A lightweight agent installs on your own machine and runs on a schedule. It transcribes call recordings and redacts PHI on-device — only anonymised text is ever sent to MyAutobot for sentiment, QA, compliance and benchmarking.
Caller: I need to rebook [PATIENT_NAME] — DOB [DOB], NHS [NHS_NUMBER].
Built in collaboration with our sister company Integrella, Trusted by
Everything sensitive happens inside your clinic. Only redacted, anonymised text is sent on for analysis.
Install the agent on a clinic machine and point it at your recordings. Set it to run nightly, hourly, or one time.
Each call is transcribed on the machine and PHI is replaced with typed tags like [PATIENT_NAME]. Audio never leaves.
Only anonymised text syncs to your dashboard for sentiment, QA, compliance and benchmarking — then writes back to your CRM/PMS.
Every call scored, with negative and high-risk conversations routed to a manager the moment they're processed.
Objective scoring against your scripts and standards turns every call into targeted coaching material.
Consent, identity and vulnerable-patient checks on every call, with a full, exportable audit trail.
See how your team compares to your industry — and to a MyAutobot AI voice agent — on the metrics that matter.
A look at a real clinic dashboard. Everything is aggregate — reported on call types and trends, never about an identifiable patient.
A single deep-dive across a year of recordings — perfect for a baseline or board review.
Strong — use as the model script for new starters.
Pharmacy redirects cause confusion — a one-page process sheet would lift this.
Callers often ring back — clearer "what happens next" wording is the win here.
Solid — payment answers vary by staff member; a shared crib sheet would help.
Friction flags: 6% of calls showed signs of caller frustration — most in prescription queries. Reported in aggregate only, never about individual patients or staff.
1 in 8 calls rings out at the Monday 8–10am peak — one extra front-desk hour would catch most of them.
Feed these straight into your website, phone menu or front-desk scripts.
In 7 of 10 flagged calls, sentiment recovers by the end — your team is de-escalating well. Agent tone stays calm even when callers don't.
Every chart is built from anonymised text. DialVue reports on call types, trends and team coaching — never on identifiable patients, and staff insights stay private to your managers.
DialVue back-tests your real calls against your industry average and a MyAutobot AI voice agent, so you get an evidence-based view of where automation helps — and where your people already shine.
Run a benchmarkFirst-call resolution across 1,240 analysed clinic calls this month.
Insights and call summaries push straight into your PMS/EHR and CRM. Build your own with a REST API and webhooks.
curl -X POST https://api.myautobot.ai/v1/transcripts \ -H "Authorization: Bearer $API_KEY" \ -H "Content-Type: application/json" \ -d '{ "call_id": "clinic-4827", "redacted_text": "Caller: rebook [PATIENT_NAME]…", "duration_sec": 252, "sentiment": "negative" }'
Transcription and redaction run on your machine — effectively free. You only pay for the cloud analysis you switch on, priced per minute.
Trial it on real clinic calls
Small clinics
Multi-site groups
Trusts & large groups
Windows desktop agent · ~40 MB · free 30 minutes included · no card required.
Runs on Intel & AMD PCs — and on Windows on ARM via built-in emulation. macOS & Linux coming soon