By Michael Quigley, President & Chief Strategy Officer, Co-Founder at Impel

There’s a fascinating piece making the rounds from a16z about the company Decagon. The argument is simple and seismic: AI collapses the cost of high-quality attention, and when that happens, every business becomes a concierge business.

I’d push back on one word. Not every business becomes a concierge business. Some businesses were concierge businesses — and then the scale got in the way.

The dealership is one of them.

What the dealership was built to be

Think about what a great dealership looked like forty years ago. The dealer principal who knew your name. Who remembered that you traded in a truck last time, that you had three kids, that you liked leather seats and hated sunroofs. Called you personally when the right car came in. Made sure the service advisor knew your history before you pulled into the lane.

That wasn’t customer service. That was a concierge. And it worked, because the relationship was the product as much as the vehicle was.

Then the pressures of scale driven largely by the internet began to arrive. The constant flow of internet leads. Multi-rooftop groups. BDC departments managing hundreds of contacts a day. Digital retailing. The dealership tried to behave like a scaled business — high volume, systematized, efficient — without the operational systems in place  that make scale a high NPS experience for consumers. And it ended up with the worst of both worlds.

Too fragmented to deliver the concierge experience. Too high-touch to benefit from pure operational efficiency.

The result? A disjointed and often impersonal experience. Missed calls. Sales responses measured in hours (sometimes days), not seconds. Generic follow-up. Service reminders that don’t know what vehicle you drive.  The relationship that was once the dealership’s greatest competitive advantage quietly hollowed out.

And as I argued in The Down Escalator Problem, the cost of that erosion isn’t static. It compounds — operationally, competitively, and culturally — every month you wait to address it.

Why generic AI doesn’t fix this

The conventional narrative about AI in customer service goes like this: deploy a chatbot, deflect tickets, reduce headcount, save money. And for large-scale industries that have millions of customers — airlines, telecom, banks — that’s a reasonable starting point.

But that’s not the dealership’s problem to solve.

The dealership’s problem isn’t volume. It’s context and human touch. A generic AI can tell a customer your service hours. It cannot tell them that their 2022 Accord is approaching its factory-recommended interval, that they declined a tire rotation last visit, that their lease ends in four months, and that based on their browsing behavior, they should consider a Honda Pilot for their next purchase. It can’t connect those signals and respond like someone who actually knows them.

That’s the difference between a scale business deploying AI generically and a dealership deploying vertical (domain-specific) AI. Generic AI makes you cheaper. Vertical AI makes you the concierge again.

In Centaurs, Chess, and the State of Vertical AI in 2026, I explored why vertical AI — purpose-built for a specific industry — consistently outperforms general models in high-stakes, context-dependent environments. Automotive is exactly that environment.

What the concierge actually looks like in automotive

The a16z piece describes the concierge as ambient, proactive, personalized, continuous. Not episodic — not only present when something goes wrong — but woven into the ongoing texture of the customer relationship.

That description maps perfectly onto what automotive retail has always needed and consistently failed to deliver at scale.

Proactive: reaching a customer before their service is overdue, not after they’ve already gone to the competitor down the street. Personalized: knowing the vehicle, the history, the driving behavior, the preferred communication channel. Continuous: a relationship that doesn’t reset every time the customer calls in or submits a lead. Ambient: present at every touchpoint, not just the ones where something breaks.

This is what vertical AI built for automotive actually does. It doesn’t replace the relationship. It restores it and enhances it at a scale no human team could sustain alone.

The stakes are higher in automotive than almost anywhere else

The a16z piece goes on to make the point that when the concierge knows you and is always on, the line between support and commerce (and, I’d posit, between departments) dissolves. But a great concierge isn’t just solving problems, it’s driving revenue.

In automotive, that dynamic is everything. The customer who feels known and cared for services with you. They buy again and refer. They don’t defect to the independent shop down the road because nobody followed up. The lifetime value of a loyal automotive customer is measured in tens of thousands of dollars and sustainable margins across years of service revenue and repeat purchases.

The dealership that delivers the concierge experience doesn’t just win the transaction. It wins the relationship. And in automotive retail, the relationship is the business.

Forty years ago, the best dealers knew that instinctively. The ones who are defining the next twenty years are figuring out how to deliver it again — at scale, powered by AI that actually understands the industry it’s serving. AI that enables humans to focus on what they’re built for and best at: fostering genuine relationships that engender long-term loyalty.

The concierge never left automotive. It just needs the right technology to come back.

Want more perspectives on vertical AI, automotive, and strategy? Subscribe to Drive the Future.