Getting traffic from ChatGPT & AI agents, where do I even start?

SEO
,
1

Hey everyone,

So I keep hearing about people getting traffic and even actual sales through ChatGPT and other AI assistants, and honestly I feel like I’m late to the party lol.

Right now pretty much all my traffic comes from the usual stuff — some Google SEO, Instagram, a bit of email. But with all this “agentic commerce” talk going around, I really want to figure out how to get my store in front of people who are shopping through AI.

A few questions for anyone who’s been playing around with this:

  1. How do I actually get more traffic from ChatGPT? Like when someone asks it “what’s the best X for Y” — how do I make sure my products are even part of that conversation?

  2. Is there anything specific I should do to get my store and products properly indexed by these AI tools? Do they look at product descriptions, schema markup, reviews…? I’m kinda lost on how this works tbh.

  3. Anyone here actually seeing real sales from this? Would love to hear what’s worked (or not worked) for you.

Any tips, apps, or resources would be super appreciated. Thanks in advance! :folded_hands:

2

There isn’t a magic submit-to-ChatGPT button. What I’ve seen matter most:

  • Make every product crawlable without JS or login. Submit your sitemap to Google Search Console and Bing Webmaster Tools, then check that product URLs aren’t blocked in robots.txt.
  • Validate Product schema. Include name, image, price, currency, availability, SKU/GTIN, brand, shipping, returns, and AggregateRating only when backed by real reviews.
  • Rewrite thin descriptions around specific use cases and buyer questions. Add size/material/spec tables, comparisons, FAQs, and clear shipping/return info.
  • Track AI traffic separately. Look for referrals from chatgpt.com, perplexity.ai, and Gemini, then ask customers how they found you because some visits show as direct.

I’m working on ClickFrom.ai, which checks this structured data and tracks AI citations, but you can do the basics above manually today.

4

@olgapapodoki There’s no submit-to-ChatGPT step - most of these tools answer shopping questions using the same crawled/indexed web as search engines (Bing’s index feeds a lot of it, plus live browsing), so it comes down to the same fundamentals done well, not a separate playbook.

Start with whether you’re even indexable: search site:yourdomain.com in Google for a few product URLs. If they’re not showing up there, they won’t show up in AI answers either - that’s step zero before anything else matters.

From there: add/validate Product structured data (JSON-LD) with price, availability, and review/rating data - these tools lean on structured data much more than they parse marketing copy. Then rewrite thin product descriptions into content that actually answers comparison questions (“best X for Y”, sizing, materials, who it’s for) since that phrasing is closer to what gets pulled into an AI answer than generic sales copy. Genuine on-site reviews help a lot too, since product recommendations from these tools skew toward items with visible review signals.

On measuring it: GA4 will undercount this badly since AI app traffic often lands as “direct,” so don’t panic if referrer data looks flat - the honest way to catch it early is asking new customers how they found you.

If this answer helped, mark it as a solution.

5

Hey @olgapapodoki!

What you’re describing is AEO (Answer Engine Optimization). It is basically SEO but for AI assistants instead of Google’s blue links. Instead of optimizing to rank, you’re optimizing to get cited - so when someone asks ChatGPT “what’s the best X for Y,” your product is what it pulls from.

The good news is if you’re already on Shopify, Agentic Storefronts (Settings > Sales channels) already syndicates your catalog to ChatGPT, Gemini, Copilot, etc. by default. So you’re not starting from zero.

But being in the catalog and being recommended are two different things. That’s where AEO comes in - structuring your product data (especially the product title & description) so AI can actually understand and trust it enough to suggest it.

:victory_hand: Hope it helps!
Ellie from BOGOS: Free Gift Bundle Upsell team

6

Hi @olgapapodoki,

Honestly, don’t worry about being “late”, most store owners are still just trying to figure this out.

AI bots like ChatGPT or Perplexity don’t really “browse” like humans. When someone asks for a recommendation, the AI looks for two things: what other sites (like Reddit or blogs) say about you, and how easy your own site is for a machine to read. If your product details are buried in messy code, the AI will just skip you and suggest a competitor instead.

As for sales, the direct traffic is usually pretty low because the AI often gives the answer right there on the screen. But getting mentioned is huge for brand trust. People see your name in the AI answer, then they usually go and search for you on Google later to actually buy.

The first thing you should do is fix your “Schema markup.” It’s basically a hidden cheat sheet that tells the AI exactly what you sell and how much it costs. If you want to handle this without messing with code, I’d suggest using SearchPie. It automatically builds the technical structured data that AI assistants need to find and recommend your products. It’s a great way to make sure your store is “AI-ready” while everyone else is still just doing basic SEO.

Hope this helps you get started!

7

It might sound like a cliché to say that it’s never too late to start haha but it’s not the most important thing when you start it but how you do it. AI traffic is blooming but it has not dominated traditional channels such as Google, social,.. so you always have room to start.

So the question that we need to answer is: how ChatGPT (or other AI engines) finds your products?

AI engines don’t rank the same way Google does, they tend to recommend products or pages that are:

  • clearly described and easy to extract trusted (reviews, brand mentions,…) because they literally “extract” rather than “read” the information.
  • already visible across the web (your site + articles, comparisons,..)
  • well strutured with clean product data (title, specs, use cases, FAQs,..)

So our goal here is to to become the obvious and citable answer. Some practices you can consider are making your product pages answer-ready, adding structured data, product schema, FAQs schema. These will make your sites and pages more extractable

image
image1968×736 83.3 KB

Next thing is building supporting content such as comparisons pages, use guide and link to relevant products (but don’t stuff link in content).

However, you should consider this channel as an additional one not switch to it to invest too much and forget about traditional SEO because SEO still dominates in this period of time. If you want to improve traditional SEO and develop AI visibility for your Shopify store, you can consider using some help like SEOWILL (here), it can support both SEO and AI visibility with on-page SEO, Product/FAQ schema, AI-friendly content structure

8

One thing that’s worked for us: being cited matters more than being indexed. AI assistants lean heavily on Reddit threads, niche blog roundups, and comparison posts when answering “best X for Y” questions, way more than they crawl your product page directly. So alongside cleaning up schema, it’s worth just showing up in those honest third-party conversations since that’s often literally the source getting quoted back.

11

@olgapapodoki Before you buy any schema app, check whether you already have schema.

Most modern Shopify themes already output Product structured data on their own. Open one of your product pages, view source, and search for application/ld+json. Or just paste the URL into Googles Rich Results Test, its free and takes a minute.

You want to see exactly one Product block. If you install an app that adds schema on top of what your theme already puts out, you end up with two Product blocks describing the same page, and that is worse than having one. Conflicting data tends to get ignored rather than merged. So check first, and only reach for an app if the check comes back empty or broken.

Second thing, and this one is a real gate that nobody has mentioned yet.

Agentic Storefronts is mostly on by default, thats right, but there are conditions attached and if you miss one you are simply not in the catalog. Go to Settings, then Policies, and make sure your Terms of service, Privacy policy, and Return and refund policy are all genuinely filled in. Not the empty template stubs, real published text. Shopify requires all three before your products are eligible for ChatGPT. Its a ten minute job and its probably the highest value thing in this whole thread.

Also worth knowing that the ChatGPT channel currently needs your store to sell to buyers in the US. Your store can be based anywhere in the world, thats fine, but if you dont sell into the US you will not show up there at all.

Then go to Sales channels, then Agentic, and look at the catalog count. If that number is lower than your active product count, some products are being rejected, usually for a missing required field like brand or GTIN. That gives you an actual list to work through instead of guessing.

One more, since a couple of people above said GA4 will hide this from you. That used to be the whole story, but Shopify now has its own reporting for this. The Agentic channel has a performance view showing sessions, referrals and sales coming from AI channels. Check there before you fall back on asking customers how they found you.

On your third question, whether anyone is seeing real sales. Be a little careful with the growth numbers floating around, they look enormous mostly because the starting point was close to zero. Worth setting up properly since the setup is free and takes an afternoon, but I wouldnt move budget away from your Google and email work for it yet.

12

The first thing I would check is of I have done GEO on my store.

14

You’re not too late honestly, this is still early. Two things matter most: external mentions (reviews, editorial coverage, third-party sites) and clean structured data (Product/Offer/FAQ schema) stores with thin external coverage rarely get cited even with great product data, so it’s less about your site alone and more about who’s talking about you elsewhere. If you’re on Shopify, check out their ChatGPT product feed integration to get in the mix directly. And yeah, real sales are happening Shopify reported AI-referred orders up 11x between Jan 2025 and early 2026 but it takes some setup, not just waiting for it to happen.

15
  1. Check that your products are listed in Shopify Catalog
  2. Run and test queries in Agentic sales channel on Shopify admin to see where you’re ranked
  3. Make sure your product are filled with relevant category taxonomy and metafileds
  4. Maintain good trust and reputation using reviews and low return rate.
  5. Fill FAQ’s in the knowledge-base app
16

I’ve been looking into this too, and I think the first thing I’d do is make sure the normal SEO side is actually solid before trying to optimize specifically for ChatGPT.

I’ve seen stores spend a lot of time adding AI-specific stuff while the important product and collection pages still have indexing issues, weak internal linking, or barely any useful information on them. If an AI system can’t make sense of the store in the first place, adding more AI-focused content probably isn’t going to fix much.

For the actual AI side, I’ve mostly been checking the same product queries manually in ChatGPT and Perplexity and seeing which stores keep coming up. Reviews and third-party mentions seem worth paying attention to as well, since the answer isn’t necessarily coming straight from the merchant’s own site.

I use SiteGuru for the regular site checks, mainly because it gives me a prioritized list rather than another huge audit to work through. I’m not treating it as an AI visibility tool, but it helps make sure the basics aren’t being neglected while this stuff is still figuring itself out.

The sales attribution is the part I’d really like to see improve. Getting mentioned in an answer is interesting, but knowing whether that mention actually brought someone to the store and resulted in an order is a much harder problem.

17

You’re not late. The thing everyone was rushing to get into last year got cancelled, and the part that actually matters is mostly already switched on for you.

Start in your admin, not with an app.

Shopify turned Agentic Storefronts on by default for eligible US-selling merchants in March 2026 as part of the Winter '26 Edition, with no opt-in required. Go to Sales channels > Agentic. If your products are in Shopify Catalog, they’re already discoverable in ChatGPT. Google AI Mode and Gemini come through the Google & YouTube sales channel, and Meta surfaces through the Facebook and Instagram channel, so those need the relevant channel installed. Same screen lets you preview how your products rank in Catalog search and see performance data. That’s your answer to question 2 and it costs nothing.

Don’t buy an app promising in-chat checkout. OpenAI ended Instant Checkout in March 2026 and moved purchases back to merchant sites while keeping discovery inside ChatGPT. Walmart had measured in-chat checkout converting roughly 3x worse than a click through to their own site, even though ChatGPT was sending about 2x the new-customer rate they see from search engines. The pattern that stuck is discover in AI, buy on your site. Which is good news, because your existing checkout, email flows and pixels all still work.

What actually decides whether you get recommended. There’s no bidding and no creative format here. The model picks products on data quality and relevance to the query, which means a small store with a clean catalog can out-surface a much bigger competitor. So the boring work is the work:

  • Fill in every product attribute, not just title and price. Material, size, colour, use case, who it’s for.
  • Availability and price accuracy. Stale inventory data gets you dropped.
  • Write descriptions that answer the query, not the brand mood. If people ask “best X for sensitive skin”, the phrase and the reason need to exist on the page in plain language.
  • Reviews matter, but third-party mentions matter more. Being named in someone else’s “best X for Y” roundup, a Reddit thread, or a YouTube comparison does more than your own page claiming you’re the best. That’s the part most merchants skip because it isn’t a toggle.

Skip llms.txt. Apps and agencies are selling it hard right now. Google’s own guidance says you don’t need llms.txt or special markup to appear in AI Overviews or AI Mode, and an Ahrefs study of 137,000 sites found 97% of llms.txt files got zero traffic in May 2026. Thirty minutes to ship if it makes you feel better, but it isn’t the lever.

On real sales. AI-referred traffic to Shopify stores grew about 7x between January 2025 and early 2026, with AI-attributed orders up around 11x. Big multiples, small starting base. Treat it as a real acquisition channel that’s still early, not a replacement for anything you’re doing. Before you optimise anything, make sure you can actually see it: check your GA4 referral sources for chatgpt.com, perplexity.ai and copilot.microsoft.com, and use the attribution in the Agentic channel in admin. Most people asking “is anyone getting sales from this” already are and just aren’t measuring the source.

One distinction worth holding onto: being mentioned is not the same as being cited or linked. A lot of AI visibility tools count mentions and call it traffic. Only one of those sends you a customer. If you drop your store URL I’ll tell you which of these is your actual bottleneck

19

Solid answer above, the data quality point is the one that really matters. one thing worth adding though that even a clean catalogue can trip over — check you’re not accidentally blocking the bots that actually do the citing!

There are two kinds of AI crawler and most stores treat them as one. Training crawlers (GPTBot, CCBot etc) build the model. The citation/inference bots — ClaudeBot, Google-Extended, Applebot-Extended, PerplexityBot, OAI-SearchBot — are the ones that fetch your page the moment someone asks a question, to quote you with a link. If you’re on cloudflare with “managed robots.txt” switched on, the default list blocks several of those citation bots alongside the training ones. so you can do all the catalogue work and still be invisible to the answer, because the bot that would quote you cant reach the page

Worth a two min check: look at your robots.txt and your Cloudflare AI crawl settings and make sure the citation bots are allowed, even if you’re blocking the training ones :slight_smile: cheapest fix on the list and the easiest to get wrong by accident

20

Shopify now offers an agentic AI sales channel to help with this. Make sure you have it set up in your channels and have filled out all of the prerequisites.

More information:

21

I can help you with this, I have got like 6000 DAU directly from chatgpt and other LLM’s

22

Hi @olgapapodoki

It’s still early, don’t worry.

  1. There’s onpage and off page like SEO. Onpage - if your getting traffic from Google its all good, but having the products, cat pages, blogs - about pages are a must too. Offpage - being mentioned in listicles and citations that keep shoing up for the keywords you want to rank for.

  2. I like to reverse engine whats working. Search the prompts, pull the pages of whats being mentioned, see the details on the product pages and if your missing anything - we built this into shop mentions to make it a bit easier than manually doing it.

  3. We’ve been seeing decent AOV and majority through ChatGPT so far, with Perplexity being second.

I built a quick ai visibility checker for Shopify stores Free AI Visibility Scan for Shopify Stores | Shop Mentions

Best, James

23

@olgapapodoki You’re not late! AI shopping still relies heavily on strong product data. Focus on clear product descriptions, structured data/schema, reviews, fast pages, and consistent product information across your store. There’s no magic “ChatGPT SEO” button make your products easy for AI to understand, trust, and recommend.