Reps rehearse the hard calls before they happen

A voice roleplay agent that lets sales reps rehearse hard customer conversations, scores each session on objection handling, rapport, product knowledge and next steps, and sends the scorecard to Slack, as Impel built on Claude

The problem

A new sales hire learns the hard conversations by losing a few real ones first, and the owner is the only coach in the company. Ramping takes months, and the objections that cost deals are the same ones every time.

How the agent works

1. It starts

   The team writes practice tracks from its own recorded calls: the objections, the awkward pricing moments, the support complaints that turn into churn.

2. The agent works

   A rep opens the agent and talks to it by voice; speech to text feeds Claude, which plays the customer in the chosen track, and text to speech answers back. When the session ends the agent scores it on objection handling, rapport, product knowledge and next steps, and sends the scorecard as a direct message in Slack. Nobody is forced to practise; reps use it on their own time, and the manager reads the scorecards and picks what to coach.

3. The result

   The human checkpoints are the manager's review of the tracks before they go live and of the scores before they count for anything.

Worked example

Anthropic's case study on Impel, which builds AI for automotive dealers and has over 150 customer facing reps, describes a dealer roleplay agent built in about 3 weeks (Britt DeJohn, SVP of Operations) after analysing over 150,000 recorded dealer calls (Benjy Kest, Chief of Staff). The study reports ramp time cut from 3 to 6 months to about 1 month, about 25% of go to market staff practising voluntarily in the past month, and up to 20% fewer escalations in the first two months of a rollout (Britt DeJohn). Scorecards go out by Slack DM. These are Impel's figures; the practice tracks for your business have to come from your own calls.

Sources

• https://claude.com/customers/impel