The 10-point checklist to get your company recommended by AI

This is the checklist we walk new customers through. Nothing on it is exotic — that's the point. AI engines assemble recommendations from public evidence, so the work is publishing the right evidence and keeping it consistent. Ordered roughly by impact for a typical trades company.

Rating, volume, and recency all feed the engines' trust model — and recency is the one most companies neglect. Build a habit: every completed job gets a review ask within 48 hours. A steady drip of fresh Google reviews does more for AI visibility than almost anything else on this list.

Same business name, phone, address, and service area on your website, Google Business Profile, Yelp, Angi, and your state license board. Engines cross-reference; mismatches read as risk, and risky candidates get dropped from answers silently.

You don't need fifty listings. You need complete, photo-rich profiles on the handful of platforms the engines cite for your trade — in our citation data that's consistently Google, Yelp, Angi, Houzz, and BBB. An unclaimed, half-empty Angi profile is worse than none: it's the version of you the engine might quote.

License number, insurance, years in business — on your homepage footer and about page, in plain text. Engines prefer recommending companies whose legitimacy they can verify against a source, and your state's contractor registry is a source they check.

"How much does X cost in [city]" is the highest-volume buying question in every trade we track, and the engines answer it by citing whoever published real numbers. Ranges are fine; honesty about what moves the price is better. This single page type is the most-cited contractor-owned content in our data.

Not doorway-page spam — a genuine page with local proof: jobs completed there, neighborhoods you know, local permit quirks. Engines resolve "in Coral Gables" literally, and a company with a real Coral Gables page beats a company whose site never says the word.

An FAQ written from real customer calls — "can you repair just the flashing?", "do you handle the permit?" — maps one-to-one onto the prompts people type into assistants. You already know the questions; you answer them on the phone every day. Write them down.

Project galleries with dates, cities, materials, and problem-solved captions. Specifics are citable; "quality you can trust" is not. One detailed case study outperforms a hundred stock photos.

Measure your current visibility across your market's buying questions — repeated runs per engine, not one search. Without a baseline you can't tell whether any of the work above moved the needle, and this channel drifts too much to eyeball.

Visibility work compounds when it's a loop: measure, fix the biggest gap, measure again. Weekly is the right cadence — fast enough to catch a competitor's move or a review-score dip, slow enough that real trends separate from noise.