframes or contentEditable regions, Merry locates the target section and writes the draft directly. It then dispatches input and change events so Practice Fusion autosave registers the edit.
Because the write occurs inside the clinician's authenticated session, the EHR audit trail records the change under that clinician. Inserted text stays visible and editable until signature. Suki's documented handoff model relies on sanctioned partner integrations. For an independent cardiology group on Practice Fusion, that usually means copying the note out of Suki and pasting it into the chart.
Heart Failure Documentation: LVEF, NYHA, GDMT
Heart failure visits carry three data points that drive phenotype, therapy eligibility, and coding defensibility: LVEF with the study date, NYHA functional class, and the current GDMT regimen at maximal tolerated doses. Suki's cardiology profile recognizes terms like ejection fraction, but its public documentation does not describe enforcing a discrete HF summary block. Merry treats that block as required for any visit tagged with an HF diagnosis.
A typical injected line reads: HF summary: LVEF 30% by TTE on 08-14-2026 (HFrEF), NYHA Class II. GDMT: sacubitril/valsartan 49/51 mg BID, carvedilol 12.5 mg BID, spironolactone 25 mg daily, empagliflozin 10 mg daily; furosemide 20 mg as needed. If the echo date or functional class is missing from the conversation, Merry leaves a highlighted placeholder rather than guessing.
Maximal Tolerated GDMT and Contraindications
The phrase maximal tolerated GDMT only holds up under review when the note explains why a dose stopped where it did. Merry captures the stated reason for each class not at target. Examples include symptomatic hypotension limiting ARNI uptitration, potassium of 5.4 mmol/L holding the MRA, or eGFR below an agent's initiation threshold. This supports quality measures and makes the longitudinal reasoning explicit for the next visit.
Cost Math and CPT G2211 Recovery
Merry AI Pro Annual costs $54 per month, or $648 per clinician per year, with no per-encounter metering. At the low end of commonly reported Suki pricing, $299 per month, a clinician pays $3,588 annually. The difference is $2,940 per clinician, roughly 82 percent lower overhead. At $399 per month the gap widens to $4,140. For reference, Freed lists around $99 per month, and Nuance DAX enterprise contracts often run near $8,000 per year.
G2211 recognizes the longitudinal relationship inherent in managing a chronic condition like heart failure. At an approximate national rate of $16 per claim, a cardiologist reporting it on 975 qualifying established-patient visits per year, about 20 per week, recovers roughly $15,600. Merry flags the add-on only when the note shows continuing responsibility for the condition. It is never applied as a default.
Same-Day Echo and Modifier 25
Cardiology clinics frequently pair visits with same-day echocardiography or device interrogation. Under current CMS policy, G2211 is not payable when the E/M carries modifier 25 for a separately payable procedure. The exceptions are preventive services, vaccine administration, and annual wellness visits. Merry separates the HF assessment from the procedural note and warns when modifier 25 would make G2211 unpayable. It never suggests modifier 59 on an E/M code, because NCCI reserves that modifier for distinct procedures.
Compliance, Attestation, and Next Steps
Every Merry note closes with an attestation block from the Clinical Logic Bridge. The block records which statements were drafted, which source utterances support them, and when the clinician reviewed and signed. California AB 3030 and Texas HB 1709 point toward disclosure and clinician accountability for AI-assisted content. Practices should still map these obligations with counsel and confirm that their BAA covers browser-level PHI access.
For the underlying evidence base on ambient documentation, burnout, and note accuracy, the literature indexed at https://www.ncbi.nlm.nih.gov/pmc/ is a reasonable place to start. We would rather you check our reasoning against primary sources than take a comparison page at face value.
If you run heart failure clinics on Practice Fusion, we can review three of your recent HF notes against this framework. We will show where the LVEF, NYHA, and GDMT documentation is complete and where it is thin. Schedule a 15-Minute Workflow Audit and bring a chart you find tedious.
