What is AEO? Answer Engine Optimization for Dealerships

Your Inventory Is Already Being Judged by AI. Here's the Rubric.

Every day, car shoppers in your market open ChatGPT, Perplexity, or Google AI Overview and ask questions your lot has the perfect answer for. "What's a reliable used SUV for a family of five in Nashville?" "Which dealer near me has the best-priced Tacomas right now?" The AI answers them — and either names you, names a competitor, or names nobody local at all.

That selection process is Answer Engine Optimization (AEO). Not "ranking in featured snippets" — that's a 2019 definition. In 2026, AEO is the discipline of making your dealership and your inventory citable by large language models across ChatGPT, Perplexity, and Google's AI surfaces. The mechanics differ by platform, and the gap between dealers who understand this and dealers who don't is already measurable in leads.

How ChatGPT, Perplexity, and Google AI Overview Pick Sources Differently

These three platforms are not running the same algorithm. Treat them as three different buyers with different tastes.

ChatGPT (with web browsing / GPT-4o) pulls from Bing's index in real time and cross-references its training data. It weights factual density — how many specific, verifiable facts are present — and entity clarity — whether your dealership, the vehicle, and your location are clearly connected. It does not simply cite the top Bing result. It cites the result with the most extractable information. A VDP that says "2023 F-150 XLT, well-maintained, one owner, priced to sell" gives ChatGPT nothing to work with. A VDP that says "one-owner 2023 F-150 XLT, 34,200 miles, tow package, clean Carfax, priced $900 below DFW market average this week" gives it four citable facts.

Perplexity is a pure real-time search engine with citations displayed to the user. It weights the opening 60 words of a page heavily — if your VDP description front-loads a specific, direct answer to a buyer question, Perplexity is far more likely to surface it. It also rewards domain freshness: pages last updated more than 30 days ago get deprioritized relative to competitors with weekly refreshes.

Google AI Overview is the most schema-dependent of the three. It feeds from Google's Shopping Graph (which reads your Google Merchant Center feed), your Knowledge Panel (Google Business Profile), and on-page Vehicle schema. A VDP without `Vehicle` schema markup is largely invisible to AI Overview's inventory-recommendation layer — regardless of how good the text description is. AI Overview also weighs Google Business Profile review signals when recommending dealerships by name.

The practical upshot: optimizing for all three is not three separate jobs. It is one job done correctly — unique narrative content, schema markup, fresh updates, and off-site authority. But knowing which lever matters most on which platform helps you prioritize.

The 4-Signal AEO Test

Run this test on any VDP right now. Open the page. Ask four questions:

1. Entity Recognition — Can an AI reading this page identify three distinct entities: your dealership (name + location), this specific vehicle (VIN-level detail, not just make/model), and the buyer context (use case, geography, or lifestyle)? If all three aren't present, the AI can't form a recommendation. It needs to know *who you are*, *what you have*, and *why it matters here*.

2. Factual Density — Count the specific, citable facts in the description. Price relative to market. Mileage. Number of owners. Option packages by name. Service history. A description with zero citable facts cannot be cited. Aim for a minimum of five verifiable data points per VDP.

3. Freshness — When was the Inventory Comments field last updated in vAuto? AI crawlers check page modification dates. A VDP that hasn't changed in 45 days sends a low-priority signal to every AI crawler. Weekly refreshes are the standard for competitive AI visibility.

4. Structural Clarity — Can the AI parse your content without guessing? This means semantic HTML, `Vehicle` schema (vin, mileage, price, availability), and a description that is written in plain English rather than lot-speak. "Must see to appreciate!" is structurally opaque. "Third-row seating, 47,000 miles, no accidents, complete oil change history at Honda of Clear Lake" is structurally clear.

Score your VDP 0-4. If you score below 3, you are not getting cited.

The VDP Rewrite: Before and After

Here is what failing the 4-signal test looks like in vAuto's Inventory Comments field — and what passing it looks like.

Before (scores 0/4):

"2023 Honda CR-V EX-L. Leather seats, sunroof, Honda Sensing. Great condition — won't last long! Come see us today at Riverside Honda!"

Entity recognition: fails (no location, no buyer context). Factual density: zero citable facts. Freshness: updated once when it landed on the lot. Structural clarity: "great condition" and "won't last long" are non-parseable claims.

After (scores 4/4):

"One-owner 2023 Honda CR-V EX-L — 28,400 miles, clean Carfax, no accidents — priced $1,100 below greater Columbus market average as of this week. The EX-L trim adds heated leather front seats and a panoramic sunroof to Honda's base CR-V, making it the practical pick for Ohio families who split time between highway commuting and weekend drives. Honda Sensing (adaptive cruise, lane-keep assist, collision mitigation braking) is standard. Full service records on file at Riverside Honda in Dublin, OH. Available for same-day test drive."

Entity recognition: ✓ (dealership + location + vehicle + buyer context). Factual density: ✓ (7 specific facts). Freshness: maintained on weekly cadence. Structural clarity: ✓ (direct, parseable, no vague claims).

ChatGPT can cite that. Perplexity can source it. Google AI Overview can extract it.

Why Most Dealer VDPs Fail the Test

The root cause is vAuto's Inventory Comments field being treated as an afterthought rather than a primary marketing channel. Most dealers either leave it blank, paste in a feature list, or use a template that replicates across dozens of similar vehicles. AI systems detect duplicate content across domains instantly — if your CR-V EX-L description is identical to three other local dealers' descriptions, none of you get cited for it.

The evidence on VDP content uniqueness is unambiguous: unique, factually dense descriptions outperform template content across every AI surface.

What to Do This Week

Monday: Check your robots.txt for `GPTBot`, `ClaudeBot`, `PerplexityBot`, and `OAI-SearchBot`. If any are blocked or disallowed, fix it immediately. A blocked crawler cannot index you. A dealer with a blocked GPTBot is invisible to ChatGPT's web browsing regardless of content quality.

Tuesday: Run the 4-signal test on 10 random VDPs. Tally your score. If your average is below 2, you have a systemic problem — not a few bad descriptions.

Wednesday: Add `Vehicle` schema to your VDP template if your website provider hasn't done it. Fields: `vin`, `mileageFromOdometer`, `vehicleModelDate`, `offers` (price, availability), `description`. This is the schema Google AI Overview reads from.

Thursday-Friday: Rewrite your five highest-priced vehicles using the after-example format above. Track your AI search citations weekly using ChatGPT and Perplexity for queries like "best used CR-V near Columbus."

If you're running 150+ vehicles, InventoryPilot AI automates this process inside vAuto — every VIN gets a 4-signal-passing description refreshed weekly, delivered directly to your Inventory Comments field. See the before/after comparison or Get a Free Audit to see your specific inventory.

For the broader picture on how these signals interact across all generative surfaces, see our GEO guide for dealerships and the practical AI search audit a GM can run in 30 minutes.