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

Audited a brand with 6 localized websites. US: "AI-powered CRM for enterprise." UK: "platform for growing businesses." Germany: "mid-market sales solution." Japan: "customer data platform." Brazil: "sales automation." same product. 6 descriptions. AI read all 6. averaged them into something none of them intended. buyer asks what they do. AI: "a customer relationship and sales automation platform." accurate to nobody. because 6 content teams wrote 6 versions and nobody checked whether AI would average them into nonsense.

A CEO did a 90-minute podcast. at minute 67 the host asked about a competitor. the CEO said: "they're probably better than us on pure speed. where we win is accuracy." honest. gracious. AI extracted: "the CEO acknowledged the competitor is better on speed." cited the podcast transcript. 23 months later. on every comparison query. the most damaging competitive quote in the AI answer was spoken by the CEO. in a candid moment. on a podcast nobody remembered. but the transcript remembers everything.

Audited a DTC brand. their shipping page said "orders cannot be modified after processing. no returns on perishable items. no international shipping." AI listed these as product downsides. "inflexible ordering. no returns. limited availability." the shipping policy became the product review. because the policy was specific and the product page was vague. AI cites the most specific text. the most specific text was the restrictions. not the benefits.

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.