Copy of a Copy: Why Weak Product Descriptions Make Your Shopify Store Invisible to AI

Steve Hutt: Welcome back to eCommerce Fastlane. I’m your host, Steve Hutt. Today’s conversation is all about keeping your product pages discoverable in Google and AI search, long after launch day. Let me set the stage. Let’s say you have hundreds of SKUs, and some of them use supplier descriptions. Maybe you’re selling a commodity product, and every other store is publishing exactly the same information from the supplier. There are also private-label products, which we’ll talk about separately. My guest today is extremely knowledgeable. Jim brings a depth of experience in organic SEO going back to the early 2000s. He worked under Bruce Clay, who’s an OG in this space. He’s now the co-founder and CEO of a company called Studio 1119.ai. I’m not sure about the pronunciation, but we’ll get to that in a minute. His team is behind Cataseo, a Shopify app that also connects to Wix and a few other content management systems. He’s been doing hands-on ecommerce SEO since the early 2000s, possibly the late 1990s, writing product content and filling in metadata by hand. For the last 15 years, he’s been helping major brands. Now, with AI, his solution offers an opportunity to refresh your product information so it’s discoverable in Google and AI assistants. This is an incredibly important topic as we head toward BFCM. There’s still time to organize your product catalog so you can start appearing in AI search. Jim, that was a long preamble, but thank you so much for coming on the show today.

Jim Boudreau: Thanks, Steve. I’m really glad to be joining you and looking forward to the conversation.

Steve Hutt: Absolutely. Let’s dig into the early days. It’s nice to set the stage: Who is Jim? Where did you get started? Help us understand the background leading up to today’s conversation.

Jim Boudreau: It goes back to 1998, when I launched a clicks-and-mortar business called ladiesgolf.com with my then-wife. These were the days of AltaVista and Yahoo, and being found wasn’t even really thought of. They hadn’t really coined the term SEO yet. You had to figure out on the fly how to get noticed, and you did it by being very complete in your writing. Whatever data field was available, whether metadata or tags, you filled it. You tried to put as much on the customer-facing page as you could. Fast-forward to the early 2000s, when I first met Bruce Clay and trained under him for a while. It became more scientific. Many people know Bruce coined the term SEO, and then Google was born. What had been a black-box art started to become more of a science. You have to stay on top of it. If you wanted to be found then, or you want to be found today, you have to do the hard work. That hasn’t changed in more than 25 years.

Steve Hutt: I totally agree. One thing I’ve learned lately is that Google’s AI Overviews are drawing from its own top results. There’s still merit in traditional organic SEO, with a decent, high-authority backlink profile, proper structured data, and time spent establishing trustworthiness. There’s an interesting acronym in the SEO world around how to be found. I still feel there are traditional SEO benefits, at least from Google’s perspective. But I’ve also learned that other large language models, whether Claude, Perplexity, or ChatGPT, don’t simply scrape Google’s results. They do their own training and choose how to parse data and what to share. If you’re researching a product, they have their own data and make their own decisions. That’s why you’re on today. It’s not just traditional SEO, Google AI Overviews, and what Gemini is doing. ChatGPT and Claude are very popular, dedicated solutions people use to research products, independent of Google. What matters is how the product information is written and then parsed into their data. I think you have an early story about how you got involved with this. I skimmed LinkedIn this morning, and I believe it involved Scandinavian Shop. Can you give us a high-level overview of the hundreds of hours of hands-on work you did to get that shop to a certain level? How does that contrast with what’s happening today through AI and your solution?

Jim Boudreau: There are a few interwoven stories here, and a lot of it comes down to what’s old being new again. Think about Google’s original intent. If you go way back, search engines such as AltaVista were trying to list websites available on a topic, when the web wasn’t so omnipresent. Google took a different approach: If you’re typing something into Google, you’re trying to solve a problem. Even today, it starts trying to finish your sentence because it’s trying to understand the problem you’re solving. That hasn’t changed, even with AI. The rules around how you get mentioned are another age-old conundrum. Before Scandinavian Shop, I had a website selling styluses. This was the era of commercial handhelds, when you’d sign a little handheld terminal for FedEx. We sold hundreds of thousands of little plastic sticks. At the time, I had one competitor with the worst-looking website and worst content I’d ever seen, but I couldn’t beat him in Google. He’d be first, and I’d be second. He’d be second, and I’d be third. Part of it was that he’d become entrenched. That’s proven true over the years in SEO: Once you’re the entrenched competitor, it’s hard to knock you off. But AI search is a greenfield opportunity because all the LLMs are taking a fresh look at everything. One lesson I learned early was that original content is king. Bruce used to espouse that original, thorough, problem-solving content is king. I learned to write that way, and we were pretty successful. When my friends started Scandinavian Shop 15 years ago, I made them go through the pain of writing original content, filling in every bit of metadata, and tagging every image. At the time, we were even putting metadata into the JPEG file headers. It was meticulous work, but it paid off over the years. They’ve averaged a couple of thousand ranked keywords in a nice niche, Scandinavian-themed gifts, and typically 300 to 500 top-10 rankings. That’s kept their lifetime pay-per-click budget to about 3%. That’s a huge benefit of investing in AI or search: Your pay-per-click budget goes down, and your credibility goes up.

Steve Hutt: Absolutely. Even when I was at Shopify, I always talked about organic search visibility as the highest-ROI marketing strategy you can work on. There’s that entrenched position you mentioned, but you still have to do the work and be found in organic search. With Google, that can also mean appearing in AI Overviews and being part of the conversation as Google tries to understand and solve someone’s problem. The next question becomes interesting.

Jim Boudreau: It really does. AI tools look at it in a very structured way, no pun intended. They need the most specific, complete answer to a question. The content that ranks highly in Google today isn’t necessarily the same content that’s going to rank highly in AI. You might do an amazing job of romancing your product, but if you’re not answering questions in a structured manner, both in your customer-facing content and your back-end data, you’re just as invisible.

Steve Hutt: There’s E-A-T, and then they added another E for experience: experience, expertise, authoritativeness, and trustworthiness. People were trying to figure out whether Google ranked certain brands highly because they had experience or expertise in what they discussed or sold, and because they were authoritative and trustworthy. Does that play a role today in how AI and large language models decide which brands to rank in search?

Jim Boudreau: I think it’s still very early. In the things we’ve tested, take a market such as Scandinavian-themed gifts. There are a few sites that have been around forever. They have the best assortments and products, and they’re the best in their category. But if you ask most LLMs for a particular Scandinavian-themed gift, none of them get mentioned. Very obscure sites get mentioned instead. Trying to connect the dots has been interesting. That’s what Bruce and others did in the early days of SEO: They tried to reverse-engineer Google’s algorithm to figure out how to leverage it. So far, nothing jumps out and says, “This is the formula.” Any SEO expert will tell you that formula is still being written. It goes back to what I mean when I say what’s old is new: fundamentals. In traditional search and AI search, the fundamental requirement has always been to do the best possible job describing the product. That’s your customer-facing description. To make it more structured so Google understands what you’re talking about, you wrap it with metadata. Now you’re going one step further and adding another layer: structured data. It’s like the old joke about cassette tapes. Remember making a copy of a copy of a copy?

Steve Hutt: Yeah.

Jim Boudreau: If your fundamental description isn’t sound, your metadata will be less sound, and your structured data will be less sound. The quality diminishes very quickly. That’s going to be the challenge for most merchants: How do they freshen up their catalogs in a way that’s friendlier to consumers, Google, and the LLMs?

Steve Hutt: Let’s talk about Shopify, and then we’ll get into your app because you do something quite unique compared with what’s natively available and what competitors are attempting to build. Let’s start with Shopify Catalog. I think a lot of people don’t realize it’s turned on by default in their admin. Can you explain what Shopify Catalog is for those who don’t know? We can discuss it because I’ve had some interesting findings, too.

Jim Boudreau: You’re probably more of an expert on Shopify Catalog itself than I am. One thing I’ve found in 40 years of launching companies is that it’s easy to get twisted around when you’re living in the competitive context instead of asking, “What’s the right thing to do?” and building the best mousetrap available. Everybody on every platform, not just ecommerce, is trying to give you AI tools to help create original content. That still doesn’t mean it’s rankable content.

Steve Hutt: Right.

Jim Boudreau: That’s the distinction with what Shopify, Wix, and many platforms are doing. They’re trying to give you an assist and help you create original content, but that doesn’t mean it’s going to rank. It’s a step in the right direction, but there’s much more to it.

Steve Hutt: Shopify Catalog is a moving target, but the catch is that it’s really a pipe. It moves whatever product data you already have. With this whole agentic storefront approach, I don’t believe it’s necessarily moving anyone toward meaningful AI traffic. Part of the reason is that their product data isn’t structured for machine consumption. The data is whatever you already have in your store. There’s Shopify Magic, which helps you write copy, but it’s one product at a time. You provide a product title and a few keywords, and it writes a description. It’s limited in what it can do. I don’t really call it magic because it’s very surgical and works one product at a time. To add context, Sidekick is more of a persistent AI agent for things such as theme edits, changes, and A/B testing. Catalog is the pipe, while Shopify Magic works one product at a time. It doesn’t provide much insight into why it’s writing what it’s writing. That’s the contrast with what you’re building. What happens when someone adds the Cataseo app to their admin? Can you walk us through what happens in the background first?

Jim Boudreau: The first thing it does is establish your goals and preferences. The most critical part of creating rankable content is understanding the context of your store. What market are you in? What tone of voice do you want? Do you want a casual style, a lighthearted style, or a technical style? What’s your brand voice? You can tell it to write like a pirate if you want. However you want to create your voice, it’ll take those inputs. Then it looks at how you want to present the information. Do you want pure narrative? Features and specifications? Bullet points? It’s trying to understand how you would write it yourself. It also looks at your optimization priorities. Do you want to optimize your biggest sellers or your weakest sellers? Products with the highest inventory or the lowest? There are 14 different preferences for optimization priority. It’s getting all of that set up. You can use traditional menus and a setup wizard if you prefer. We also have an MCP chatbot called Katie that you can have a conversation with, and it’ll onboard you. It establishes those preferences, which you can change later, then loads your catalog and starts preparing your store for optimization.

Steve Hutt: I loaded it into my demo store, and what stood out was how it closes gaps at scale. Shopify Magic writes from your title and takes you through that journey one product at a time. Your tool looks at who’s ranking in the top 10 for each product. I’m not sure which database you’re using, whether it’s Moz or something else, but you’re looking at who’s already ranking and making sure the content is the best it can be based on competitors and current rankings. I’d call that competitive grounding, and I think it’s unique to your solution. Then there’s the scale. You can batch 50 or 100 SKUs together. From what I could see, you get a side-by-side review and an approval step. It’s not a black box that just goes off and makes changes. It can handle these updates at scale, which is pretty amazing.

Jim Boudreau: At the end of the day, this is about getting ranked and making your store visible. Without competitive context, you might have the most beautiful description in the world, but no one’s going to find you. You have to be better than your competitors. We use Moz, Google’s tools, and a couple of other tools to identify the 10 highest-ranking pages for a given product. If you’re the only one with that product, we look at similar alternatives. What are the 10 closest products available? How do those people talk about the product, and how do we make sure you’re a hundred times better? We also weave in FAQs. If you’re using bullet points and features, which we suggest, those FAQs become customer-facing, and that’s helpful. If you look at what Shopify does with structured data, it’s going to work from what you’ve provided in your main product description. That core description is critical. Taking the title and building down goes back to making a copy of a copy of a copy.

Steve Hutt: Yeah.

Jim Boudreau: Right?

Steve Hutt: Yeah.

Jim Boudreau: The core description you write for each product has to be rock solid. Otherwise, everything is derivative, and it gets weaker as it filters into the layers that Google, LLMs, and agents are reading.

Steve Hutt: That’s another benefit of your solution. Shopify Catalog syndicates your product data, however you created it. You could have used Magic, written it by hand, or used your solution. That data is now going out to ChatGPT and Copilot by default, which means people can shop and buy within these chat solutions, provided you have the appropriate data and they’re recommending you. That’s the challenge: gaining visibility within these tools so people can buy there. That’s where your tool fits in, by moving beyond generic information. Can you talk about that?

Jim Boudreau: Absolutely. If you’ve been in ecommerce for a while, you’ve received a hundred emails saying, “For $500 a month, we’ll get you on page one of Google.” It’s the same playbook: Create landing pages and blog posts. For the first six months, you get a blip, and then it fades. Google got smart and said, “You have this beautiful content page, but it’s leading me to the same unoriginal product description everyone else has.” You could approach it that way in the old days. Today, the tail wags the dog. You need the core description, and from that, you need structured data. If you don’t have that foundation from the start, all the custom data fields and metafields will be nebulous and won’t do anything for you.

Steve Hutt: Another question, more of a side question: A lot of brands listening have a multichannel approach. They’re not just on Shopify. They might have an Etsy shop or be on Amazon because its organic reach is enormous. Some don’t put their full catalog on Amazon. They use it as a brand play, generating sales while trying to bring people back to their main website. What are your thoughts on Amazon? Is Cataseo thinking about organic search within Amazon and its product descriptions? That’s another growing part of a brand’s multichannel strategy.

Jim Boudreau: You bring up an interesting topic. Going back to the early days of Scandinavian Shop and my experience selling styluses, sometimes you have an A-tier keyword where a competitor is entrenched. For Scandinavian Shop, there’s a hand-carved wooden horse that’s ubiquitous in Swedish households called a Dala horse. One competitor owned the domain for that term and covered anything and everything related to Dala horses. They weren’t going to beat them in a frontal assault. They were a new player entering a mature market 15 years ago. Instead, we picked a dozen B-tier keywords that were just as relevant to Scandinavian heritage but didn’t have a dominant player. If you searched them in Google, the results weren’t consistent. They were a mishmash. We saw that as a greenfield opportunity and made sure we dominated those B-tier keywords. Collectively, that dominance helped them overtake competitors on the A-tier keywords. Going back to E-A-T, they became established across a broader range and did an end run around the competition. I call it this: If you can’t change the planet, change the universe.

Steve Hutt: Yeah.

Jim Boudreau: If you’re a brand competing against other brands, you have an opportunity with your Amazon brand store and your other stores. You’re writing about your products in your voice and your way in different places, reshaping Google’s and the LLMs’ view of those products. As a rough example, imagine three competitors with three pages about product X, perhaps a TV. But you have 10 pages: blog posts, a landing page, a category page, and different but similar product descriptions on Amazon, Shopify, and Etsy. Suddenly, you’re repainting the universe in your image and changing the competitive landscape around you.

Steve Hutt: It’s amazing.

Jim Boudreau: It’s an old SEO trick.

Steve Hutt: I love it. I’m all in on that. I also think about Amazon’s halo effect. Even though Amazon takes merchant fees and you want to bring people back to your own audience and Shopify store, there’s still a benefit when people search organically or through chat solutions. If Amazon appears, that doesn’t necessarily mean they’ll buy there. It can be a trust signal when Amazon and these chat solutions surface your product. Someone might look to check the price, then go back to the mothership, the main brand, and have a look. That’s often how I think the workflow works. Amazon gets a large share of ecommerce revenue, though I don’t know the exact percentage. This halo effect has benefits, but the question is whether you can appear organically within these chat solutions. The answer is yes.

Jim Boudreau: One interesting angle is figuring out how to work the system in your favor. You can’t promote your own store on Amazon, but your brand might also happen to be your domain name. Scandinavian Shop is the brand name of my friends’ Amazon store. If you Google Scandinavian Shop, you end up at their store. How do you make it easy to be found outside Amazon and benefit from that halo effect? If it takes too many clicks or too much hunting, people will give up and buy on Amazon. That’s another important thing: Have your own brand on Amazon. There are all sorts of games people play. You have overseas competitors who hijack UPC codes. I’ve seen it happen to people who bought their own UPC codes and spent money to preserve them for unique products, only to have them hijacked by bad actors overseas. The way to protect yourself on Amazon is to have a brand, have that brand own the UPCs, and have the brand control those products. My long-winded answer is to gain control and keep control of your destiny however you can.

Steve Hutt: Absolutely. I did some research on your framework and how Cataseo works. You have a strong opinion about visibility optimization, and I love strong opinions. There’s SEO, GEO for generative engine optimization, AEO for answer engine optimization, social search, and agentic shopping. You call these five surfaces a fragmented visibility stack. Your framing is different. Instead of treating search as fragmented, you’ve developed a solution to that fragmented visibility problem. Can you walk us through what’s fragmented and what you believe the solution is?

Jim Boudreau: I always try to look at it from the business owner’s perspective.

Steve Hutt: Okay.

Jim Boudreau: If you’re running an ecommerce store and start hearing this alphabet soup of acronyms, your eyes glaze over. Most operators have finally accepted SEO as a painful but necessary reality in their business, and that took a decade.

Steve Hutt: Right.

Jim Boudreau: They’ve been sold a bill of goods by all kinds of providers. If they have visibility in search, it’s because they scratched, clawed, and fought to get there. When you throw half a dozen new acronyms at them and say they’re all part of this black box called AI, their eyes glaze over. They think, “Now I don’t have to learn one thing; I have to learn five things? Are you kidding me?”

Steve Hutt: Yeah.

Jim Boudreau: The idea behind visibility optimization is to simplify it. At the end of the day, you have one goal: Be as visible as possible on as many platforms as possible. Few SEO experts I know would argue with getting the foundational 80% right first. Get the description, metadata, and structured data right. Include FAQs in your descriptions. Make sure your image alt text isn’t just a copy of the product name but actually describes the image. Doing that 80% puts you ahead of 95% of your competition because it’s work most operators don’t have time to do. If you’re an enterprise-level brand or an ecommerce operator doing $50 million and up, you have teams that can do it. But if you’re among the millions running a $500,000 or $2 million store, you don’t have time to learn a new science and rebuild hundreds or thousands of items, much less categories and brands. We haven’t even talked about those yet. It’s hierarchical. Your category descriptions have to represent the products, and the same goes for brands. That’s what we’ve tried to do with Cataseo: Make it all-encompassing. It handles products, categories, brands, and blog posts. We want to make your site visible and appealing to customers when they arrive, as well as appealing to LLMs and Google. You’re the site that has the answers. That’s what visibility is about.

Steve Hutt: I appreciate that. You’ve framed it perfectly. Getting that foundational 80% right, including structured data, is a high-leverage technical move and a smart business strategy. Your solution does that at scale and sets people up for success. The extra 20% opens a can of worms around whether your Shopify store itself is AI-friendly, including your robots.txt file. I wouldn’t be surprised if Shopify’s defaults affect GPTBot, search bots, ClaudeBot, and PerplexityBot. I’m going to look into that and put it in the show notes. With bots scraping so much, I suspect platforms are either blocking some access or reducing how much they can crawl because it taxes the network. They may introduce delays. Have you heard much about those layers, along with LLM information files and llms.txt? People are trying to share information with large language models about who they are and whom they serve, but there’s also the robots.txt file. Is your store friendly to these bots?

Jim Boudreau: You bring up a good point. Everybody is throttling access. Some are limiting requests to every 10 seconds. With a big catalog, that can mean weeks or months before an AI engine fully crawls your site. They’re also being selective. Different ecommerce platforms throttle at different levels, or not at all, depending on who’s accessing them. One might throttle Google, while another might not. We haven’t seen consistency. One thing we generally recommend to customers is continuous improvement. We’ve tested this: If you optimize your entire store and publish your entire catalog, you get a nice blip for about 90 days. Then, 90 days after that, you’re back to your baseline. You need to keep refreshing. It’s not all or nothing. We’ve found that refreshing 10% of your catalog every month, continually over time, works. Ironically, your timing can coincide with crawler limits. You’re doing something more productive and pragmatic for traditional search while getting into a cadence with the LLMs. I wouldn’t necessarily call it a cat-and-mouse game, but there’s a push and pull. If you’re Shopify, Wix, or anyone running a platform, you have to protect yourself during this giant AI land grab, with all the tools out there. Throttling is one way they’re doing that.

Steve Hutt: Let’s talk about next steps for listeners wondering what to do. They go to the Shopify App Store and search for Cataseo. I was spelling that C-A-T-A-S-I-O.

Jim Boudreau: S-E-O.

Steve Hutt: Yes, S-E-O. Sorry, I apologize.

Jim Boudreau: All good.

Steve Hutt: They install it, the ingestion happens, and they go through the onboarding you mentioned. How is the tool priced? Is it ongoing or a one-off? What should someone expect, and how quickly can they get up and running?

Jim Boudreau: They can be up and running in 10 minutes. The pricing is straightforward, and you can operate in one of two models. If you want a monthly batch that handles a set number of products, $25 a month covers 50 products. It’s basically 50 cents a product for however many you want to do in a batch. You can also buy express tokens for one-off work on a category, brand, or blog post. You get 25 free to start with the trial, and a few more if you’re one of Steve’s listeners. You can buy express tokens in batches of 10 and work on things ad hoc. Depending on your scale, one thing we love to do is get to know our customers. I love doing white-glove onboarding: getting on a quick Zoom, walking through setup, and helping people understand best practices. We can also analyze their site’s SEO and establish a baseline. We track their metrics and have an analytics toolkit that presents ongoing metrics for AI and traditional search visibility. At the end of the day, we want to help our merchants be more successful. We like to lock arms and do that together.

Steve Hutt: I love that. Having a business partner and white-glove onboarding is a great feature. The timing is good, too. As this episode is released, we’re still a few weeks away from code freezes, including Shopify’s. There’s time to install this and set yourself up for success by getting that foundational 80% in place. Shopify Catalog is still pushing the data out to places such as Gemini and ChatGPT. You’re setting yourself up for success, and it’s affordable based on the size of your catalog. I’ll put everything in the show notes. What are your final takeaways and next steps for listeners?

Jim Boudreau: This is the first greenfield opportunity in a long time. In traditional search, the winners in most categories have been established. You know who occupies the top 10, and you can fight your way in. But in AI search, that’s no longer the case. If you’ve had a hard time overtaking your competition or gaining ground in traditional search, AI search gives you an empty whiteboard. It’s a good time for any ecommerce operator to reset their competitive posture.

Steve Hutt: Even for more advanced users, we didn’t get into the MCP you have through ChatGPT or Claude. For a lot of people, those tools are their source of truth, and now they can query your solution there. We mentioned Katie, but it’s interesting to have that available. If one of those tools is your workhorse, connecting through MCP comes at a relatively small monthly cost compared with the value. You can query it in natural language, have a source of truth, and set up monthly refreshes. It’s really cool that you’ve made this available.

Jim Boudreau: Thank you. It’s certainly been a labor of love. I have a lot of friends in ecommerce, and I’ve always wanted to see them win. When we developed the proof of concept two years ago, we saw huge leaps were possible just by filling in the blanks and doing a better job with core content. We saw 25% ranking improvements in six weeks, which is unheard of in SEO. When we saw the potential, we knew we had to commercialize it and get it out to people.

Steve Hutt: That’s amazing. It’s cataseo.ai, and it’ll be in the show notes. There’s also a companion Shopify app. Jim, thank you so much for coming on the show. This was a little longer than the episodes I typically record, but it’s an incredibly important time for people to get their structured data in place so large language models and Google can understand what they’re all about. You’re setting yourself up for success in Q4 and into next year. I appreciate that you’re building this tool and making it available in the Shopify ecosystem. That’s why I wanted you on today. Kudos to you and the team, and thank you for recording.

Jim Boudreau: Thank you, Steve. Best wishes for continued success to you, eCommerce Fastlane, and all your listeners.

Steve Hutt: Lovely. Have yourself a great afternoon.

Jim Boudreau: You too.