It was a heavy couple of weeks for anyone trying to keep up with this industry. A keynote at Amplify about whether AI finally breaks the dealer profit ceiling, a US jobs report that went the wrong way, a Canadian one that went the right way, a new tracker on how many dealers are actually using AI, and the usual turnover numbers that everyone quotes and nobody does much about.

I read all of them, and I kept having the same reaction — separately they're headlines, but stacked together they point at something fairly specific. And it isn't really about AI.

Are dealers using AI to cut headcount?

No, and that's the part worth sitting with.

Steve Greenfield of Automotive Ventures opened a keynote panel at Reynolds and Reynolds' Amplify 2026 with a chart showing the average American dealer hasn't posted a losing year in fifty. His question to the panel was whether AI permanently lifts dealers above the historic 1.5% to 2.5% net profit range, or whether the industry drifts back to the mean the way it always has.

The three dealers landed in genuinely different places. Jeff Swickard's stores price over 80% of used inventory with AI and have lifted volume 20%, with an auto-replenishment system now buying auction vehicles with no human involved. Howard Henneman of Keys Motors held the line that cars are sold, not bought, and nothing has replaced the person closing the deal in his stores. John Keoughan swapped bulk texting for AI-personalized outreach and nearly doubled his open rates.

What struck me wasn't the disagreement, it was what none of them was doing. Not one was using AI to reduce headcount. Every one was using it to get more out of people already on the payroll — a much more demanding strategy than it sounds, because it only pays off if those people are still there in eighteen months.

Why aren't dealers seeing revenue from AI yet?

Cox Automotive's AI in Auto Retail Tracker, published August 11, 2026, found that 82% of dealers are using AI, most commonly for customer follow-up, content generation, and task automation. 69% expect AI to grow sales and revenue. Only 22% report seeing that growth today. The benefits dealers do report are improved customer experience (28%) and improved employee productivity (26%).

I don't read that as AI disappointing anyone. The return is real and it's getting stuck one step short of the P&L, because the step that converts a tool into revenue is a person. A pricing engine doesn't make money on its own. A used car manager who has spent four months learning when to trust it and when to overrule it makes money. Those are two different assets, and only one of them appears on the invoice.

What did the July 2026 jobs reports mean for dealership hiring?

The US numbers were rough. Payrolls fell by 23,000 against a forecast 83,000 gain, with May and June revised down by a combined 103,000. Unemployment fell to 4.1%, but for an unwelcome reason: roughly 264,000 people left the labour force, pushing participation to 61.4% — the lowest reading outside the pandemic since 1976. Retail alone shed 19,000 jobs.

That distinction matters more than it usually gets credit for. When unemployment rises because of layoffs, candidates land in the market and you can hire them. When it falls because people have exited, those candidates are simply gone. One is a hiring opportunity and the other is a hiring problem wearing a good headline.

Canada looked much better on the surface. Statistics Canada reported 75,000 jobs added in July against expectations of 15,000, with unemployment falling to 6.4%, a two-year low, and wholesale and retail trade leading gains at 21,000. But that same sector remains down roughly 50,000 jobs year over year, against slower population growth and rising retirements. A good month sitting on top of an unchanged decade.

If your 2027 staffing plan quietly assumes a looser labour market will send you people, I'd stress-test that now.

What is dealership employee turnover in 2026?

Total dealership turnover sits at 42%, with sales roles near 60%, according to the NADA Dealership Workforce Study. Entry-level technician turnover runs about 67% with median tenure of 0.9 years, and Gen Z turnover rose six points year over year, double the movement of any other generation. On the technician side, the industry needs to replace roughly 76,000 people annually while US training programs graduate about 39,000.

Our own data says something sharper. In HR4's State of the Automotive Workforce, built from 800+ dealerships across North America, 27.6% of all exits happen before day 90 and 58% happen inside the first year. Sales turnover in our data runs materially higher than the NADA figure, which I'd put down to methodology rather than disagreement — the two studies count differently, and either way the direction is the same.

How should a dealer group calculate ROI on AI?

The way I've started saying this internally, deliberately bluntly: the return on an AI tool is roughly its productivity lift multiplied by the average tenure of the people running it. Double the lift and you double the return. Halve the tenure and you hand it straight back. Most of us are spending real money on the first half of that equation and treating the second half like weather.

When someone in an AI-enabled role leaves, what walks out isn't the seat, because seats are refillable. It's the judgment about when the model is wrong, which is learned and almost never written down. The workaround they built that nobody else knows about. The customers sitting mid-conversation inside the automation. The two newer people on the floor who were quietly learning from them.

We see this at rooftop level constantly. A group rolls something out, two stores make it look easy, and everyone credits the market or that GM's enthusiasm. Usually it's neither. The stores where it works have people who've been there long enough to figure it out, and the ones where it doesn't have been starting over every nine months.

Four questions to ask before your next AI rollout

None of these are about the tool, which is sort of the point.

1. How long until a new hire in this role is fully productive? If that's 90 days and the role turns over at 50%, a real share of the department runs at partial output permanently. Shortening the ramp is often a bigger lever than the software.
2. Does the workflow live in a person's head or in the system? That's the difference between a store that survives a resignation and one that resets.
3. Can you see productivity by role and by rooftop, or only in total? Most groups can't, because hiring sits in one system, training completion in another, reviews in a spreadsheet, and time and attendance somewhere else.
4. Do you know why people left, or only that they did? Exit data is the cheapest research available and the most consistently ignored, and people rarely leave for the reason written on the form.

The boring part is the part that lasts

The fifty-year streak isn't proof dealers pick the right technology every time. It's proof the business is resilient and operators keep adapting. AI gets absorbed the way the internet was — something everyone eventually has, where the difference comes down to how well it's executed at store level.

Which puts us back where dealership performance has always actually been decided. Hiring well, getting people good quickly, giving managers something better than instinct, and giving people a reason to stay past the point where they've become genuinely valuable. That's what we've been building toward at HR4, the HR and health and safety platform built for auto dealer groups across Canada and the US, and honestly it's the least exciting sentence in this post. It's also the one I'd bet on.

AI raises the ceiling on what a good team can do. It does nothing for a team that's turning over.

Frequently asked questions

Are dealerships using AI to replace employees?

‍Not at present. Dealers presenting at Amplify 2026 described using AI to raise output per employee rather than reduce headcount, applying it to used-vehicle pricing, inventory replenishment, and customer outreach while keeping sales staff in place.

What percentage of car dealerships use AI?

‍Cox Automotive's AI in Auto Retail Tracker, published August 11, 2026, found 82% of dealers using AI in some form, most often for customer follow-up, content generation, and task automation.

What is the average employee turnover rate at a car dealership?

‍Roughly 42% across all dealership roles, rising to about 60% in sales and about 67% for entry-level technicians, whose median tenure is under one year.

How much dealership turnover happens in the first 90 days?

‍27.6% of all dealership exits occur before day 90, and 58% occur within the first year, according to HR4's State of the Automotive Workforce study of 800+ dealerships across North America. Early attrition concentrates the loss in the period before a new employee reaches full productivity.

Why is dealership technician hiring so difficult?

The industry needs to replace approximately 76,000 technicians each year while US training programs graduate around 39,000, producing a structural annual shortfall of tens of thousands of people.

How does employee turnover affect return on AI investment?

‍AI returns accrue through trained employees, so they scale with tenure. A tool delivering a productivity gain returns roughly that gain multiplied by how long the operator stays, meaning high turnover erodes AI ROI even when the technology performs as promised.