My products are invisible to ChatGPT. I tried three Shopify apps to fix it, and this is what works

I’m setting shop on Shopify, and optimising asap so AI Agents can find and recommend my products before anyone else does catches up on this TAM.

I’ve been selling pop-culture illustration prints on RedBubble for years. Throw blankets, prints, stickers, T-shirts. The work is mine, the catalogue is mine, and after RedBubble’s cut and their platform fees, I keep around 25% of what a customer pays. On that $40 throw blanket, I see roughly a tenner.

You can do the maths on my earning, it isnt great. And at any volume — it’s not worth getting out of bed for. This is why i had a bright idea setting up a new store with Shopify.

A months ago I started moving the catalogue to my own Shopify store. The plan was simple — I keep the art the same, keep the products the same, but own the storefront + own the headache of print + distribution. Margin doubles or triples on most items. Email list is also mine. In The worst case I keep RedBubble running and Shopify becomes a side experiment.

https://www.redbubble.com/people/illustraman/shop

I underestimated one thing. AI Shopping Assistants!

On RedBubble I had a built-in discovery engine. Type “Mandalorian throw blanket” into Google or RedBubble’s own search and my products show up in the listings. Years of marketplace SEO and category indexing did the heavy lifting. On Shopify, it’s day one. I’m a new domain with no authority, no inbound links, no category placements. All the storefront customisation in the world doesn’t help if nobody finds the store.

I’ve ground through enough Shopify SEO blog posts to know the playbook. Think meta titles, alt text on every image, internal linking, schema markup, product feeds for Google Shopping. I ticked all those boxes. 🫣

Then I noticed something that wasn’t in any of those blog posts.

The thing nobody is writing about yet

Out of curiosity I tried “show me the Mandalorian-inspired throw blanket under $50” into ChatGPT.

It recommended five products. Obviously of them were mine. One was from another store as a competitor. Their product page was, generously, fine. They were getting recommended — ChatGPT didnt even have a close my store existed, — well that hurt 😢.

I tried Perplexity. Same shape, different competitor. I tried Gemini. err, that was a whole diffent issue with google search recommendations.

This was a visibility problem. You see, The AI shopping agents like ChatGPT Shopping, Perplexity, Gemini, Google’s AI Overviews — could obviously see my competitor’s catalogue — just not mine (not that i pay for sponsorting me). The volume of product searches happening through these surfaces is climbing fast. **Shopify themselves announced in March 2026 that AI-attributed orders on their platform grew 11× year-over-year. Eight times the AI traffic**.

Catalogue-readiness for AI agents went from “interesting in a year or two” to “if you dont get this done, AI Shopping assisntat wont know or showcase your inventory, today.”

I started looking for apps that could fix this.

Let’s get “agent ready” & what does that mean?

I’d never heard the phrase before — Agent Shopping Ready. Sounds cool, but isnt this another SEO buzz word? It turns out there are many ingredients, and almost no one talks about these together. Here just a few to consider;

• Bot access. Your store’s robots.txt file decides which crawlers get in. Shopify's defaults sit in a frustrating middle ground — they let some bots through, block others, and most store owners never check. If you've blocked GPTBot, ClaudeBot, or PerplexityBot in there (intentionally or not), your products will never be in their training data, never be in their search results, never be in their answers. Period.
• Structured data — Schema.org JSON-LD. Robots loooove structure. When the bot does crawl your product page, can it understand what it’s looking at? A page with text saying “Mandalorian Helmet Sketch on a 60×80 inch throw blanket, soft fleece, $42.99” is one thing. The same page with a JSON-LD block declaring Product with name, image, offers.price, offers.priceCurrency, aggregateRating, etc., is entirely different. The AI can lift the second one into a product card. The first, it has to guess about — and it'll often skip pages it can't parse cleanly when there are cleaner alternatives elsewhere.
• Universal Product Identifiers — GTIN, EAN, UPC, MPN. This one I didn’t appreciate at all coming in. Every retailer’s product needs a unique global ID so AI agents can match it to the same product across the web — pulling in reviews, comparable prices, category placements. No GTIN means your throw blanket …well no knows there IS a throw blanket... And the likes of ChatGPT Shopping won’t ever show anything, even if your that product category is ripe.

Besides, competition is the name of the game. You wont get your Store AI Readiness Audiet and organised — someone else will. The game will carry on. Long story short, I went looking for an app that audited or fixed all three. I expected to find ten of them. And I did find the few apps that took the question seriously, each with a different theory of the answer.

AI Shopping Readiness requires more… inventory optimisation (work)

Here are the features compared from Schema App, TinyIMG and Proofy v2

Shopify Apps for AI Agent Readiness

Each of those has decent (sometimes long) list of features and benefits. If you already KNOW exactly which bit, SEO feature, or GAP your store is missing, the specialist tools win. If you don’t know, or if you suspect the answer is “more than one thing,” and that is the system app that you should go for. Let me walk through what i found and what i tested in each of my candidate apps, as follows.

Schema App

Schema Plus App is the established name in the Shopify schema/JSON-LD space. They’ve been at this for years, mostly servicing larger merchants and enterprise. The product is mature, decent feedback.

Installing that also took a few moments. The dashboard threw me into schema configuration straight away — Product schema, Breadcrumb schema, Review schema, FAQ schema. I picked sensible defaults and let it generate markup across my product pages. The JSON-LD it produces is genuinely good. Comprehensive, valid, the kind of thing Google’s Rich Results Test gives you a green tick for.

The issue is that nothing in the app told me whether ChatGPT, Claude, or Perplexity could actually crawl my store in the first place. I’d asked the JSON-LD question and gotten an answer, but the bot-access question hadn’t been asked. And it didn’t audit my GTIN coverage either so I went and pulled my Shopify product CSV and discovered roughly half my products didn’t have GTINs at all (RedBubble assigns its own internal SKUs; when I imported, GTINs came across as blanks).

Pricing. Aactually more affordable than I expected — Lite at $10.99/mo, up to Premium at $28.99/mo. Importantly the pricing isn’t SKU-scaled, so a 2,000-product catalogue doesn’t cost more than a 50-product one. All tiers include a 14-day free trial, which I appreciated.

The Schema Plus pitch in one sentence: if JSON-LD coverage is your only gap and you want the most thorough structured data generation on the market, this is the specialist tool — but it won’t tell you anything about bot access or product identifier gaps.

TinyIMG

TinyIMG started life as an image optimisation app — compress your product photos so pages load faster. They’ve added an AI angle in the last year, including an LLMs.txt feature that signals to AI crawlers what content on your store is allowed for training/use.

The free tier is generous on the image-optimisation side, which is genuinely useful. I let it batch-process my product images and got a nice page-speed bump. The LLMs.txt feature was a one-click setup that placed the file at the right path. Less straightforward was understanding what it actually does versus the older robots.txt approach. (It's a separate convention from robots.txt, and AI bots vary in whether they respect it. Useful, but not a replacement for getting robots.txt right.)

What I didn’t get from TinyIMG, this JSON-LD coverage report (the markup it does ship is basic and not really their focus), no GTIN audit at all, and nothing on product description quality or fixing missing fields.

Pricing. Not a bad stard, free plan available (50 image optimisations per month), paid from $14/mo (Beginner). The pricing scales by image optimisation volume, not product count — so a large catalogue with few new images isn’t penalised. The LLMs.txt feature requires the $23/mth for Advanced tier though.

The TinyIMG pitch in one sentence: if page speed is your biggest pain and you want a capable SEO bundle that includes LLMs.txt generation as one feature among many, it’s solid — but agent readiness is one checkbox in a broader tool, not the core product.

comparison. tiny bias, but i worked these hard.

Proofly v2

Proofly is the newest of the three (it went live on the App Store in May 2026). Its whole positioning is around the agent-readiness question — the dashboard literally has a thing called the “Agent Readiness ” front and centre. https://proofly-shopify.com/

I installed it and the first thing it did was run an audit. Two minutes later it told me three things I hadn’t known:

• Three of the nine major AI bots it checks were blocked in my robots.txt. I had no idea. (Shopify themes vary, and mine had inherited blocks from somewhere.)
• JSON-LD coverage was 23% across my product pages. I’d assumed Shopify did this for me automatically. It mostly doesn’t.
• 47% of my products had no GTIN. The CSV import from RedBubble had brought across SKUs but no globally unique identifiers, so to a search engine my products were anonymous.

That was a useful 120 seconds or

The fix loop was the part that surprised me. For each finding it would let me preview the fix — see the exact diff, the new JSON-LD block, the new ALT text, the rewritten description — before anything touched my live store. And every applied change went into a “Proof Ledger” so I could roll it back inside 30 days if I changed my mind. Coming from RedBubble — a marketplace where you essentially have no control over what’s published — this kind of safety rail felt like the right amount of cautious.

The free 14-day trial covers 100 products and includes 500 fix credits, which was enough to find out whether the diagnosis was correct before I committed to a subscription. I appreciated not being asked for a card upfront.

Pricing. Now, i will level with you — thats the cheapest of the bunch. $9.99/mo flat (the GROWTH tier covers up to 2,000 products) — starts with a generous 14 day trial with a 100 products test run. For my catalogue size, that’s the friendliest of the three… and i had a lot of fun figuring out the GTN numbers (or there-lack-of) to fix up. — sadly its a paid teir, but worth he money.

The Proofly pitch in one sentence — if you don’t yet know where your gaps are and want one app to scan and fix the whole agent-readiness surface, this is the one obviously designed for that exact job.

The honest verdict (and what I actually picked)

Don’t read this as me crowning a winner — Pick the right tool for the job. Your store may have different products and thus challenges. However, here are some 2 cents i have to offer;

• Pick Schema App if you already know JSON-LD is your only gap (you’ve audited the others), budget isn’t tight, and you want the most thorough markup generation on the market. The specialist wins when you know what you need.
• Pick TinyIMG if image performance is your bigger pain, you want a generous free tier, and you’ll add other tools alongside it for the rest of the agent-readiness picture.
• Pick Proofly if you’re like me — moving from a marketplace, new Shopify domain, not certain where the gaps are — and you want one app that audits the whole surface and ships a fix loop. The 14-day no-card trial is the only way I would have figured out my actual gaps without committing to a subscription.

I went with Proofly — as you may have guessed.. Mostly because the audit told me things I genuinely didn’t know about my own store, and because I’d otherwise have ended up stacking three apps to cover what one was covering. At my catalogue size it was also the cheapest path. Six weeks in, my JSON-LD coverage is at 100%, robots.txt is open to the bots I want, GTINs are sorted on the products that warranted them — and there is a a feature to re-use other (previous catalogue context) for self resolving GTIN — super handy feature!

If you have direct experience with any of these I’d love to hear it. Especially if you found a fourth app I missed — I’m genuinely curious whether anyone else is treating agent readiness as a thing… rather than a another SEO element we must add-onto..

if you’re starting from scratch

Lets consider your options. You will need to describe your store now for Humans Shoppers and AI Shopping assistants. as such you need to also to adhere to new robot-ai shopping assistant structure… Hence, here are my hot tips;

1. Run a robots.txt check for the major AI bots — GPTBot, ClaudeBot, PerplexityBot, Google-Extended at minimum. Any of the three apps does this. Proofly does it free.
2. Sample ten of your product pages and view-source. Search for application/ld+json. If the block is missing or thin, that's your starting gap.
3. Pull your Shopify product CSV and check the GTIN column. If a meaningful chunk is empty, that’s a Google Merchant Center silent-rejection waiting to happen — and AI agents skip what they can’t anchor.
4. Pick one app and run it for a month. Don’t stack overlapping tools (Schema App + Proofly together is wasteful). Find out what’s missing before you spend money to fix it.
5. Don’t try to fix everything in week one. Ship the JSON-LD across your top-50 SKUs first. Open robots.txt. Add GTINs where they exist. The compounding curve on agent visibility takes a few weeks to show up; one big push followed by patience beats grinding daily.

If you’re moving from a marketplace to Shopify like I did, double down on these. You’ve already given up the marketplace’s built-in discovery - you need to rebuild discovery from scratch, and AI agents are now a meaningful slice of that surface.

Dont you worry, I’ll write a follow-up in 60 days with what’s actually changed. If there are specific things you’d like me to test (other apps, other AI surfaces, other product categories) leave a comment.

Will keep you posted on my developments trying to launch my Drop-Ship T-shirt business. JF