AI phone and chat agent that captures leads after hours

The problem

The client helps seniors and their families plan, downsize, pack and settle into a new home. Their inbound line ran through a consumer voice service, and every call the small team could not pick up went to voicemail. We were brought in to make sure no enquiry was lost outside business hours.

• 01

  Calls arriving after hours, at weekends, or while the team was on a move went to voicemail and often were never returned.

• 02

  Callers are frequently seniors or family members under stress, and a voicemail prompt was a poor first experience at a difficult moment.

• 03

  Lead details arrived as unstructured voicemail audio, so nothing could be reported on or routed automatically.

• 04

  Every new enquiry had to be re-typed by hand into the industry CRM the team runs their jobs from.

• 05

  Consultation bookings depended on a human being free to call back and find a slot, adding days to the response time.

What we built

We built the agent inside the CRM platform the client already paid for, rather than assembling a separate voice stack, so the team had one place to see calls, contacts, calendars and automations. The agent was designed to behave like a calm team member rather than a phone tree: it collects contact details first, answers only from a curated knowledge base, and never quotes prices or invents a service area. A parallel website chat agent shares the same knowledge base and capture rules so both channels give the same experience.

• Voice agent

  Answers inbound calls 24/7, collects contact details, qualifies the enquiry, and books a consultation call live on the phone.

• Website chat agent

  Same persona and knowledge base on the client's site, with matching capture and booking behaviour.

• Knowledge base

  Curated FAQ and company overview so the agent answers from approved wording only and defers anything it does not know.

• Field extraction layer

  Writes ten structured fields per qualified enquiry onto the contact record after each call.

• Booking integration

  Round-robin consultation calendar that respects each staff member's connected Google Calendar, so blocked time is never offered.

• CRM bridge

  Post-call automation that pushes booked leads, with full transcript, into the client's industry CRM.

How it works

From trigger to result.

The client's industry CRM had no public API at the time of build, so the bridge was built through a workflow automation connector. A REST API was released by that vendor afterwards, which is the planned upgrade path.

Our approach

How we worked.

1. Discovery

   Reviewed the client's existing call handling, service area, pricing rules and CRM structure, and confirmed what the agent was allowed to say.

2. Agent build

   Wrote the persona and prompt, loaded the knowledge base, and wired field extraction onto the contact record.

3. Booking and calendar

   Built the shared consultation calendar, connected each staff member's Google Calendar, and wired live booking into the agent.

4. CRM integration

   Built the post-call automation and middleware bridge into the industry CRM, including transcript delivery.

5. Testing and iteration

   Ran repeated live calls and chats with the client, tightened capture accuracy, and applied client feedback before go-live.

The outcome

What changed.

• Calls arriving after hours, at weekends, or during a move are now answered and qualified instead of going to voicemail.

• Callers can book a consultation during the call itself rather than waiting for a return call.

• Booked leads appear in the CRM automatically with the full call transcript attached.

• The team decides what the agent may say on pricing and service coverage, so it never guesses or over-promises.

Questions

What people ask.