How does a D2C brand show up in ChatGPT, Perplexity and Google's AI results?

Why is the brand suddenly asking about ChatGPT?

It usually starts with a forwarded screenshot. A customer asked ChatGPT or Perplexity for the best option in the brand's category, and the answer named rivals, not the brand. Or Google's AI Overview linked to a marketplace listing instead of the brand's own page. The founder asks: why are we missing, and what are you doing about it?

The honest answer has two parts. Some of how these systems choose sources is publicly documented; much of it is not. This guide separates the two, gives you the fixes that are documented to matter, and shows how to measure progress without inventing a ranking factor.

What does it cost to get AI search wrong?

Doing nothing has a cost. When a shopper asks an AI assistant for a shortlist, the brands it names get considered. And if the brand's facts are inconsistent across the web, an answer can repeat an old price or the wrong return policy, and the brand hears about it from an annoyed customer.

Overselling costs more. Special files, rewritten copy, bought mentions, guaranteed placement: Google's own guidance calls much of this a myth. If the agency sells a hack and nothing moves, the retainer is at risk, and so is its credibility on everything else it reports. Some call this work GEO (see generative engine optimisation in the glossary); Google treats it as SEO for a new search experience, and that is the right way to scope it.

How do AI answers decide which brands to name?

The middle column below was checked against each platform's own documentation in October 2026. The right column is inference: patterns worth acting on, but not documented ranking rules.

Swipe the table sideways to see every column.

Sources: Google's AI features and your website and guide to optimising for generative AI search; OpenAI's overview of its crawlers and shopping results in ChatGPT search; Perplexity's crawler documentation.

Two points follow. Crawler access is the one hard requirement you can verify today: if robots.txt or the CDN blocks Googlebot, OAI-SearchBot or PerplexityBot, nothing else matters. And answers vary with phrasing, location, user history and the day, so one screenshot proves nothing. You need a fixed set of prompts, checked on a schedule.

What does Google say about AI Overviews and AI Mode?

Google's Search Central page on AI features is short and direct. The lines worth quoting to a brand:

• There are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary."
• "There's also no special schema.org structured data that you need to add."
• The fundamentals still apply: allow crawling in robots.txt and CDN settings, link internally, keep important content in text, and keep Merchant Center and Business Profile data up to date. Structured data must match the visible page.
• To limit what is shown, use the normal preview controls: nosnippet, data-nosnippet, max-snippet or noindex. Google-Extended is a separate control for Google's other AI systems, not a switch for AI Overviews.

Google's newer generative AI optimisation guide calls unique, first-hand content the factor most likely to shape a site's presence in generative AI search, and warns that mass-producing pages for query variations can breach the scaled content abuse policy. It also says no third-party tool has access to Google's internal ranking or AI systems. Remember that when a vendor shows a brand a "Google AI score".

How do you make the brand's facts consistent everywhere?

The cheapest fix is removing contradictions, so any system reading about the brand finds one consistent business. Build a one-page fact sheet with the brand, then check every surface against it.

Fix the fact sheet first

Agree the trading name, a one-line description, category, address and phone where relevant, price band, shipping regions, return policy and main ranges. One version of each, signed off by the brand.

Check each surface against it

• Website: About page, footer, contact page, policy pages and Organization markup.
• Google Business Profile: Google's guidelines for representing your business say the name should match what appears on the storefront, website and stationery, with one profile per business and the most specific category that fits.
• Marketplaces: Amazon, Flipkart or wherever the brand sells. Name, description and policies should match the site.
• Social profiles: bios and links on every active channel.

Then connect them with the sameAs property in Organization markup, which Google's Organization structured data documentation says helps it tell an organisation apart from others with similar names.

Which structured data matters for a D2C site?

Google says structured data is not required for AI features. It is still worth doing, because it makes pages eligible for rich results and gives every system a clean copy of the facts.

• Organization, on the home page or About page. Use the most specific subtype, such as OnlineStore. Include name, url, logo (at least 112×112 px), contact details and sameAs. Business-wide return policies can be nested here.
• Product with an Offer, on every product page where shoppers can buy. Google's Product structured data overview explains the split: product snippets for pages where people cannot buy, merchant listings for pages where they can. See the schema.org Product type for the full vocabulary.
• FAQ markup, with a caution. Google deprecated the FAQ rich result and, per the Search Central documentation updates log, it has not appeared in Search since 7 May 2026. Keep visible FAQ sections that answer real buying questions, for readers, not for a rich result.

How do merchant listings and product rich results fit in?

For a D2C brand, product data is where AI search and classic search overlap most. Google's merchant listing documentation requires a product name, image and an Offer with a price above zero and a three-letter priceCurrency. It recommends brand, gtin, availability, shippingDetails and hasMerchantReturnPolicy. Google's AI guide names Merchant Center feeds as a way to help products appear in AI responses, and OpenAI lists availability and price among its merchant ranking factors. So keep stock and prices correct on the page, in the markup and in the feed.

Check two Search Console reports: Product snippets and Merchant listings. On Moolwan, the Bengaluru home décor brand run by the ai-dm founders, 1,400 products are eligible for rich results, with 24 of 1,424 flagged (Search Console · Product snippets · 26 Sep 2026), and there are 1,380 valid merchant listings, 1 invalid (Search Console · Merchant listings · 26 Sep 2026). The flagged items are the work queue.

What content answers buying questions?

AI answers draw on pages that answer the question asked, and D2C buying questions are predictable. List them with the brand's sales and support team, not from a keyword tool alone.

• Choosing: what to look for, which size or material suits which use.
• Comparing: this type versus that type, this range versus a rival's.
• Trusting: is the brand genuine, how do returns work, how long does delivery take.
• Using: care, installation, what to pair it with.

Write each answer from first-hand knowledge: the brand's own testing, measurements and photos. Put the direct answer near the top. One strong page per question beats thin variants, and Google says it needs no special phrasing, chunking or ideal length.

Moolwan's own guides show what this kind of page can do. Two of them earned 8,479 and 7,853 impressions in Google's AI results (Search Console · Generative AI features · 25 Jun–23 Sep 2026). ai-dm drafts SEO articles and schema for the team to edit and approve before anything is published.

How much do reviews and third-party mentions matter?

This is the least documented area. OpenAI ranks merchants partly on "quality", without defining it; Google says Business Profile data helps local businesses appear. Neither says how reviews are weighted. The reasonable inference is that genuine reviews and credible mentions make a brand easier to recommend, while fake ones carry risk: Google says seeking artificial mentions is "not as helpful as it might seem".

• Ask for reviews on Google Business Profile and the marketplaces after delivery, and reply to them, good and bad.
• Pitch real roundups, press and creator reviews, with clear disclosure where products are gifted.
• Monitor mentions for wrong facts and correct them at the source.

Should the brand publish an llms.txt file?

llms.txt is a proposed plain-text file that points AI systems to a site's key pages. It is not a standard. Google's AI optimisation guide says Google Search does not use it and it will not help or hurt rankings, and we found no OpenAI or Perplexity documentation saying their search uses it. Publishing one is cheap; never sell it as a fix or report it as progress.

How do you measure AI search visibility?

Use three layers, because no single tool sees everything.

Search Console's generative AI performance report

Google's Search generative AI performance report shows impressions from AI Overviews and AI Mode by page, country, date and device, and has been available to all websites since 31 August 2026. It has no clicks and no query dimension, so it shows which pages are used as sources, not which questions triggered them. Clicks from AI features sit inside the normal Performance report under the Web search type, and Google's Performance report help notes that an AI Overview occupies a single position, shared by all its links.

GA4 referrals

OpenAI's publishers and developers FAQ says ChatGPT adds utm_source=chatgpt.com to links in its search results, so those visits show chatgpt.com as the source; Perplexity visits usually show as a perplexity.ai referral. Check session source for visits and first user source for new users; Google explains the difference between traffic-source scopes. On Moolwan, GA4 shows 368 new users from chatgpt.com (GA4 · First user source · 29 Jun–26 Sep 2026). Not every AI answer produces a click, so treat this as a floor.

Prompt tracking and AI share of voice

Write a fixed set of prompts that mirror real buying questions, run them in ChatGPT, Perplexity and Google on a schedule, and record which brands are named and which pages cited. The share of prompts naming the brand, against rivals, is its AI share of voice. Keep the prompt set stable so month-on-month changes mean something.

What are the steps, and how do you check each one worked?

Swipe the table sideways to see every column.

What do Moolwan's numbers show, and what don't they?

Moolwan is where ai-dm is built and run every day, so its numbers are the honest test case.

They show the brand is used as a source in Google's AI results and that ChatGPT sends new users. They do not show which questions triggered the impressions, or that any single fix caused them. Say both when you present numbers like these. See them on the Moolwan live numbers page.

How does ai-dm help an agency run this?

ai-dm tracks AI search visibility across ChatGPT, Perplexity and Google AI results for each brand, so the prompt set runs on a schedule, not by hand. Its site and SEO audit ranks findings by severity, including schema and crawl problems. It monitors reviews and mentions, and drafts articles, schema and replies for the team to approve. ai-dm recommends; the team decides.

What should you do this week?

Pick one brand that has asked about AI search and run the first three steps before writing a proposal. Its robots.txt, fact sheet and Merchant listings report will teach you more than any screenshot.

Related pages

• Why did a brand vanish from ChatGPT answers?

  A brand stopped showing up in ChatGPT answers. How to confirm it with a fixed prompt set, find the likely cause, fix it in order and track it from there.

  Read the post →

• Impressions up, clicks down: is it AI Overviews?

  Search Console shows more impressions and fewer clicks. How to tell whether AI Overviews explain it, what to tell the brand and what to do next.

  Read the post →

• How should an agency scope, price and report AI search as a service?

  Scope AI search work an agency can deliver: audit, prompt set, schema, content, product data and reviews, priced in tiers, with no promised placement.

  Read the guide →

More guides

• Agency growth9 min read

  How should an agency scope, price and report AI search as a service?

  Scope AI search work an agency can deliver: audit, prompt set, schema, content, product data and reviews, priced in tiers, with no promised placement.

  Read the guide →
• Reporting10 min read

  How can an agency cut month-end reporting time without losing the story?

  Find where reporting hours go, time your own process, automate the pulls and keep the story human. A weekly rhythm and narrative template inside.

  Read the guide →

Questions agencies ask