Case study
Voice agent demos that turned prospects into customers
I built 20+ voice agent demos in four months, each shaped around a real prospect. Two of those prospects signed, one of them a US enterprise.
- Role
- AI automation consultant
- Built
- 20+ demos in 4 months
- Result
- 2 prospects became customers
- 01A call comes in, or the agent dials out
- 02The agent pulls the customer’s record from Salesforce
- 03It talks to them in their language
- 04It books an appointment or hands over to a person
- 05The outcome goes back into Salesforce
The ask
The sales team at a voice AI platform kept meeting prospects who wanted to see their own use case working before they would sign. My job was to build those demos.
One of them was a US company. They wanted an agent that could call their existing customers and also pick up inbound calls, using the customer data they already kept in Salesforce.
What I built
I built every demo on the platform itself, shaped around the calls that prospect actually handles. For the US company, the agent looked up the caller in Salesforce before it said a word, and wrote the outcome back once the call ended.
Across the demos I covered outbound and inbound calls, several languages, handing the call to a human, and booking appointments.
How I tested them
Voice agents break in ways you don’t see when you test by reading a script. People interrupt, change their minds and go off topic. I used Cekura to run simulated calls against the agents and catch those problems before a prospect did.
Result
I was on the demo calls with the sales team, so I heard the buyers’ questions first-hand. Two of the prospects I built for signed on as customers, including the US company on an annual contract.
The platform, its customers and the deal figures are confidential, so they’re left out here. The flow above is a general picture, not a copy of any client build.