Six Months In: What Dealer Voice AI Rollouts Actually Look Like

The first 30 days of a dealer voice AI rollout, in honest detail

The first 30 days of a dealer voice AI deployment are the part the vendor demo does not prepare you for. The demo is the polished version. The first month is the rough version. Both are real.

Most of the month is calibration. The team is learning what the AI does well, what it does badly, and what kinds of calls should still be routed to a human. The prompts need adjustment. The routing rules need adjustment. The escalation paths need adjustment. None of this is hard, but it is all-consuming, and it does not happen on its own.

There are usually three surprises in the first month, and the dealerships that get through the period well are the ones who see the surprises coming.

Surprise one: the AI is great at the calls you thought it would be bad at.

Phone tag, reschedules, the third confirmation call. The kinds of conversations that take a human BDC agent eight minutes and produce almost nothing. The AI handles them in under a minute and never gets frustrated. The team expected this to be the hard part. It turns out to be the easy part.

Surprise two: the AI is bad at the calls you thought it would be great at.

The first conversation with a hot lead. The lead who is ready to buy and has one specific question. The lead who is on the fence and needs a moment of human warmth. The AI handles the easy confirmation calls but stumbles on the lead who is one sentence away from booking the appointment. The dealership that wins the first 30 days is the one that recognizes this and routes those calls to a human, fast, before the AI has a chance to lose the lead.

Surprise three: the customer experience is uneven, and it shows up in the data, not in the complaints.

Customers do not call the dealership to complain about the AI. They just do not book the appointment. The show rate on the AI-set appointments in the first 30 days is usually lower than the show rate on human-set appointments, and nobody knows why, because the customers never said anything. The data tells the story, if anyone is looking.

The dealerships that come out of the first month in good shape are the ones who are listening to the AI's calls, every week, and writing down what they hear. The dealerships that come out of the first month in bad shape are the ones who put the AI in production, glanced at the dashboard, and moved on.

What months three through six actually look like

Months three through six are the part nobody talks about, and they are the part that determines whether the deployment is going to last.

By month three, the team has stopped treating the AI like a side experiment. It is part of the operation. The dashboards are not the focus anymore. The team is looking at the show rate on the AI-set appointments, the customer experience scores on the AI-handled calls, and the breakdown of which kinds of calls are being routed to the AI versus a human.

This is also the window where the prompt changes start to compound. A small change to the AI's greeting in month two, plus a routing rule update in month three, plus a new escalation path in month four, are not big moves on their own. Together they produce a different AI by month six than the one that went into production on day one. The dealerships that win this window are the ones that are tracking those changes and being deliberate about them.

A few patterns show up in the dealerships that get the most out of months three through six.

The team has built a weekly review cadence for the AI's calls. Twenty calls a week, picked at random, listened to by a senior manager. The same cadence the human BDC team gets. The AI's quality is held to the same standard as the human team's quality, and the prompt changes that come out of the review are written down and tracked.

The escalation rules have matured. The team is no longer escalating every complex call to a human. They have identified the specific moments in a conversation where the AI should hand off, and the rest of the conversation is handled by the AI. This is the difference between month one (escalate everything) and month six (escalate only the moments that matter).

The handoff to sales is integrated. The AI sets the appointment, the AI sends the confirmation, the AI handles the day-before call, and the handoff to the salesperson is structured so the salesperson knows what the lead talked to the AI about. The salesperson is not starting from scratch. The dealership that has this right in month six is the dealership that has stopped losing appointments between the AI and the sales floor.

The numbers move, and the team can explain why. The show rate on AI-set appointments is approaching the show rate on human-set appointments. The customer experience scores on AI-handled calls are stable or improving. The lead-to-appointment rate is the number it was on day one, but the appointment-to-show rate is meaningfully better. The team can point to the specific prompt changes and routing changes that produced the lift. That is what month six looks like when it goes well.

What we got wrong about the rollout, in the early days

In the early deployments, we thought the work was mostly in the first 30 days. Get the prompts right, get the routing right, get the escalation right, and the rest would take care of itself. The numbers would move, the team would get comfortable, and the deployment would stabilize.

That framing was wrong. The first 30 days are the setup. Months three through six are the work. The dealerships that get the most out of the deployment are the ones that are still actively tuning the AI in month six, and the dealerships that get the least out of it are the ones that stopped tuning in month one.

It took us longer than it should have to internalize this. We had dashboards that emphasized the first 30 days, success metrics that were set in the first month, and a support model that de-prioritized the post-month-one work. We changed all three. The deployments that started after the change look different from the ones that started before, and the lift is real.

The honest version of what to expect

A dealer voice AI deployment that goes well produces a different operation in month six than the one that existed in month one. The BDC team's time is spent on the calls that require a human. The AI is handling the calls that do not. The customers who reach the dealership are getting to the right person, human or AI, faster than they were before. The numbers move, and the team can explain the moves.

A deployment that does not go well looks the same in month six as it did in month one. The AI is still making the same mistakes. The team has stopped reviewing the calls. The show rate on AI-set appointments is stuck at the level it was at in week three. The dealership is paying for a tool that is doing the work of a tool that was never properly set up.

The difference between the two outcomes is not the AI. The difference is the operational discipline around the AI, applied for the full six months, by a team that is willing to keep listening.

Where to go from here

If you are evaluating voice AI for your dealership, the most useful thing you can do is ask the vendor what month six looks like for their deployments. Not month one. Month six. The vendor that has good answers to that question is the one that has been through enough rollouts to know what actually happens. The vendor that only has month-one stories is selling you the demo.

If you are already running voice AI in production and you are not sure whether the deployment is in the good month-six or the bad month-six, the test is simple. Pick twenty AI-handled calls from the last week. Listen to them. Write down what you hear. If you cannot honestly describe what the AI is doing well and what it is doing badly, the deployment is in trouble. The first step back is to start listening.

We work with dealerships that want help getting to the good month-six. The work is real, the lift is real, and the dealerships that commit to the six-month discipline are the ones that get the most out of it. If you want to talk about what that looks like for your store, book a call and we will walk through it.