How to Get Recommended by ChatGPT (and Other AI Assistants)

To get recommended by ChatGPT, make it easy for AI to understand exactly what you do, who you serve and why you are a good choice, then make sure that information is consistent across your own website and the other places AI learns from: reviews, directories, comparison articles and community discussions. AI assistants recommend businesses they can describe confidently. Your job is to give them clear, repeated, trustworthy evidence.

How ChatGPT decides who to recommend

When someone asks an AI assistant “what’s the best alarm app for heavy sleepers?” or “which accountant in Kraków handles freelancers?”, the assistant builds an answer from what it already knows and, increasingly, from what it can look up on the web in real time. It tends to name businesses that are clearly described, frequently mentioned in relevant contexts and supported by third-party sources.

No one outside the AI companies knows the exact ranking rules, and they change. But the pattern is consistent: assistants favour businesses whose purpose is obvious and whose reputation is visible in more than one place. A vague homepage and no outside mentions make you easy to skip.

Step 1: Find out what AI says about you today

Before changing anything, ask the assistants the questions your customers ask. Not “what is [your brand]?” but the buying questions: “best [product] for [use case]”, “[service] near [city]”, “alternatives to [competitor]”. Ask ChatGPT, Perplexity, Gemini, Grok and Claude, and note:

• Whether you are mentioned at all.
• Which competitors are named instead.
• What the assistant says about you, and whether it is accurate.
• Which sources it cites, when it shows them.

This is your baseline. It is also exactly what the free AISeesYou report does for you.

Step 2: Say clearly what you do, for whom

Rewrite your homepage and key pages so a stranger (or a model) can answer three questions in one read: what is this, who is it for, and what makes it different. Use the words your customers use. If people search for “alarm app that makes you get out of bed”, say that, in plain language, near the top of the page.

• Put a one-sentence description in your page title, meta description and first paragraph.
• Create a dedicated page for each main product, service or use case.
• Name the specific problems you solve and the audiences you serve.

Step 3: Answer the questions people actually ask

AI assistants are built to answer questions, so pages that directly answer questions are easy to use as sources. Add an FAQ section, comparison pages (“[you] vs [competitor]”), pricing explanations and short guides. Lead each answer with a direct sentence, then add detail. Avoid hiding key facts in images, PDFs or scripts that are hard to read.

Step 4: Make your site easy for AI to read

• Allow AI crawlers such as GPTBot, OAI-SearchBot, PerplexityBot and Google’s crawlers in your robots.txt, unless you have a reason not to.
• Make sure important text is in the HTML, not only rendered by JavaScript.
• Add structured data (schema.org) for your organization, products, FAQs and articles.
• Keep a clean sitemap and consider an llms.txt file that summarises your site.

Step 5: Get mentioned where AI looks

Your own site is only one voice. Assistants also lean on what others say: review sites, app stores, industry directories, “best of” lists, news coverage, Reddit and other forums. Work on being genuinely present there:

• Ask happy customers for reviews on the platforms relevant to your industry.
• Get listed in reputable directories and comparison articles.
• Take part in community discussions honestly, where your product is a real answer.
• Keep your name, description and details consistent everywhere.

Step 6: Measure, fix, repeat

AI answers change as models update and as new content appears. Re-check the same questions regularly, see whether you now appear, and compare against the competitors that are recommended instead. Treat it like SEO: steady, ongoing work rather than a one-off trick.

Common mistakes to avoid

• Only asking about your brand name. Assistants will often describe you if asked directly; what matters is whether they name you when a customer asks a general buying question.
• Keyword stuffing. Repeating phrases doesn’t help an assistant understand you. Clear sentences that explain what you do work better.
• Fake reviews or spammy mentions. They can damage trust with customers and platforms, and they are not a reliable way to earn recommendations.
• Checking once and stopping. Answers shift over time and differ between assistants, so a single check can mislead you.
• Ignoring accuracy. If an assistant describes you wrongly, fix the source: update your site and the listings it is likely drawing on.

Start with the questions that matter most to your revenue, fix the obvious gaps, then work outwards. Small, consistent improvements across your site and your wider presence add up.

A real example

AISeesYou’s founder, Jacob, built Awaken, an alarm app that makes you dance or solve puzzles to turn it off. After optimizing it for AI search, ChatGPT started recommending it, and 26% of Awaken’s users came from ChatGPT and other AI, for free. That experience is why AISeesYou exists.

FAQ