How to get ChatGPT to recommend your business

There are two separate problems hiding in this question. The first is whether an AI engine can read your website at all, which is technical and fixable in days. The second is whether it will name you when someone asks who they should use, which is a question about evidence and reputation. Most advice conflates them.

Start by reading the answers, not your rankings

Write down the ten questions a buyer asks before choosing someone in your category. Ask each engine, in a fresh session, and record three things: which companies are named, in what order, and which sources are cited underneath. That is your baseline, and it usually contains a surprise. Very often the cited sources are not company websites at all but review platforms, comparison articles, and community threads.

Why third-party sources dominate recommendations

A recommendation is a comparative judgement, and a company describing itself is weak evidence for it. Engines lean on sources that compare several options, because those are the ones that answer the question being asked. This is why a business can be perfectly readable, technically flawless, and still never named.

• Be present where your category is already compared: directories, review sites, and the listicles that rank for “best X” in your market.
• Community threads carry real weight in recommendation answers, and they cannot be bought without it backfiring.
• Case studies with named outcomes give an engine something concrete to attribute to you.

Give the engine something specific to say

Adjectives are unquotable. “Leading provider of innovative solutions” gives a model nothing it can attribute. A specific claim does: what you do, for whom, where, at what scale, with what result, and what you do not do. Stated limits are unusually persuasive to an answer engine because they read as calibration rather than marketing.

Then fix the plumbing, because none of it counts otherwise

All of the above assumes the engine can fetch your pages. The AI crawlers do not run JavaScript, so client-rendered content is invisible to them. Serve your key pages as real HTML, publish structured data describing your organisation, allow the AI crawlers in robots.txt, and make sure Bing has you indexed. Then the evidence work has somewhere to land.

One caution: engines change retrieval behaviour between releases. A page cited in March can go unmentioned in June with nothing changed on your side. Treat this as a monitored practice rather than a project with an end date.