Our own Michael Cirillo sat in the room this week at Reynolds and Reynolds' Amplify 2026, where Automotive Ventures' Steve Greenfield opened a keynote panel with a fifty-year chart that made the whole room go quiet: the average American dealer has not lost money in any year over the last five decades.
Greenfield moderated a conversation with three dealers who've built very different businesses on top of that same stubborn fact: Jeff Swickard of Swickard Auto Group, Howard Tenenbaum of Keyes Automotive Group, and John Keoughan of Temecula Valley Buick GMC.
The premise of the session was simple and a little provocative. If dealers have held steady at 1.5% to 2.5% net profit before tax for fifty years, with 2025 coming in at an elevated 3.3%, is AI about to create a permanent step up in that number, or is the industry just going to drift back toward the mean like it always has?
We're like cockroaches. We're resilient. We keep changing, and you have to adapt.
The Real Money Was Never in the Car
Greenfield's data made one thing plain before the panel even sat down: new vehicles generate the most revenue in a typical store, but F&I and service are where the actual profit lives.
Personnel costs eat up 47% of a dealership's overall cost structure, which is exactly why every panelist kept circling back to the same idea from different angles: AI's real opportunity isn't flashier tech, it's freeing up the people already on payroll to do the parts of the job only a person can do.
What Three Very Different Dealers Are Actually Doing About It
Jeff Swickard lets AI buy his cars. His stores price over 80% of used inventory with AI, which pushed volume up 20%, and he's rolling out an auto-replenishment system that detects low inventory, builds a shopping list, bids at auction, and arranges transport without a human touching any of it.
Howard Tenenbaum isn't chasing the next shiny tool. His focus is doubling down on things that already work, like service-drive videos, and he's blunt that no AI system has replaced the person needed to actually create a sale. "Cars are sold, not bought," he said, and he's not backing off that line for anyone.
John Keoughan swapped bulk texting for something with a pulse. Moving from generic mass SMS to a personalized outreach system took his open rates from 2.5% to 4.5%, and a new AI phone system now measures sentiment on 100% of calls instead of the old system, where the only feedback a store ever got was from customers angry enough to complain.
All three agreed on where AI still hits a wall. Nobody on that stage believes the showroom floor gets fully automated anytime soon. The tools get sharper every year. The person closing the deal is still the person closing the deal.
A Ten-Year-Old Prediction Just Became Reality
Swickard's auto-replenishment system landed differently once Paul made the connection out loud on the podcast this week. He recalled a conversation with Dale Pollak, founder of vAuto, roughly a decade ago, where Pollock described a future where the inventory a store needs simply shows up at the door before anyone on staff even realized they needed it. The data would source it, the data would put it on a truck, and it would arrive already matched to that store's market and goals.
That's not a hypothetical anymore. It's Swickard's actual workflow: a system that detects low inventory, builds the shopping list, bids at auction, and arranges transport with no human touching any step of it. The prediction and the practice have finally caught up to each other.
So Is the Profit Ceiling Actually Moving, or Just the Tools Underneath It?
Nobody on stage claimed to know for certain. But the direction of the conversation was consistent: this isn't about replacing headcount, it's about making the headcount you already have generate more per person than it ever has.
One detail from the panel makes that explicit. AI is not being used by any of these three dealers to cut staff. As Kyle put it, discussing the session, the more likely outcome is a filtering effect: AI makes the great operators greater, and the dealers who can't close the same efficiency and market-penetration gap start to fall behind. Not fewer people doing the same work. The same people doing work that actually compounds.
Swickard's 20% volume lift from AI-priced inventory and Keoughan's near-doubled open rates aren't hypothetical; they've already happened. Whether that adds up to a permanently higher industry-wide baseline or just a temporary edge for the dealers moving fastest is the real question worth sitting with.
What This Means for Your Store
Look at where your own personnel hours are going toward repetitive tasks that a tool could absorb, then ask honestly whether that freed-up time is actually being redirected toward higher-value work, or just disappearing. The dealers on this stage weren't using AI to do less. They were using it to make sure their best people spent their time doing the one thing AI still can't: making a customer actually want to buy.
And if our filtering theory holds, sitting out that shift isn't a neutral choice. It's a competitive one.
