15+ years in marketing. Traditional playbook is retired. I run marketing with AI. I also do marketing for AI. Because agents now buy and people ask AI who to trust before they spend a dollar. The game flipped on both ends. This is where I break it down. Let's #connect

A customer success platform was showing up in AI answers as help desk software. For 3 years. I found this during an AI visibility audit. The company had a solid product. Health scoring. Retention. Expansion revenue. Strong market position. But when I searched “best customer success platforms” across AI engines, they were nowhere. Not one engine. Not one query. Then I found them under “best help desk software.” Right next to Zendesk and Freshdesk. Products they don't compete with. In a category they don't belong in. So I traced it back to G2. Their profile was listed under “help desk software.” Set in 2023 by whoever created the profile. They picked “help desk” instead of “customer success.” Nobody caught it. Nobody checked. G2 is one of the sources AI uses to understand and categorize software products. G2 said help desk. AI said help desk. For 3 years. I told them: “You've been invisible in your own category for 3 years because of a dropdown menu on a site you forgot about.” We recategorized them on G2. Updated Capterra. Added schema. Eight weeks later, they started showing up in customer success queries for the first time. One dropdown menu. Three years of misclassification. Go check your G2 category right now. Then Capterra. The category AI puts you in might not be the category you think you belong to. It's the one your data keeps telling it.

The difference between your pitch and AI's pitch. your pitch: crafted. rehearsed. 45 slides. builds a narrative. addresses objections. ends with a CTA. delivered by your best presenter. AI's pitch: assembled from 12 sources in 3 seconds. no narrative arc. no objection handling. no CTA. a flat summary of whatever data was available. delivered to millions simultaneously. your pitch reaches 50 people per week. AI's pitch reaches 50,000. yours is better. AI's is bigger. and the buyer hears AI's version first. before they hear yours. which means your pitch isn't introducing the company anymore. it's correcting the introduction AI already made. the best pitch in the world can't overcome a bad AI introduction at scale. because the AI introduction happens before the pitch gets a chance.

The honest version of your company is the AI-optimized version. this is the part nobody wants to hear. the best AI description isn't the one that makes you sound amazing. it's the one that's accurate. specific. verifiable. because AI doesn't reward superlatives. it rewards claims it can check. "the leading platform" gets ignored. "$49/month for up to 50 users with Salesforce integration and 4.7 stars from 340 reviews" gets cited. the specific version sounds less impressive to a marketer. it sounds more useful to a machine. and the machine is the one writing your introduction to 51% of buyers. stop optimizing your description for how it sounds. optimize it for how it checks. the version that survives fact-checking is the version AI keeps. everything else gets rewritten.

every brand has an AI reputation. most have never read theirs. you have a Google reputation. you check it. a Glassdoor reputation. you check it. a Trustpilot reputation. you check it. a social media reputation. you monitor it daily. your AI reputation reaches more people than all of those combined. 600 million ChatGPT users alone. you've never read it. the reputation that reaches the most people is the one nobody monitors. not because it's hard to check. 90 seconds. incognito. brand name. because it never occurred to anyone that AI had an opinion about them. it does. it's specific. it's public. it's being read right now by someone considering your product. and you've never seen it.

90% of your marketing effort is invisible to AI. the brand campaign. the event sponsorship. the podcast ads. the influencer partnership. the conference booth. the swag. the billboard. the retargeting pixels. the email nurture. the webinar series. AI sees none of it. AI sees: the H1 tag. the meta description. the schema. the product page text. the FAQ. the comparison page. the review schema. the structured data. the 10% of your effort that lives in HTML is the 100% of what AI reads. the other 90% builds human awareness. important. valuable. but invisible to the machine that increasingly decides who gets recommended. you're investing 90% of effort in the channel AI can't see. and 10% in the channel 51% of buyers now use.

A bakery in Paris is monitoring what ChatGPT says about it for €99 a month. a Series C SaaS company with $40M ARR has never checked once. the bakery owner knows what AI tells customers about their croissants. the SaaS CEO doesn't know what AI tells buyers about their platform. €99 a month vs $0 invested. the bakery is ahead. not because they're more sophisticated. because they checked. the bar for starting isn't budget. isn't technology. isn't headcount. it's curiosity. and right now a bakery in Paris has more of it than most tech companies.

Your Black Friday discount could be invisible to AI shopping agents. Shopify just changed that. Shopify Catalog now shares compare-at prices with AI shopping agents. Announced October 9. Live as of this week. Merchant-controlled setting. Before: agents saw one price. The current price. No context. After: agents can see the current price AND the original price. The markdown. The discount. The deal. This matters for Black Friday. Massively. When an AI shopping agent compares 50 products for a buyer, it can now see which ones are on sale and by how much. It can factor discounts into its recommendation. Tell the buyer, “This product is 40% off right now.” Even help make a purchase when the price hits the target the buyer set. The agents that couldn't see your discount last month can see it this month. Every Shopify merchant who turns on compare-at pricing has a chance to make their deals more visible to AI shopping agents. Five weeks before Black Friday. The timing isn't accidental.

Your company is probably grieving its AI visibility problem. And yes, there's a five-stage process. 1. Denial: “AI doesn't matter for our industry.” 2. Anger: “Why the hell is AI saying THAT about us?” 3. Bargaining: “Can we just fix the one thing the CEO saw?” 4. Depression: “There are 47 things wrong, and nobody owns this.” 5. Acceptance: “Give me the spreadsheet. I'll check every Monday.” Every company seems to go through some version of these five stages. The only difference is how long they stay in each one. Some go from denial to acceptance in a single meeting. Others spend 18 months debating whether AI visibility deserves a budget. Meanwhile, AI keeps answering buyer questions. It keeps recommending competitors. It keeps repeating outdated information about your brand. And your buyers keep making decisions based on what they see. The market doesn't wait for your company to accept that AI visibility matters. It just keeps talking about you while you're still in denial. Which stage is your company in right now?

A brand listed 11 features on the product page. the first 3 were visible on load. features 4-11 were behind a "show more" button. JavaScript expand. AI read 3 features. the competitor listed all 9 features in static HTML. no button. AI compared: "competitor offers 9 features. [brand] offers 3." the brand had more features. AI counted fewer. because 8 were behind a click. every "show more" button on your site is a wall between AI and the content behind it. count what's behind the click. that's how much is invisible.

I've shared dozens of audit findings. client stories. fixes. patterns. all free. all public. now I want to learn from you. who's doing AI visibility work right now? what tools are you using? what's your weekly process? what was the weirdest thing you found? what fix produced the biggest change? reply with one thing. one finding. one fix. one screenshot. one process. whatever you have. I'll engage with every single reply. this thread should be the largest collection of real AI visibility practice on the internet. help me build it.

“AI only notices big brands” is one of the most expensive myths in startup marketing. AI doesn't know how many employees you have. It doesn't care how much funding you've raised. And it doesn't automatically recommend the biggest company in the category. It needs information it can actually understand, access, and trust. A 5-person startup with clean schema, a useful comparison page, and structured FAQs can earn mentions in AI answers. A 500-person company with a JavaScript-heavy website, missing product information, and no structured data can struggle to get mentioned. The difference isn't headcount. It's how clearly each brand communicates what it does, who it's for, and why it's worth recommending. That's the opportunity. A smaller brand can compete for AI visibility without matching a larger competitor's marketing budget. But it still needs the right information, credible sources, and content that answers real buyer questions. “We're too small” isn't a strategy. It's a reason not to start. Your competitors' size isn't what makes them visible to AI. So what is your website making harder for AI to understand?

A company's support team was documenting AI's mistakes three times a month. Nobody realized it. In Slack, the same complaints kept appearing: "Another customer thinks we have a feature we don't offer." "Customer expected the price to be lower." "Prospect compared us to a competitor we've never heard of." Support treated each one as an isolated incident. Customer confusion. Miscommunication. Bad expectations. But nobody asked the obvious question: Where did all these customers get the same wrong information? We traced the pattern back to AI. Different customers. Same AI answer. Same outdated source. The answer described a feature the company didn't have, quoted the wrong price, and positioned the product alongside an unexpected competitor. AI was setting expectations before the sales team ever spoke to the buyer. And the evidence had been sitting in the company's support Slack for months. The fix started with one question added to the support intake form: "Where did you first learn about us?" We added "AI/ChatGPT" as a trackable source. Within a month, the pattern became visible. The wrong expectations mapped back to the AI answer. We corrected the information feeding that answer. The following month, wrong-expectation tickets dropped by 60%. The lesson isn't just about AI visibility. It's about where you look for the problem. Your support team might already be documenting what AI gets wrong about your company. It's in the tickets. The sales call notes. The Slack messages nobody connects. Before you spend more money correcting customer expectations, find out who's creating them in the first place.

There's a 60-second test that separates companies that act on AI visibility from companies that just talk about it. The clock starts when you first learn about the problem. Do you open an incognito tab? Or do you save the post for later? I've noticed a pattern. The people who check immediately tend to find problems faster. They spot the missing price. The wrong product description. The competitor AI keeps recommending instead. Then they fix something. The people who bookmark the post and plan to investigate later often never get around to it. Not because they don't care. Because learning about a problem feels productive. Actually confronting it is uncomfortable. And that gap between learning and doing is where momentum dies. You don't need a six-month AI visibility roadmap to start. You need 60 seconds. Open ChatGPT. Search for your brand alongside your biggest competitor. Read the answer. Is your brand mentioned? Are the facts accurate? Which competitor gets recommended instead? Now you have something real to work on. You can build the roadmap after you've seen the problem. Stop collecting advice about AI visibility. Go find out what AI actually says about your brand. The first check takes minutes. Putting it off can cost you months.

the Monday check that pays for the whole week. 90 seconds. incognito. brand name. read. compare to last Monday. anything change? if yes: investigate. if no: move on. the 90 seconds isn't productive because of what you find most weeks. most weeks nothing changed. it's productive because of the one Monday where something did change. a competitor appeared. a feature disappeared. a price changed. a comparison shifted. that one Monday saves a month of damage. because you caught it on Monday instead of discovering it 4 weeks later when a prospect mentioned it on a sales call. 51 uneventful Mondays and 1 that saved a deal. the 90 seconds was worth it every week. because you never know which Monday is the important one until you check.

Nobody knows what this discipline will be called in 5 years. AEO. GEO. AI visibility. AI search marketing. agentic readiness. brand intelligence. machine perception management. the names are still competing. the work is the same regardless of the name: check what AI says about you. fix what's wrong. keep checking. the brands that wait for the name to settle before starting the work are losing to the brands that started the work before the name was decided. SEO didn't have a stable name in 2002 either. the people doing the work in 2002 didn't care what it was called. they cared that it worked. it's 2002 again. the name doesn't matter. the check does. start the work. the name will catch up.

the brand that stopped creating content and started curating data. they published 3 blog posts a week for 2 years. 300+ articles. solid content. good traffic. AI descriptions: wrong. they stopped publishing new content for 6 weeks. instead: updated schema on every product page. restructured 50 old blog posts to match current positioning. added FAQ pages matching real buyer queries. published 2 comparison pages. deleted 80 outdated posts. redirected 40 stale URLs. created zero new content. AI description improved more in those 6 weeks than in the previous 2 years of publishing. because AI doesn't reward volume. it rewards accuracy. and accuracy comes from curation. not creation. sometimes the best content strategy is to stop creating and start fixing what you already have.

I've shared dozens of specific things AI gets wrong about brands. now I want to hear yours. what's one thing you found that AI gets wrong about your company that you traced back to a specific source? the specific source is the key. not just "AI says X wrong." but "AI says X wrong because of Y page." the story with the source is 10x more useful than the complaint without it. drop yours below.

For 11 months, a competitor was writing a brand's product description for every buyer asking AI, "Which is better?" Here's how. The competitor published a page titled: "[competitor] vs [brand]: the honest comparison." They chose the features. They chose the benchmarks. They chose the framing. Their strengths got highlighted. Their weaknesses got buried. The brand's weaknesses got a whole section. And the page was good. Clean HTML. Structured comparison table. Specific claims. Easy for AI to read. The brand had no comparison page at all. Its product page described its own product. The competitor's page described both. So when a buyer asked AI, "[brand] vs [competitor]," the competitor's page became the obvious source for the answer. AI used the competitor's benchmarks. The competitor's framing. The competitor's version of the brand. Not necessarily because the comparison was fair. Because it was the most detailed comparison available. For 11 months, the brand was effectively absent from its own competitive narrative. The fix took an afternoon. We published: "[brand] vs [competitor]: features, pricing, and performance compared." A structured comparison. Specific benchmarks. Clear trade-offs. Honest about where each product wins. Now AI had another source to work with. The competitor no longer had the only detailed comparison in the conversation. The brand's perspective finally existed in a format AI could use. Eleven months late. But live. Here's the uncomfortable part: If your competitor has a comparison page about you and you don't have one about them, you're letting them define the terms of the comparison. Every time a buyer asks AI "Which is better?", their version is available. Yours isn't. You don't control what your competitors say about you. But you can stop leaving them as the only source explaining why someone should choose them.

A company's customer success emails were convincing AI that its product was hard to use. The emails were doing exactly what they were supposed to do. After every onboarding, customers received a friendly message: "We know the first few weeks can be challenging as you learn the platform. Don't hesitate to reach out if you hit any bumps." Empathetic. Helpful. Completely standard customer success communication. But customers forwarded the emails. Some quoted them in forum posts. Others pasted them into Reddit threads about the product. The wording spread across public, crawlable pages. AI found the same sentence on 14 different pages. Then synthesized those mentions into a product assessment: "Users report that the platform has a challenging learning curve during the first few weeks." The CS team meant: we're here to help. AI read: confirmed learning curve. The empathy became evidence of a product weakness. And the strange part? Nobody had complained about the email. Nobody had written a negative review. The brand's own attempt to reassure customers had helped create the negative framing. Here's what we changed. We rewrote the onboarding email: "Most customers are fully up and running within 5 days. Our team is here to make sure you're one of them." Same warmth. More specific. A clear outcome instead of an assumption that getting started will be difficult. The old version acknowledged the struggle. The new version set an expectation of success. Both were supportive. Only one gave AI an obvious phrase to associate with a learning curve. Check your templated customer emails. Look for phrases like: "We know it can be tough." "Don't worry if it takes time." "The first few weeks can be challenging." Your customers might quote those words publicly. AI might later reuse them without the reassurance that surrounded them. Your support team is shaping customer experience. Its words may also be shaping how AI describes your product.