Walmart Sparky users spend 35% more. whether it's friction reduction, better discovery, or higher-intent buyers self-selecting... the outcome is the same. buyers who let AI help them shop spend more. the cause is debatable. the 35% isn't.

the question used to be "will AI change commerce?" then "how fast?" now it's "what did we miss while we were deciding whether to start?" the answer is always the same: time. the one resource you can't recover. every week of debate was a week of compounding someone else earned.

access looks clean... all the major AI bots getting through. the one to watch is OpenAI MCP client at the bottom. that's not a crawler... that's an agent trying to DO stuff on your site 👀

Shopify AI traffic grew 8x. orders grew 13x. the traffic is converting better than human traffic. and most brands still haven't checked if they show up in the answer that's sending it.

AI visibility work has made me more obsessed with simplicity, but LESS patient with marketing that sounds good and says nothing.

the entire AI visibility discipline in 11 words: make it easy for a machine to recommend you accurately.

The two most useful AI visibility tools cost $0. One shows you what you're sending AI. The other shows you what AI is telling your buyers. 1. Right-click → View Source This shows you what an AI crawler can actually read on your page. Is the price in the HTML? Is the schema correct? Does the first sentence explain what you sell? Or does it just say “Welcome”? That's the input check. What are you feeding the machine? 2. Open ChatGPT in incognito Type your brand name like a stranger would. Then read the answer. What does AI think you sell? Who does it think you're for? What does it get wrong? That's the output check. What is the machine telling your buyers? One shows you what's going in. The other shows you what's coming out. Most brands haven't done either. Not because they're difficult. Because nobody on the marketing team thought to look at the source code. And nobody thought to log out before checking ChatGPT. Two checks. $0 cost. Start there. Everything else is refinement.

perfectly said. the mirror was always there. AI just removed the option to ignore it.

uncomfortable and necessary. the words AI uses are the words the internet taught it. if you don't like them, the fix isn't arguing with AI..... it's changing what the internet says about you.

"the API itself is becoming content" ....that's the next layer and almost nobody's ready for it. product pages talk to AI today. APIs talk to agents tomorrow. the brands building both at the same time are ahead of everyone building one at a time.

yep. the screenshot is the symptom. the source is the fix. most brands react to what AI said instead of tracing WHY it said it. find the page, fix the page, the mirror changes.

that's the nuance nobody talks about. AI shows you clearly but it weights whatever evidence it found first. clear and fair are two different things. your job is making sure the evidence it finds is actually representative, not just whatever showed up first.

exactly. most brands manage the narrative they control. AI shows you the narrative you don't. the feedback loop between what the internet says and what AI repeats was always running.... most brands just never had to look at it before.

There are only 3 types of brands in AI right now. And most brands have no idea which one they are. 1. Invisible AI doesn't mention you. Buyers asking questions in your category never see your name. The fix is straightforward: Add the data. Get crawlable. Get seen. 2. Visible but wrong AI mentions you. But it gets something wrong. Wrong price. Wrong features. Wrong positioning. This is more dangerous than being invisible. Because the buyer sees the answer, trusts it, and acts on it. The fix: Correct the source. 3. Visible and accurate AI describes you correctly. Right product. Right audience. Right price. This is the rarest category. And even here, you're not done. Because accuracy today doesn't guarantee accuracy next month. Sources change. Competitors change. Products change. AI changes. So you have to keep checking. Most brands think they're in category 3. Most are actually in category 2. And they don't know. Because they've never checked what AI says about them.

This B2B SaaS company spent serious money building a website AI couldn't read. Beautiful design. Clean layout. Clear messaging. Compelling product thesis. Feature pages. Use cases. Team section. A great website. Built on a JavaScript rendering framework. No Product schema. No Organization schema. No FAQ markup. No structured data. The site rendered almost entirely client-side. So when an AI crawler arrived, the initial HTML was basically a shell: Script references. And nothing else. We tested 44 non-branded queries across three AI engines. Zero citations from the company's own domain. Not because the content was weak. Because the content was unreachable. The crawler arrives. Sees an empty container. Moves on. We saw something similar in a consumer ecommerce audit. A product page exhausted the AI crawler's token budget on navigation menus before it reached the actual product facts. That brand at least had partial schema, some indexable text, and a crawlable catalog. This one had none of it. The website existed for humans. For AI, it was an empty container with a loading spinner that never resolved. We've seen versions of this across a lot of companies we've audited in the last six months. Beautiful modern websites. Impressive when a human evaluates them. Invisible to the systems increasingly deciding what buyers see. The site looks like it cost $100K to build. To an AI crawler, it looks like: <div id="app"></div> The most expensive blank page on the internet.

The most important content on your website might be the content nobody ever reads. “Content” in 2020 meant blog posts. Articles. Videos. Thought leadership. All still valuable. Still drives backlinks. Still gets cited. Still builds authority. But “content” in 2026 has another layer most teams haven't built yet. Structured data. Schema markup. Comparison tables. FAQs with your brand in the answer. Static HTML pricing. Machine-readable product information. The old content talks to humans. The new layer talks to machines. Both matter. But most brands are missing the second layer entirely. Blog posts build authority. Schema tells AI what your product actually is. The article gets you crawled. The structured data helps determine what AI can extract when it arrives. The strange part? Some of the content humans never read is becoming the content that determines whether machines send buyers to you. And most content teams don't even know this layer exists.

AI is the most honest mirror your brand will ever face. it doesn't care about your intentions. doesn't read your brand guidelines. doesn't attend your all-hands. it shows you exactly how the internet describes you when you're not in the room. most brands don't like what they see. that discomfort is the starting point. not the problem. the problem was always there. AI just showed it to you without being polite about it.

daily schema scans is smart. most brands set schema once and forget it exists for two years. the features change, the schema doesn't, and AI keeps citing the old version. automated checks catch the drift before it costs you.

5 AI visibility checks. 10 seconds each. Do them right now. 1. Check your price. Right-click your product page → View Source → Ctrl+F your price. Is it actually in the HTML? If not, AI may not be able to reliably see it. 2. Check your AI crawlers. Go to yourdomain.com/robots.txt. Search for GPTBot or ClaudeBot. See “Disallow”? You may be blocking the engines you're trying to rank in. 3. Check your first sentence. Read the first sentence of your product page. Does it clearly say what you sell, who it's for, and what it does? Or does it say “Welcome”? 4. Check your meta description. Open View Source and find the meta description. Was it written for 2026? Or is it still describing the product you sold in 2021? AI can still use it as a summary signal. 5. Check your feature count. Count the features on your product page. Now ask ChatGPT what your product does. Count the features it lists. Do the numbers match? 50 seconds. At least one of these will probably surprise you. Reply with the number.

Same conversation. Every week. Different client. “We need an AI visibility strategy.” “Have you checked what AI says about you?” “No.” “Let’s do that first.” We check. They go quiet. Then: “How long has it been saying this?” Every time. The check replaces the strategy conversation. Because the strategy is usually the same: Fix what’s wrong. The only variable is what’s wrong. And you can't know that from a meeting. You can only know it by checking. The strategy starts after the screenshot. Never before it.