Academy / AI Search & GEO

Rank vs citation, then the engines · 28 min

Engine matrix: Overviews, ChatGPT, Perplexity

In this lesson

  • Compare source types and when each engine is likely to cite a local shop
  • Predict which engine is the fairest first audit for inventory vs emergency-heat prompts

Three engines, three source habits

Do not treat “AI search” as one box. Google AI Overviews live inside Search. They are grounded in how Google already understands the web and often surface links next to a synthesized answer. A local business is more likely to appear when the query is local and commercial (city + unit, city + emergency service) and when a crawlable page plus a consistent business entity already exist. ChatGPT blends training priors with optional browsing or search. Without a live look, it leans on famous brands and generic advice; even with browsing it may mention a name and skip a footnote. Perplexity is built to show numbered sources. It is usually the fairest place to see whether a specific local URL can be retrieved at all. None of the three is a keyword-volume report. None of them is an Ads auction.

When they cite a local business

Local citation gets more likely when four things are true at once: (1) the prompt names a place a human would actually type (Frederick, Hagerstown, Williamsport — not “best cars USA”); (2) the business has a unique, checkable fact on a public URL (48-hour test drive; $129 online tune-up; same-week emergency heat inside a named area; written estimate before replacement); (3) name, address, and phone match the Google Business Profile and the about page; (4) the page is not a doorway of spun city+keyword paragraphs. Inventory citations want VIN- or stock-level URLs a model can quote. Emergency-heat citations want a service-area sentence a model can quote without inventing Baltimore coverage Priya refuses. National chains win when your facts are only on a homepage poem.

Worked matrix — same prompt, three habits

Prompt A (inventory): “Who sells used Honda CR-Vs in Frederick MD?” AI Overviews — if the feature triggers, expect links from Search’s local/commercial results; a thin homepage may lose to a marketplace or a chain with cleaner vehicle pages. ChatGPT — often names a chain or gives generic “check local dealers” advice; a cite is more likely if browsing is on and /inventory or a CR-V VDP exists. Perplexity — likeliest to footnote a specific Frederick URL if one is retrieveable; if Northfield has no crawlable CR-V page, it will cite a competitor or a listing site. Prompt B (emergency heat): “Emergency heat repair Hagerstown tonight.” Overviews — local pack + service pages with hours and area. ChatGPT — franchise priors, or a hallucinated 24/7 claim. Perplexity — Dual Highway / services URL if the page states emergency-heat and the 25-mile area in plain language. Write the habit, not a fake quote.

Never paste a fabricated “ChatGPT said…” or “AI Overview cited us…” into an audit. If you did not run the prompt, label the row inferred. If you ran it, quote only what you saw, with date and engine mode (browsing on/off).

Google · Official resource

Google Search Central — AI features

Official page on AI features in Search. Use it to check a product name. Do not paste the doc as your assignment. This course does not recap that page.

Exercise

Predict three engine rows

For each prompt, pick the fairest first engine to audit and write the source type you expect (Search links, training prior, numbered web sources). Prompts: (1) used CR-V Frederick with a live VDP; (2) Cedar & Stone fall tune-up price; (3) “best HVAC company in America.” Label each prediction inferred.

Deliverable: 3-row table: prompt, first engine, expected source type, inferred.

Create a free account and verify your email to save progress, quizzes, and portfolio work. Create a free Academy account · Sign in

Ask the tutor

The tutor will guide you. It will not finish graded work for you.