Head-to-Head Architecture Comparison
When a telehealth medical director evaluates ambient documentation, the decision rarely turns on whether a note reads well. Both Merry AI and Nabla Copilot produce a clinician-reviewed draft from captured audio. The real friction lives in three places: how the finished note gets into the EHR, how coding complexity is captured without a second pass, and how the record defends itself during an audit. This review works through each of those in the way a colleague would walk a chart with you.
The clearest structural difference is the write-back path. Nabla's published EHR integration guidance describes a front-end integration that exports the note through JavaScript events, or a back-end Core API integration requiring server-side development on the EHR. Both are export patterns. Merry AI writes directly into the focused EHR field through browser DOM injection, in the same tab, with no export button to configure.
| Feature | Merry AI | Nabla Copilot |
|---|---|---|
| Monthly Cost (Annual) | $54/mo Pro | $99/mo standard |
| DOM Injection Speed | Sub-second, direct to focused field | Export event or API queue |
| Multi-Speaker Group Note-Splitting | Speaker diarization, per-participant notes | Single clinician-reviewed draft |
| CPT G2211 Complexity Prompting | Real-time visit-close attestation prompt | Manual add-on beyond E/M suggestion |
| Attestation Logs | Cures Act aligned timestamping | Basic text export |
| Data Retention | Configurable, note-in-EHR default | Reported 14-day transient window |
| Specialty Templates | 1,000+ in Template Center | Specialty-gated E/M coverage |
Note Insertion and Workflow Friction
Telehealth documentation lives inside a browser tab, which is exactly where the write-back friction concentrates. Nabla's own documentation is candid that web-based EHRs use the existing Nabla UI and replace the copy or export action with an Export button that your EHR must listen for and map into the correct fields. That mapping is a project, and it is a project you repeat for every EHR variant on a distributed telehealth panel.
Merry AI removes that project by targeting the DOM element that currently holds focus. The note lands in the assessment field, the plan section, or the visit summary box precisely where your cursor sits. There is no server-to-server round trip, so the insertion is bounded by local rendering rather than an API queue. For a director standardizing behavior across many clinicians, the operational value is uniformity: the workflow is identical regardless of which web EHR a given clinician uses.
Zero-IT Deployment Across a Distributed Panel
Distributed telehealth teams rarely share a single IT footprint. Some clinicians run one EHR, others run a second, and credentialing spans multiple sites. Because Merry AI operates as a browser extension performing DOM injection, there is no per-EHR server integration to commission before a clinician can start. That matters for a medical director whose rollout timeline would otherwise be gated by each EHR vendor's integration backlog.
Latency Where the Patient Is Still On The Call
The goal for telehealth is finishing the note before the call ends. Sub-second insertion means the draft is in the chart while the patient is still visible, so review and sign-off happen in the same window as the encounter. This is where the estimated 2.5 hours of daily savings originates: not from faster typing, but from removing the reconciliation pass that export-based flows leave for end of day. You can pressure-test that estimate against your own panel in a 15-Minute Workflow Audit.
Coding Accuracy and CPT G2211 Capture
Coding is where telehealth revenue quietly leaks. Nabla surfaces an E/M code suggestion with a full Medical Decision Making breakdown, gated to a set of outpatient specialties, which the clinician reviews and copies into the EHR. That is a useful E/M aid. The gap is G2211, the longitudinal-complexity add-on, which is not an MDM computation but a determination about the continuity of the care relationship.
Merry AI treats G2211 as a first-class prompt at visit close, keyed to the continuity signals already present in the record. Because the prompt appears while the encounter is open, the clinician makes the attestation in context rather than reconstructing it later. Reimbursement values should always be confirmed against the current CMS physician fee schedule and the AMA CPT reference, but conservative panel math puts recoverable annual value above $15,600 per clinician.
Why Longitudinal Complexity Is Structurally Different
Single-visit MDM scoring answers a bounded question about one encounter. G2211 asks whether this visit is one node in an ongoing relationship, which requires the system to reason across the patient's history rather than the transcript alone. Merry AI's prompt is built to reference that continuity, which is why it is a distinct step rather than a line item folded into an E/M suggestion.
Compliance, Retention, and Audit Defensibility
Compliance posture is table stakes and then some. Nabla publishes HIPAA, SOC 2 Type II, ISO 27001, and GDPR coverage, and its correct framing is a clinician-reviewed documentation tool rather than an autonomous decision-maker. Merry AI shares that framing: every note is reviewed and signed by the clinician before it enters the record, under BAA coverage for every practice.
Retention is where the two diverge in a way a director should note. A privacy assessment in one Nabla deployment reported a 14-day transient retention window for transcripts and structured notes. Merry AI's default keeps the note in the EHR as the system of record and minimizes transient retention, so the durable copy lives where your governance already applies. Each note also carries a Cures Act aligned attestation block recording generation, review, and sign-off events, which is the trail a payer audit actually asks for.
Pricing and Total Cost of Ownership
Price should be read as total cost, not headline rate. At $54 per month on the Pro annual plan versus a $99 standard rate, the direct software line is roughly 45% lower, but the larger difference for telehealth is the absence of per-encounter API metering. Export-and-API architectures can accrue variable costs and require IT time to maintain each integration path; a flat per-clinician rate makes budgeting across a growing panel predictable.
For a medical director running the numbers, the honest comparison is the sum of software cost, integration engineering, and recovered coding revenue. You can review the tiers directly at Compare Practice Partner Plans, stock your specialties from the 1,000+ Specialty Clinical Prompts library, and then bring your own panel data to a 15-Minute Workflow Audit so the estimate is grounded in your encounter volume rather than ours.
