Blog

How to get your business recommended by ChatGPT (and other AI assistants)

GroundReach

When someone asks ChatGPT “who’s the best bookkeeper in Melbourne?” or “what’s a good CRM for a small agency?”, the answer usually names a handful of businesses. Getting your business into that answer is what people mean by “ranking in ChatGPT”. There’s no ranking page to climb, and no one can guarantee a mention, but there is a lot you can do to make one more likely, and you can measure whether it’s working.

This guide covers how AI assistants choose who to name, the steps that make a difference, and how to track the result.

The short version

  • AI assistants with web search (ChatGPT, Perplexity, Gemini, Claude) look things up much like a search engine, then write an answer from what they find.
  • They can only use pages they’re allowed to read, and they lean heavily on what other sites say about you.
  • So: let AI crawlers in, describe your business clearly, answer your buyers’ real questions on your own pages, get mentioned on the sites assistants already cite, and check regularly.

How AI assistants decide who to name

When web search is on, an assistant turns the question into searches, reads a set of pages, and summarises what they agree on. If several trusted pages mention your business for the thing being asked about, you’re more likely to appear. If your site is the only place that says you’re good at it, you’re less likely to.

Without web search, an assistant answers from what it learned in training. For newer or smaller businesses, that usually means it doesn’t know them at all. When we asked an AI model, without web search, our own buyer questions in October 2026, it named HubSpot and never named GroundReach, because our site was days old. That’s normal, and it’s why the live web is where the work pays off.

Step 1: Let AI crawlers read your site

Each AI company runs its own crawlers, for example OpenAI’s OAI-SearchBot, PerplexityBot and Anthropic’s Claude-SearchBot. If your robots.txt blocks them, or your firewall does, those assistants can’t read your pages and are less likely to recommend you.

  • Check robots.txt for rules that block AI search crawlers. Blocking training crawlers (such as GPTBot or Google-Extended) is a separate choice and doesn’t stop AI search.
  • If your site sits behind Cloudflare or a similar service, check its bot settings too: they can block crawlers that robots.txt allows.

Step 2: Say clearly who you are and what you do

Assistants have to be confident a page is about your business before they name it.

  • Use the same business name, location and description everywhere: your site, Google Business Profile, Bing Places, Apple Business Connect and directories.
  • Add structured data (schema.org Organization or LocalBusiness) to your homepage, with links to your official profiles (sameAs).
  • Say plainly who you serve and what you offer, near the top of the page, in words a buyer would use.

Step 3: Answer your buyers’ real questions

Assistants quote pages that answer the question directly. Write down the questions your buyers actually ask (sales calls, emails and Google’s “People also ask” are good sources) and answer each one:

  • Put the direct answer first, in two or three sentences, then the detail.
  • Use the question as a heading.
  • Keep facts specific and current: prices, locations, who it’s for, how long it takes.
  • Mark up genuine question-and-answer sections with FAQPage structured data.

Step 4: Get mentioned where AI assistants look

This is the step most businesses skip, and it often matters most. Assistants rely on other sites to decide who’s worth recommending: reviews, directories, industry publications, comparison articles and community discussions.

When we checked ten AI-marketing keywords in Google Australia in October 2026, LinkedIn appeared in the top ten for seven of them and Reddit for six, and both were among the sources Google’s AI Overviews cited. Your market will have its own equivalents. Find the sites that keep coming up when assistants answer questions in your category, and earn a genuine presence there: a listing, a review, an expert answer, a guest article.

Step 5: Measure, then repeat

Because assistants don’t give the same answer twice, one check isn’t enough. A useful routine:

  1. Keep a fixed list of 5–10 buyer questions.
  2. Ask each assistant every question several times, and record whether you’re named and who’s named instead.
  3. Note which sources the assistants cite when they name a competitor and not you: those are your next outreach targets.
  4. Repeat weekly or monthly, and look at the trend, not a single result.

This is what GroundReach’s AI visibility check does: it asks ChatGPT, Perplexity, Gemini and Claude your buyer questions with web search on, shows a mention rate with an honest range, and lists the sources behind your competitors’ mentions.

Common mistakes

  • Blocking AI search crawlers by accident, often through a security setting rather than robots.txt.
  • Only optimising your own site. Without mentions elsewhere, assistants have little reason to trust it.
  • Chasing a single result. A mention one day and none the next is normal variation; track it over time.
  • Writing for bots. Pages that are clear and useful to people are the ones assistants quote.

Frequently asked questions

Not in its regular answers. What you can do is improve the signals assistants rely on: a readable, clearly described site, direct answers to buyers’ questions, and mentions on trusted sites.

Is SEO dead now that people use AI?

No. AI assistants with web search still find pages through search-like retrieval, so a crawlable, well-structured site with useful content helps you in both. What’s added is the weight of being mentioned on other sites and of answering questions directly.

How long does it take?

It varies. Changes to your own site only count once assistants and search engines re-read your pages, and building mentions on other sites takes longer still. Measure at least monthly and judge the trend.

Can a beginner do this?

Yes. Most of the steps above are about clear writing and consistency, not code. The technical parts (robots.txt, structured data) are small and well documented.


Want to see where you stand? GroundReach checks how often AI assistants name your business and what to fix next.