AI Should Surface the Issue. Leadership Should Own the Decision.

Most dealerships now run at least one AI tool: automated bidding, chat, follow-up, merchandising, or reporting. The platform turned it on by default. The vendor left it that way. Nobody at the store was asked.

Whether to use AI is not a live question anymore. What is still open is who reads its output, who checks that output against what landed in the CRM, and who has the standing to shut it off on a Friday afternoon.

AI is the smoke detector. It tells you something changed. A person who has run a dealership marketing department decides what to do about it.

Automation Drifts

Automated campaigns learn from whatever you feed them. Feed them stale inventory, a bad conversion signal, or an unfiltered radius and they optimize hard toward the wrong buyer, quietly and at speed.

Meta's automated campaigns ran 9% ahead of hand-built ones at the halfway point of controlled testing — then finished 12% behind them. 58% of advertisers got a better return from manual. Left alone, automation drifts.

Source: Haus, The Meta Report — 640 controlled geo-lift experiments over 18 months. Sample is direct-to-consumer advertisers, not dealerships. The mechanism carries; the exact figures will not. We measure yours against your CRM.

It Grades Its Own Homework

The platform reports the conversions it wants credit for. We check those against what actually landed in the CRM and what actually sold, because those are different numbers and only one of them pays you.

That same automation over-reported its own contribution by about twelve points against measured incremental lift.

Guardrails Are Not Automatic

Geographic exclusions, negative keywords, placement blocks, frequency caps, suppression lists — so you stop paying to advertise at your own customers and staff. No AI sets these for you. Somebody has to.

Your AI Vendor Has No Counterpart

The tools get installed with a demo and a login. After that, nobody on the dealer side is briefing them, testing their output, or holding them to a number. They just run.

Automated Account Health Checks

Spend pacing, delivery, and budget burn checked every morning against the plan. A campaign that swallowed three days of budget over a Saturday shows up Sunday, not on the invoice at the end of the month.

Feed & Tracking Monitoring

Inventory feed status, conversion tag health, call tracking uptime, and a real submission through the site's own forms. When tracking breaks, bidding keeps spending — it just stops learning from anything real, and the report still looks fine.

Spend-Pacing Alerts

Pacing against the month's plan by campaign and by funding source, with an alert when a campaign runs hot or goes dark. Co-op-funded spend gets watched on its own line, because that money has a deadline attached to it.

Reporting & Proof Workflows

Screenshots, invoices, tear sheets, and platform exports collected while the month is running, so a co-op packet gets assembled out of files that already exist instead of reconstructed the week it is due.

Human Review Before Material Changes

Automation proposes; a person signs off. Budget reallocation, campaign restructuring, creative swaps, and bid strategy changes wait for somebody who knows what is sitting on your floor this month and what the OEM is pushing next.

Data & Governance

Who holds account access, who approves changes, what gets logged, and the written list of what the automation is never allowed to do on its own. The dealership owns the accounts and the data. The dealership approves the changes. We run the system and answer for it.

The strongest use of AI in a dealership marketing operation has nothing to do with a chatbot standing in for your sales team. It looks like this:

• Catching the drift early. Spend walking away from plan, a feed that stopped refreshing, a tag that stopped firing, a lead source that quietly renamed itself in the CRM.
• Hours instead of weeks. The gap between a tracking break and somebody noticing it is where most of the wasted money in a dealership account lives.
• Paperwork that assembles itself. Invoices, proofs, and screenshots captured as they happen, so a claim does not run out its deadline while somebody rebuilds a PDF from memory.
• Patterns nobody has time to sit and find. Which search terms are pulling service traffic into a sales campaign, which model family gets views but no leads, which ZIP codes keep buying without ever being targeted.

None of that makes the call. It puts a better set of facts in front of the person who does, sooner.

• Run it end-to-end. We set up the monitoring, write the guardrails, read the output every morning, and bring you the short list that needs a decision. Nothing material moves without your approval.
• Strengthen your current setup. We inventory the automation already running in your accounts — bid strategies, CRM follow-up, chat, merchandising — flag what has no supervision, stand up the alerting, and teach your marketing person what to watch and when to pull the handle.
• Add the monitoring layer. Your team runs the campaigns. We watch the signals and call you when something needs a human.

• Daily monitoring: spend drift, feed failures, tracking loss, disapprovals, and source changes caught before they compound
• Human approval gates: no budget reallocation, campaign restructuring, or bid strategy change goes live without review
• A monthly operating review: what the automation flagged, what the guardrails caught, and what its reported numbers look like next to your CRM and sold units
• Clear data boundaries: the dealership owns the data, approves the changes, and decides what AI is authorized to do