Platform / Coaching and training

On the roadmap

Practice on an AI customer first.

New hires and veterans will be able to practice a real dealership conversation against an AI customer that pushes back, then hear direct feedback on the call about what they could have done better. It is planned training, not a customer call.

Book a walkthroughResources for coaching and training

PinpointAI customer: the buyer who must talk to a spouse
AI customerI like the Highlander, but I am not committing to anything until I talk to my husband.
RepOf course. Would it help if you both came in Saturday morning? I can hold 10:30 and have it ready for a drive.
AI customerMaybe. Let me ask him and get back to you.
Feedback on this call
You offered a specific day and time and made the visit about both of them.
You let "let me ask him" end the call. Ask when they talk and hold the time until then.
You did not ask what he would want to know. That is the reason for the visit.
Planned: the AI plays the customer, the rep takes the call, and direct feedback follows. Synthetic example.

What ai roleplay training is

AI roleplay training is practice against a voice agent that plays a customer: the price shopper, the payment buyer, the one who needs to talk to a spouse. Pinpoint has this on its roadmap. The rep will run the conversation out loud and get direct feedback before hanging up, and training will stay separate from real customer conversations, so practice never enters your customer records, your metrics or your evidence.

On the roadmap: roleplay training is not available today, and coaching from real conversations is what is available now.

How it works

  1. 01

    Pick the customer type

    A manager will choose the scenario: a price shopper who opens with the lowest number, a marketplace lead who wants photos, a buyer who says they need to talk to their spouse. The scenarios will come from the kinds of conversations dealerships actually take.

  2. 02

    The rep takes the call

    The rep will talk to the AI customer out loud, on the phone, the way they would with a real one. The agent will push back, go quiet, ask the awkward question and end the call early if the rep gives it a reason to.

  3. 03

    Feedback on the call

    When the roleplay ends, the rep will hear what worked and what to change while it is still fresh: the question that was missed, the number given too early, the time that was never asked for. The same practice call will be there for a manager to replay.

  4. 04

    Practice stays separate

    Practice calls will live in their own place. They will not appear in your customer records, your call metrics or your evidence, and an AI customer will never be pointed at a real buyer. Training and the real floor will stay two different things.

What a manager sees

Evidence, approvals and permissions

This training is planned to sit apart from your customer conversations. Practice calls will not be mixed into customer records, metrics or evidence, and an AI customer will only ever talk to your own people. Feedback will point to the moment in the practice call it came from. Who can assign practice and who can replay it will follow the permissions your dealership configures at onboarding.

The AI layer never acts outside the approvals and permissions a dealership configures at onboarding. Inbound voice agents answer only missed calls; outbound calls are drafted by the AI and sent only when a manager approves them.

Questions about ai roleplay training

Does the AI customer use our real customers' conversations?

No. Training is planned to stay separate from observed customer conversations. The scenarios will be built from the kinds of situations dealerships face, not from a particular customer's call, and your conversations are never used to train AI models. Practice calls will also stay out of your customer records, your metrics and your evidence.

Can my team practice on it today?

No. Roleplay training is on the roadmap. Today a rep is coached from their own real conversations, with the moments marked and the recording ready to play. When practice arrives, it is built to sit beside that coaching rather than replace it, so the same habits get worked on both ways.

Could a practice agent ever call a real customer?

No. An AI customer is built to talk to your own people only. It is a training partner, not a line into your customer base, and it will have no way to reach a buyer. Voice agents that speak with real customers are a separate capability with its own permissions and its own approvals.

What will a practice call not prove?

How someone handles a roleplay is not proof of how they handle a buyer with a trade and a payoff. Practice results will stay out of your metrics for that reason. Treat a practice call as a rehearsal: useful before the real one, and never a substitute for what the real calls show.

How is this different from the voice agents that talk to customers?

One plays a customer for your rep; the other speaks with your customers. Roleplay training is on the roadmap. Customer facing voice agents are being built now: inbound, they are being built to answer only the calls your team missed, and outbound calls will be drafted for a manager to approve. Different jobs, different permissions, different stages.

Go deeper

See ai roleplay training on your own dealership's conversations.

Book a walkthrough