Software & Tools

Best AI Scribes for Dietitians: Heidi, Twofold, and the Note-to-Claim Gap

An honest comparison of AI scribes for dietitians — what Heidi and Twofold do well, how to evaluate them, and the note-to-claim gap that insurance practices still pay for.

By Alva Health Team · Published July 12, 2026 · ← All articles

Disclosure: this article is written by the team at Alva, which offers AI charting as part of a broader billing platform — so we compete with the products below on one feature. We've tried to be fair; verify details with each vendor.

AI scribes solved a real problem for dietitians. The 15–20 minutes of charting after every session — reconstructing a 60-minute conversation into an ADIME or SOAP note while the next client waits — was the single most hated block of time in private practice. Tools like Heidi and Twofold genuinely fixed it: record the session with consent, get a structured draft note minutes later, edit, sign.

If you run a cash-pay practice, this article ends happily: buy a good scribe and enjoy your evenings. If you bill insurance, there's a second question the scribe marketing never raises — what happens to the note after it's written? — and the answer determines whether the scribe actually gave you your time back.

What scribes do well

Credit where due. The better scribes on the market, Heidi and Twofold included, are good products:

• Audio to note, reliably. Modern medical-grade transcription handles clinical vocabulary well, and the structuring layer produces coherent SOAP or ADIME drafts rather than raw transcripts.
• Templates. Both support customizable templates, so you can shape output toward nutrition assessment, PES statements, and follow-up formats instead of a generic medical H&P.
• BAAs. On appropriate paid tiers, established scribe vendors sign Business Associate Agreements — the non-negotiable requirement for putting PHI through any tool.
• Speed. The realistic outcome is turning 15–20 minutes of charting into 2–5 minutes of review and editing. Multiplied across 20+ sessions a week, that's real hours.

How to evaluate a scribe

Pricing and features change often, so treat this as a rubric to fill in with current vendor information rather than a fixed scoreboard:

Both Heidi and Twofold score respectably on this rubric for dietitian use; the right pick depends on which templates match your documentation style and what your budget tier includes. We've written more on what to look for in AI charting for dietitians.

The note-to-claim gap

Here's the architectural point this category doesn't advertise: a scribe's pipeline ends where an insurance practice's pipeline begins.

The note is not the deliverable. The paid claim is. Between them sits everything the scribe leaves untouched:

1. Choosing 97802 vs. 97803 and counting time-based units correctly
2. Pairing ICD-10 codes to the payer's coverage policy
3. Confirming the documentation supports medical necessity
4. Filling and submitting the CMS-1500 through a clearinghouse
5. Tracking the claim, catching rejections, working denials
6. Posting the ERA and reconciling the payment

The scribe compresses step zero — writing the note — from 20 minutes to 3. Steps one through six, which commonly take 15–25 minutes per claim plus batch time for tracking and posting, remain exactly as manual as before. For an insurance-based RD, the scribe optimized the cheapest segment of the workflow and left the expensive one alone.

Worse, the gap is where the money leaks: a beautifully written note with the wrong unit count still gets denied, and a denial costs a full session's revenue until someone reworks it.

A decision framework

If you're cash-pay: a scribe is probably all you need. Your documentation burden ends at the note; there is no claim pipeline to feed. Pick whichever of the credible vendors matches your template style, confirm the BAA, and stop reading vendor comparisons.

If you bill insurance (or plan to): do the math on the whole pipeline, not the note. A $75/month scribe that saves 5 hours of charting while leaving 15 hours of coding, submission, and posting untouched has solved a quarter of your problem — and you'll still be shopping for the rest of the stack, with yourself as the glue between the pieces.

If you're hybrid: weight by revenue. If insurance is where your growth is, the note-to-claim gap grows with you.

Note-to-paid-claim, as one system

The alternative architecture is a platform where the recorded session doesn't stop at a note: the same pipeline that drafts the chart also generates the CPT/ICD codes and units from it, validates the claim against payer rules, submits it, tracks it, and posts the payment when the ERA arrives. That's the design gap Alva was built to close — the scribe is one stage of the pipeline instead of the whole product.

That doesn't make standalone scribes bad products. It makes them partial products for one specific audience: the insurance-based RD, for whom the note was never the finish line.

How Alva helps: Alva includes AI charting from your recorded sessions — but keeps going, turning the note into validated codes, a submitted claim, and a posted payment without you re-keying anything. For $99/month you replace the scribe and the manual billing work that follows it. Start a 7-day free trial — no credit card required.