Frequently asked questions

Most practices are running their first workflow within a few days of contract. There's no IT project, no integration build, and no on-prem install — patient lists go in via CSV, and the workflow uses an off-the-shelf clinical template tuned to your protocol.

No. Healthline AI is intentionally designed to run without EHR integration so you can pilot it without a procurement project. We work from patient lists you upload and produce structured outputs (call summaries, outcomes, escalations) that your team can paste, attach, or import back.

For larger deployments we can integrate, but it's never a prerequisite.

Live in production: statin adherence, cervical cancer screening, bowel cancer screening.

Coming soon: chronic disease management (diabetes, COPD, hypertension), NHS health checks, pre/post-operative care, and inbound triage. If you have a specific workflow you'd like prioritised, tell us — we shape the roadmap with our partner practices.

We price per workflow, scaled to your list size, with a single fixed onboarding fee. We don't publish pricing publicly because the right structure depends on your scale and which workflows you run — talk to us and we'll send a clear quote within a working day.

A natural-sounding, human-paced voice agent that introduces itself as calling on behalf of your practice, follows your clinical script, and handles the conversation in real time. Patients can ask questions, decline, or request a callback — the agent adapts.

Average patient satisfaction across our live workflows is 98%.

The agent escalates immediately, on rules you configure — for example, side-effect concerns, safeguarding flags, or anything outside the protocol. The escalation appears in your queue with the full transcript and a structured summary so a clinician can pick it up cleanly.

Yes. The agent identifies itself transparently at the start of every call, in the wording your practice approves. Patients can opt out at any point.

All data is processed under a Data Processing Agreement with the practice as data controller. Storage is UK-based, encrypted in transit and at rest. We do not use patient data to train external models.

We are working through Class I SaMD classification with a parallel pathway toward Class II. Every workflow includes a full clinical risk file, audit trail, and post-market surveillance loop.

Yes — every interaction is structured and exportable as CSV or PDF, including transcripts, outcomes, escalations, and timestamps. The audit trail is designed to plug straight into CQC evidence and NHS reporting requirements.