How to Get Your Brand Mentioned by AI: A Step-by-Step Plan

Getting your brand mentioned by AI is mostly about what the rest of the web says about you. ChatGPT, Gemini, Claude, Perplexity, Copilot and Google’s AI features build answers from what they learned in training and what they find through live search, and both draw heavily on independent sources. McKinsey estimates that brand-owned sites make up only about 5 to 10% of the sources AI search references. So the plan below spends one step on your own site and most of its effort on the web around it.

The order matters. Each step makes the next one count for more, and skipping the first step means you will never know whether the rest worked.

Why does the order of these steps matter?

There are two ways an AI engine learns about a brand, and the plan has to serve both. The first is training data: text the large language model read before a fixed knowledge cutoff. Anthropic lists its newest Claude models with a training data cutoff of June 2026, and OpenAI’s current API models have knowledge cutoffs in spring 2026. Anything that happened to your brand since then is invisible to that built-in memory.

The second route is live retrieval, also called retrieval-augmented generation: the engine runs a search, reads pages and writes an answer from them. This is where new mentions can show up first, because a page can be found as soon as it is published and indexed. Our guide on how AI chooses which brands to mention explains both routes in detail.

The steps run in this order because measurement gives you a baseline, crawler access and consistent facts make every later mention easier to read and trust, and recommendations only become possible once the engines know you exist.

Step 1: Measure where you stand before changing anything

Write down 20 to 50 questions your buyers actually ask. Mix the shapes: “best project management tool for a ten-person agency”, “is [your brand] any good”, “[competitor] alternatives”, “where to buy trail running shoes in Austin”. Run each prompt in ChatGPT, Gemini, Perplexity and Google, and note whether your brand appears, where, and whether your site is cited.

Run each prompt more than once. In a SparkToro study with Gumshoe, ChatGPT and Google’s AI had less than a 1 in 100 chance of returning the same brand list in any two runs of the same prompt. A single screenshot proves nothing. What you want is a mention rate: the share of runs that name you. Our full guide on how to measure AI visibility covers prompt sets, scoring and tracking, and the free AI visibility self-check gives you a ready-made grid.

Step 2: Make sure AI crawlers can read your site

A crawler is the program an engine uses to fetch web pages. Several AI companies document theirs, and some of them decide whether you can appear in answers at all. OpenAI states that sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, while GPTBot is a separate crawler used for training. PerplexityBot surfaces and links websites in Perplexity results. Anthropic says blocking Claude-SearchBot or Claude-User can keep your content out of Claude’s search answers.

Open your robots.txt file and check that none of these are blocked by an old rule or a security plugin. For Google, a page has to be indexed and eligible to show a snippet before it can be linked in AI Overviews or AI Mode. Blocking Google-Extended affects Gemini training, not Search inclusion, so that choice is yours to make on other grounds.

Step 3: State your brand facts clearly on your own site

Your site is a small share of what AI reads, but it is the one source you fully control, and other writers copy from it. Write one about page that answers, in plain sentences: what the company does, for whom, where it operates, when it started, and what it is not. Take an illustrative company: “Northfield Labs makes bookkeeping software for UK sole traders” is far easier for a model to quote than “We empower the future of finance.”

Keep product pages specific too. Name the use cases and the types of customer, because those words become the topic context engines associate with your brand.

Step 4: Make your facts identical everywhere

Engines compare sources. If your LinkedIn page says you were founded in 2019, your site says 2020 and a directory lists an old address, you give the model a reason to hedge or get it wrong. Go through your social profiles, review sites, industry directories, app store listings and, for local businesses, your Google Business Profile, and make the name, one-line description, category, location and key facts match.

Local businesses should treat name, address and phone (NAP) consistency as basic hygiene. Google’s own Business Profile rules say your listed name should match how you appear on signage and stationery, and adding extra terms to the name can get the profile suspended.

Step 5: Earn and place mentions in context on real publications

For most brands this step has the most room to move the numbers, and it is the one most often skipped. In an Ahrefs study of 75,000 brands, how often a brand was mentioned across the web had the strongest correlation with AI Overview visibility (0.664), well ahead of backlink counts (0.218). That is correlation, not proof of cause, but it is consistent with how retrieval works: an engine can only name brands that appear in the pages it reads.

What counts is a brand mention in context. “Northfield Labs, a bookkeeping app for UK sole traders” teaches an engine far more than a bare name in a sponsor list. Good mentions sit on real, indexed publications that write for their own readers, use accurate facts, and place your brand next to your category, use case and market. Our guide on why context beats volume shows how the surrounding words shape answers.

There are several honest ways to get there: digital PR and expert commentary, partnerships and customer stories, genuine community participation, podcasts and YouTube, and placed mentions on publications that accept sponsored or partner content. If a placement is paid, follow the rules. Google’s spam policies say paid links must carry rel=”sponsored” or rel=”nofollow”, and disclosure law applies in each market. The piece itself should be written for that publication’s readers; content placed mainly to borrow a host site’s rankings is what Google calls site reputation abuse. This is general information, not legal advice; our guide to paid mention rules covers the details.

Never buy fake reviews, post as fake customers on Reddit, or mass-produce AI-written pages. Fake reviews break consumer law in the US, UK and EU, fake posting breaks platform rules, Google treats mass-produced pages as scaled content abuse, and all three can poison what AI says about you.

Step 6: Give engines clean entity data

An entity is a thing a search engine or AI system recognizes as distinct: your company, as opposed to another business with a similar name. Help engines identify it.

• Add Organization schema to your home page or about page, with your legal name, logo, URL and founding details.
• Use sameAs to list your official profiles: LinkedIn, YouTube, Crunchbase, Wikidata and similar.
• Check whether your company qualifies for Wikidata. An item can be accepted if serious, publicly available references describe it, which is a lower bar than Wikipedia.
• Do not write or edit a Wikipedia article about your own company. Its conflict of interest policy asks you to suggest changes on the talk page instead, and paid editing must be disclosed.

Our guide on making your brand a known entity goes further.

Step 7: Re-measure with the same prompts

Rerun your prompt set every few weeks, with the same wording and the same number of runs per prompt. Compare mention rate, position in lists, and how often your site or your mentions are cited. Track your main competitors in the same runs so you can see share of voice, not just your own number.

Expect noise. AI answers vary from run to run, so look for a trend across several checks before drawing conclusions, and treat any single good or bad answer as one data point. Results are measured, never promised.

Step 8: Move from mentioned to recommended

Being named is step one. Being named first, and suggested for the buyer’s exact need, is a different problem that depends on how strongly and consistently the web connects your brand to that need. BrightEdge found that commercial wording such as “where to buy” produced 4 to 8 times more brand mentions in ChatGPT than informational questions, so those buying prompts are where position matters most.

When your mention rate is steady, shift effort toward your most valuable topics and toward the comparison and “best for” prompts your buyers use. Our guide on going from mentioned to recommended explains when you are ready. For ChatGPT specifically, The Blue Ocean GPT is the step 2 service built for that job.

Where does The Blue Ocean Brand Mentions fit in this plan?

It covers steps 1, 5 and 7 for brands that want them done for them. We start with a free AI visibility check, agree your brand facts with you once, then place mentions in context across a large private network of real, indexed publications in many niches and countries. First mentions typically go live within 30 minutes of kickoff, with more added over time. You get the live URL of every placement and re-checks across ChatGPT, Gemini, Perplexity and AI Overviews. Whether and when an engine picks those mentions up depends on indexing, retrieval and model updates, so we report what changes rather than promise it. Our methodology page sets out what we do and what we will not do.

What to do next

• Write your prompt list today and run it at least three times per engine to get a baseline.
• Check robots.txt and your about page this week; both are quick fixes.
• Get a free AI visibility check if you want us to run the baseline across engines for you.