Skip to main content

Voice agents

Custom AI voice agents for the calls your team can't get to

An AI voice agent answers your calls in natural speech and captures caller details in your systems. Sometimes called an AI receptionist, ours are custom built around your systems, and you sign off the wording it uses.

Claims line

  1. AgentGood afternoon, you're through to the claims team's AI assistant. How can I help?

  2. CallerHi, I'm just ringing to check on my claim. I put it in last week.

  3. AgentOf course. Could I take your claim reference, please?

  4. CallerYes, it's CL 40291.

  5. AgentThank you. And your date of birth, just to confirm it's you?

  6. CallerFourteenth of March, 1982.

  7. AgentThanks. Your claim is with a handler, who's waiting on the repair quote. I've noted that you called and I'll ask them to ring you back.

Example call. Written example, not a recording.

What an AI voice agent does

  • Answers calls when your team is unavailable

  • Captures caller details and passes them to your systems

  • Leaves notes for your team after each call

What your team gets

Each call leaves a record and notes for your team, and anything outside the rules is set to go to a person.

Record written

Call type
Claim status enquiry
Claim ref
CL 40291
Details checked
Yes
Next step
Handler to call back

Record and notes sent to your team

Example record. Fictional details.

How it runs

  1. 01Listen

    We hear which calls you miss and why.

  2. 02Map

    We agree its wording and when it hands over.

  3. 03Build

    Tested on calls written in advance, then alongside your team.

  4. 04Review

    We review handed-over calls and retest every change.

Built around your systems, with controls

Each agent is built around your systems, data and processes, so calls land where your team expects them. GME was founded by Greg East, a chartered accountant, so we start with controls and risk.

  • Wording you sign off

    Anything outside it is set to go to a person.

  • Built to reduce hallucination

    A narrow job and a hand-over rule, to reduce the risk of it making up an answer it does not have. No voice agent is error-free.

  • Controls and risk management

    What it can say, which systems it can reach and when it hands over are agreed with you.

  • Call analysis after go-live

    We review handed-over calls with your team and change the agent when they show a gap.

How we build and test a voice agent

Voice agent use cases

Claims intake

Captures the first report of a loss out of hours, in a set format, for your claims team.

Bookings

Books jobs and maintenance visits where your system allows it. Urgent calls are set to go to your on-call person.

Enquiries

Answers common questions in the wording you approve. Anything outside it is set to go to a person, and each call is logged.

Messages

Captures the caller's details and reason, so the right person can call back.

Questions

Will callers know it's AI?

Yes. It is set to introduce itself as an AI assistant.

Can it make things up?

Any AI can. We give it a narrow job and a hand-over rule for anything outside your scripts.

Will it work with your systems?

Usually, yes. We check exactly how before any build starts.

Where is call data kept, and who can see it?

Calls are recorded and logged where you choose. We agree who can see them before any build starts.

Is a voice agent the same as an AI receptionist or an IVR?

An AI receptionist is one kind of voice agent: it answers calls, captures caller details and relays messages. An IVR plays a menu and asks callers to press keys. A custom voice agent understands what callers say in their own words and can log a claim or book a job in your own systems.

Tell us about your calls

Tell us about the process you want to improve and we'll say whether we can help.