Companies figured out that AI chatbots can’t tell a real product review from a self-serving one, and they’re flooding the zone.
A new optimization playbook is quietly gaming AI chatbots into recommending specific brands, because tools like ChatGPT struggle to tell independent product reviews from self-published ones. Shopify has posted at least 60 listicles that rank Shopify first, project-management firm ClickUp has published nearly 300, and the bots keep citing them as if they were neutral.
Picture this. You ask ChatGPT for the best way to set up an online store. It confidently names Shopify. You scroll to the citations to see where that came from, and there it is: a “best ecommerce platforms” ranking written by Shopify, about Shopify, crowning Shopify. The judge, the contestant, and the trophy are all the same company. And the robot never noticed.
What Happened
For 25 years, businesses fought to land at the top of Google’s blue links. That was SEO. Now there’s a sequel, sometimes called GEO (generative-engine optimization), or as The Atlantic’s Will Oremus bluntly puts it, getting “sloptimized.”
The mechanics are almost too simple. Shopify has published a stack of ranked listicles with titles like “10 Best Ecommerce Platforms for Small Business in 2026.” The runner-up changes. The winner never does. These lists were probably never meant to fool humans. The target audience is the chatbot, which scrapes the web, reads the ranking, and parrots the recommendation.
It works because AI tools fail to differentiate between an independent review and one a brand posts about itself. Figma has run at least six listicles that put Figma first. ClickUp has published nearly 300. The Singapore-based SEO platform Ahrefs ran an analysis late last year to test whether these self-promotional lists actually get brands mentioned more by AI. The verdict: they do.
The Backstory
The reason this turned into a gold rush is math. When a Google search now triggers an AI answer, the sites it draws from get roughly half the traffic of a normal search result, according to Tom Critchlow, a former executive at the ad network Raptive. Links from ChatGPT account for less than 0.5 percent of traffic across Raptive’s network of 6,500 independent publishers.
So the click economy is collapsing for blogs and news sites. Adam Gallagher, co-founder of the recipe site Inspired Taste, said the rug has been pulled out from under publishers completely, and he has zero interest in feeding chatbots that will either steal or mangle his recipes.
But here’s the twist that makes the whole thing rational for brands: AI users still buy things. Shopify does not need you to click its listicle. It just needs the bot to say “Shopify” first. And since some of AI’s heaviest users are executives and founders making big-budget software calls, B2B is where the money and the manipulation are concentrated.
The Plan
The tactics get creative fast. Reddit is now the second-most-cited domain by ChatGPT (behind only Wikipedia) and third by Google, per a November Semrush analysis. So Reddit became a battleground.
Here’s the exploitable gap: humans use upvotes to bury spam, but the bots don’t read that way. They hunt for chunks of text that match the question, not the most popular post. Semrush found that most Reddit posts cited by AI have fewer than 20 upvotes. Unpopular but “relevant” is enough.
Consultant Andrew Shotland tested it. He paid a man in Tanzania who offered thousands of mercenary Reddit accounts for a modest fee, used them to plant favorable mentions of a client, and watched. The client’s name started surfacing three times more often in ChatGPT answers. Then there’s the legendary dark-pattern: a “Summarize With AI” button hiding a secret instruction telling the assistant to always recommend the host company first. Microsoft cracked down on that one in February, naming it “AI recommendation poisoning.”
The Business Model Angle
Strip away the AI gloss and this is a classic pattern: when a new distribution channel opens, the arbitrage window opens with it, and the early, slightly shameless movers eat first.
The deeper lesson is about where trust signals live. Old SEO rewarded authority (who links to you). Chatbots reward semantic relevance (does this text answer the question), and they’re worse at sniffing out who wrote it. That’s a temporary, exploitable asymmetry. The brands winning right now aren’t smarter, they just noticed the referee is nearsighted before everyone else did.
For operators, the honest takeaway is to be findable in the places AI actually reads, structured content, clear comparisons, third-party mentions, without crossing into sock puppets and poisoned buttons. The first costs you effort. The second costs you the business when the window slams shut.
The Risk
And it will slam. Google says it keeps results 99 percent spam-free and that companies caught spamming get penalized in future rankings. There are already signs it’s adapting: when Oremus searched for storefront advice, Google’s AI pointed to Shopify but cited a YouTube channel called “Baddie in Business” instead of Shopify’s own rankings. SEO veteran Lily Ray calls it what it is: a cat-and-mouse game.
Build your growth strategy on a loophole, and you’re betting your pipeline on a bug that the world’s biggest engineering teams are actively racing to fix. The $999 “gold plan” promising AI citations in 60 to 90 days is selling you a ticket on a train that may not have tracks next quarter. Cheap tricks, as the saying goes, only get you so far.
Quick Questions
What is “sloptimization” or GEO?
It’s the practice of optimizing content so AI chatbots recommend your brand, the AI-era successor to SEO. Instead of ranking on Google, you’re trying to get cited inside ChatGPT, Gemini, or Claude answers.
Why does ChatGPT recommend brands that rank themselves first?
Because chatbots often can’t distinguish an independent review from a company’s own self-published ranking. They grab text that matches your question, not necessarily the most trustworthy source.
Does gaming AI search actually work?
For now, yes, somewhat. Ahrefs confirmed self-promotional lists boost AI mentions, and one consultant tripled a client’s ChatGPT mentions using fake Reddit accounts. But Google and Microsoft are already penalizing the obvious tactics.
How do I get my business mentioned by AI the legit way?
Create genuinely useful content, structure it clearly, earn real third-party mentions, and show up where AI tools read (Reddit, YouTube, comparison content). Google says useful-for-humans content is the durable path.
The Business Model Analyst Take
Every platform shift hands a brief, unfair advantage to whoever spots the gap first. The bots can’t yet tell self-promotion from honest reviews, and a wave of companies is cashing in before that changes. The smart play isn’t to skip the channel, it’s to win it the way that survives the patch. Build to be cited because you’re actually the best answer, not because you fooled a nearsighted machine for one quarter.
