AI voice agents for patient calls

The problem

Clinical staff spent hours on routine calls: pre-op intake, post-procedure follow-up and appointment scheduling. The calls had to be handled without moving patient data outside the provider's perimeter.

What we built

Voice agents that place and take the calls, transcribe the conversation, determine the next action and write it back to the EMR. A person stays in the loop on anything clinical.

What we owned in production

Turn-taking and interruption handling. Escalation rules for anything outside the script. Scripted conversation evals on every prompt change. Audit logging of every decision. Operated under HIPAA and SOC 2 controls.

Scope

Workflow-readiness assessment, agent design and build, evaluation and guardrails, hybrid deployment, production rollout and support.

Client name withheld under NDA.

Have something similar on your roadmap?

Start a conversation →

GOOD TO KNOW

Frequently asked questions

What is the AI voice agents for patient calls project?

Production voice agents that take routine patient calls off clinical staff and write the next action back to the EMR.

More selected work

What problem did this work address?

Clinical staff spent hours on routine calls: pre-op intake, post-procedure follow-up and appointment scheduling. The calls had to be handled without moving patient data outside the provider's perimeter.

What did Digital Gateway build?

Voice agents that place and take the calls, transcribe the conversation, determine the next action and write it back to the EMR. A person stays in the loop on anything clinical.

What was included in the delivery scope?

Workflow-readiness assessment, agent design and build, evaluation and guardrails, hybrid deployment, production rollout and support.

Which sector was this project in?

The case is categorized under healthcare.

Browse cases by work

Which technologies are named in the case?

The published stack lists Azure AI Speech (STT/TTS), Azure OpenAI, Azure Bot Framework, Azure Communication Services, Azure Health Data Services / FHIR, ONNX Runtime, Databricks, Python, .NET, HL7 / FHIR.

What team is listed for the project?

The published case lists PM, AI engineer, back-end engineer, evaluation engineer, QA.

How delivery works

How was the system deployed or operated?

Hybrid. PHI-touching steps on local open-weight models; only de-identified reasoning in a HIPAA-covered Azure OpenAI tenant.

Is the client identified by name?

The case describes the client as Healthcare provider, United States. The public case withholds the client name under NDA.

Which Digital Gateway services relate to this work?

This case is tagged to AI agents and voice and Embedded engineering.

AI agents and voice Embedded engineering

Can Digital Gateway help with a similar requirement?

Yes. Share the intended outcome, existing systems, constraints and operating requirements so we can discuss a suitable delivery scope.

Discuss a project