Sellers ask us about ChatGPT, Perplexity, Gemini, Etsy, Shopify and schema. They also ask whether any of this is real. The answers share one idea. Shops of any size get named when AI can read them clearly. We're building Herald so every shop can have a place in AI answers, today and as AI agents start to shop.
Swipe the diagram to see all of it →
The questions cover ChatGPT, Etsy, schema and attribution. Most answers come back to one step. Make your facts clean, complete and easy for AI to read. Herald does that work for you.
Once these three make sense, most of the 50 answers follow.
Swipe the diagram to see all of it →
AI recommends what it can read and confirm. Great products get recommended when their facts are easy for a machine to read. Start there, with complete Product schema, answer-format content and consistent facts across the web.
The engine picks sources it can read and trust. Your quality shows when AI can pick it out, so spell it out with clear attributes and quotable specifics.
No. Organic AI recommendations aren't for sale. Selection runs on relevance, readability and trust, so a small, prepared seller can win.
There are two systems. The shopping surface pulls from enrolled product feeds. Conversational recommendations draw on search indexes, editorial reviews and forum discussions. It pays to be strong in both.
Guide: Shopify & ChatGPT shopping →
It helps a lot, because ChatGPT's live search leans on Bing's index. Submit your sitemap to Bing Webmaster Tools and make sure your best pages are indexed there as well as on Google.
Engines repeat the most consistent story the web tells about you. Publish one clear, structured source of truth and keep your facts the same everywhere. Corrections spread as models and indexes refresh.
Yes. Question-and-answer content with FAQPage schema is some of the easiest content for AI to use. Write real buyer questions, answer each in 40 to 60 factual words, and mark them up.
Yes. AI engines care more about clear structure today than about domain age. A new shop with clean, readable pages and a few real mentions elsewhere can get cited.
Perplexity searches live and cites sources it can parse. That means fast pages, direct answers, clear facts and clean markup. Pages that open with the answer do best.
AI Overviews combine regular ranking strength with how easy a page is to quote. Keep your Google SEO healthy. Then add the answer layer, with question-style headings, direct answers and complete Product/Offer/FAQ schema.
For informational queries, clicks often drop. For product queries, the cited source gets the click. So the opportunity is to become the citation.
The basics are the same everywhere. Make your pages easy to read, structured, answer-first and backed up by other sources. One great machine-readable surface works for all of them.
Guide: What is AEO? →
Yes, more and more. AI shopping is going visual-first. Descriptive alt text is one of the easiest wins you control right now.
Each is small on its own. Together they're part of the same shift. Pages built well for the big three are ready for the rest too.
SEO earns ranked links, and it's measured by position. AEO and GEO are two names for being cited inside AI answers, and they're measured by citations.
Guide: AEO vs GEO vs SEO →
It's fair to ask. Some vendors relabeled old audits, and the shift itself is real and measurable. Judge any tool by what you can see, like real citations and real referred visitors.
Both. SEO still drives most discovery today. Keep SEO healthy and add the answer layer on top.
They're useful when you can see what's behind them. A good score breaks down into parts you can act on.
See how the Herald Score decomposes →
Most run buyer-style prompts across engines, record whether you're mentioned and chart it. The content, schema and publishing work is usually left to you.
How Herald does the work →
You can check structural readiness the day it ships. Citations grow over weeks. Be careful with anyone who promises citations by a specific date.
Much of it, yes. Write answer-format content, add complete schema, publish llms.txt, allow AI crawlers and earn real mentions. The harder part to do yourself is a dedicated, machine-readable page for every product.
The AI-readiness checklist →
It's a plain-text file at your site's root that gives AI crawlers a map of your key content. It's quick to add and more engines read it, so it's worth doing.
Yes. AI conversations already drive a meaningful share of Etsy's referral traffic. Shops get named when their listings are clear and easy for AI to read.
Guide: Etsy & AI search →
Etsy's templates are Etsy's. You control plenty, though: conversational listing language, complete attributes and the pages you own outside the marketplace.
No. The two work together. Conversational, specific, attribute-rich listings help Etsy's own search and AI engines alike. An extra published layer lives outside Etsy.
Yes, indirectly. Reviews back up your claims, and engines value that. Reviews that crawlers can read, outside marketplace widgets, help the most.
A full second store means syncing inventory and more upkeep. A lighter option is a published discovery layer. It gives you the AI-readable presence of a website without running one.
How the published layer works →
Especially to you. AI questions are long and specific. That makes one-of-a-kind details a real advantage in answer search.
It's real and measured. Analytics firms tracking large retailers put AI conversations at a double-digit share of Etsy's referral clicks, growing quarter over quarter.
Complete the feed basics. Use descriptive non-branded titles, accurate live pricing and stock, quality images and a connected catalog. Listing happens automatically. Recommendations depend on data quality.
Being connected makes you eligible. Recommendations go to the catalogs AI can read best. Check what the engine sees, like your titles, attributes and schema.
Guide: connected ≠ chosen →
It ships basic Product markup, which is a good start. Adding Organization, FAQPage, AggregateRating and complete Offer details fills in the rest. It's one of the most straightforward AEO wins.
If you sell products, yes. AI needs to crawl your store to recommend it. Allowing these bots opens up the fastest-growing discovery channel.
It can. Fewer widgets between a crawler and your product facts make them easier to read. You can keep your site as it is for shoppers and give AI its own clean pages.
Pages that answer real buyer questions tend to earn more citations than generic product pages. The work is in building and maintaining them for every product.
AI visibility builds over time, and ads run as long as you pay. Starting early helps, because early citations in AI answers tend to grow.
Product and Offer (price, availability, condition), Organization (who you are), FAQPage (Q&As AI can quote) and AggregateRating (reviews), plus BreadcrumbList for structure.
The readiness checklist →
Not when it's done right. That means new answer-format content, correct canonical tags pointing to your store, and its own published layer.
Open with the answer in 40 to 60 factual, quotable words under a question-style heading. Say what it is, who it's for, what it's made of, what it costs and what makes it different.
Guide: What is AEO? →
Yes. Crawlers have limited time for each site. Fast, clean pages get read in full, so they have the best chance of being used.
It's worth fixing. AI reads text much more reliably than images. Put every attribute a buyer might ask about in text and in structured data.
Guide: helping AI recommend your shop →
They can. Near-duplicates split your signal, so the engine is less sure which page to trust. One strong page per product with clear variants works better.
Engines favor current sources. Accurate prices and stock keep shoppers' trust, so they should sync continuously.
You can get close. Load your pages with JavaScript turned off, run schema validators and see what's left. An AI-readiness score is meant to show the full picture.
The Herald Score →
Check referrers for chatgpt.com, perplexity.ai and gemini.google.com. The real number is likely higher, since much AI-referred traffic shows up as “direct.”
Referrer data gives you a rough picture. A published discovery layer can show the full path, from engine to cited page to click to your checkout, per product and per question.
How attribution works →
Agents research and compare products for shoppers, and more and more they buy too. Agentic checkout protocols are already live on major platforms.
Guide: agentic commerce →
Agents need to be able to act on your store. That means stable product URLs, machine-readable offers and policies, live availability and a clear path to checkout. It builds on AEO.
The AUX layer →
The advantage gets smaller and the need stays, just like SEO, where the basics still matter every day. Starting early gives you more time to build.
Give AI one place it can read perfectly. Build a clean, structured layer for your top products, with real answers to real buyer questions.
Most of these fifty questions come back to one step. For ChatGPT today or a buying agent tomorrow, clear and complete product data, backed up by other sources, is what works. Do it by hand with our guides, or, once Herald opens, let it build and publish that data for you.
Herald is in early access. Tell us about your store and we'll be in touch.