Pre-launch draft. Demo encounters are synthetic and have not yet been reviewed by a clinician; accuracy figures are not yet published.

P1 · Administrative

Bolo Front Desk

A receptionist that answers in the caller's language.

Book a pilot call

Inbound call · 00:09 elapsed · Telugu + English

Not clinically reviewed

Anything clinical transfers to a human immediately. Front Desk does not triage and does not answer “should I come in?”.

Synthetic demo call. Not a real caller.

1st

place the language gap costs the practice money

The front desk is the first place the language gap costs the practice money. Calls take longer, get handed between staff to find someone who speaks Gujarati, and are abandoned when nobody does. The patients who hang up are the ones least able to get care another way.

How it works

  1. 01

    Detect the caller's language from the first utterance, at frame level — not from a menu.

  2. 02

    Route intent through a deterministic state machine. Booking is not an LLM with function calls.

  3. 03

    Offer a genuine trade-off: a language-matched provider later, or an earlier slot with an interpreter.

  4. 04

    Read the appointment back in the caller's language, take an explicit confirmation, then commit with an idempotency key.

What it will not do

Stated plainly, because a limit you find out about later is a limit you stop trusting the rest of the page over.

  • It does not triage. It never answers “should I come in?” and never says a symptom sounds fine.
  • It never approves a refill. A refill request becomes a task for staff.
  • It never routes a patient to a provider on the basis of presumed ethnicity or language. Preference is stated by the patient, captured explicitly, and is never a constraint on earlier care.
  • Anything clinical transfers to a human immediately, from any state in the call.

Integration

Every buyer asks within ninety seconds of the demo starting. A plain FHIR path exists today — “here is the bundle” is a real answer, and “we are in the marketplace review process” is not.

FHIR Appointment, Slot
Idempotency key on every write, so a retry cannot double-book.
FHIR Task
Refill requests and callback capture.
SIP / PSTN
Barge-in supported throughout — elderly callers interrupt.
DTMF fallback
Three failed recognitions fall back to keypad, then to a person.

What we measure

These are the numbers for this product. None of them are published yet — see the methodology for why an unstratified figure would be worse than silence.

Containment rate
Percentage of calls resolved without a human.
Mis-routed clinical calls
Target is zero. Any non-zero value is reported.
Booking accuracy after read-back
Measured after confirmation, because that is what the caller agreed to.
Latency p50 / p95 / p99
Above roughly 800 ms callers start talking over the system.

What you are probably about to ask

Our patients hang up on phone robots.

They hang up on phone robots that do not understand them. They also hang up when they cannot interrupt, so barge-in is supported throughout, and after three failed recognitions the call falls back to keypad and then to a person. The call has to succeed even when the speech layer does not.

Is routing patients by language a discrimination problem?

It would be if we inferred it. We do not. Language preference is stated by the patient, captured explicitly and stored as a preference — and it is never a constraint. The system offers the earlier appointment with an interpreter alongside the later one with a language-matched provider, and the patient chooses.

What if someone calls with chest pain?

It transfers to a human immediately and does not attempt to assess anything. The red-flag lexicon fires on colloquial phrasing in every supported language, not just textbook clinical terms — that lexicon is built by linguists and reviewed by clinicians per language.

Pilot

Book a pilot call.

Thirty minutes. We will run the demo on the languages your panel actually speaks, show you the eval methodology, and tell you plainly which parts are not ready. If we are not a fit we would rather find out on this call.

One reply from a person. We do not run a nurture sequence and we do not share your address.