PARCEIT Blog
Technical, evidence-based articles on how AI search engines actually work, and what to do about it. No fluff. No buzzwords. Just crawl data and code.
Cloudflare's three-tier crawler defaults went live September 15. Behind them sits a bigger plan: infrastructure for AI agents to discover, authenticate, and read your site. Part 1 of 2 covers the readiness layer and the protocols behind it.
PwC surveyed 4,093 U.S. consumers in June: 29% will use AI somewhere in their holiday shopping, but only 4% expect to buy inside an AI tool. What that split means for a small store before Black Friday.
Merchant Center is the mailbox where your product data enters Google's world. What your feed controls, what it does not, and a ten-minute check every Shopify merchant should run after the September 22 AI checkout enrollment.
Shopify's free agentic-readiness scanner tells Shopify stores 'Good news' without running a single check. We ran it against three real stores, then ran the same pages through the Parceit audit engine. Here is what each one found.
On September 22, 2026, Google auto-enrolled eligible US Shopify merchants into native checkout inside AI Mode and Gemini. What changed, the measurement gaps it opens, and how to turn it off.
A zero count in the Agentic Storefronts screen means a requirement is not met, not a broken channel. Every requirement on Shopify's list, in checking order, with the fix for each.
AI visibility on Shopify is two problems: being in the catalog, which Shopify handles for free, and being cited in AI answers, which is yours to build. Telling them apart decides where your time and money should go.
Your product schema is generated by your theme, not by Shopify. On vintage and custom themes that code may never have arrived, and nothing goes back and adds it. How to check your store this week, and three fixes if the check comes up empty.
Shopify's Agentic Storefronts put your products inside ChatGPT, Copilot, Google AI Mode, and Gemini. What the new admin screen actually does, what it leaves to your theme, and the four checks to run this week.
91% of AI citations appear in only one search engine. If you track one engine, you are blind to most of your AI visibility. Here is what the data shows and what to do about it.
The EU AI Act requires AI output to be marked, not anchored to a blockchain. A chain proves a file existed at a timestamp, not who created it. Here is why the tamper-proof option cannot prove authorship, and what to do instead.
62% of ChatGPT referrals now land on homepages. Stores with clean Merchant Center feeds get product-page visits converting at 6.2%. Here is what to fix.
A claim made the rounds this week that Amazon had unblocked its AI crawlers. We checked. It is false, and the real story in AI shopping is bigger than a robots.txt edit.
Google swapped the model behind AI Mode and told the press before it told your dashboard. If your vendor's report does not stamp model versions, your trend line just quietly broke. Here is what reproducible AI visibility measurement requires.
The EU AI Act demands AI content markings be "sufficiently reliable and robust." A July paper tested the three major text watermarking method families against the US courtroom standard for scientific evidence. Under paraphrasing, two lost the mark 100% of the time. Here is what the evidence gap means for your business.
The IAB Tech Lab's new Agentic Audiences standard lets advertising AI agents trade anonymous audience summaries instead of lists of people. Here is what it means in plain English, what it cannot do, and why it matters for AI visibility.
OpenAI's in-house search index keeps your page title plus about 200 characters from the top of the page. Those characters decide how ChatGPT represents your business. Here is the ten-minute check.
A survey of 163 AI visibility practitioners found they rate the data 4.20 out of 5 and their willingness to fund the platforms selling it 3.19 out of 5. The complaints are specific and they are correct. Here is what decision-grade measurement would have to look like.
A straight answer for anyone searching for free AI audit tools: what a free audit checks, the four options worth your time in 2026, and exactly where the free tier stops working.
ChatGPT is moving its Sources button behind the three-dot menu for some users. The citation still exists. Reaching it now takes a click most users never make. Here is what that does to citation measurement.
An analysis of 60 ChatGPT conversations shows the model writes brand names into its own search queries before it retrieves anything. Being named is worth roughly 33x more than being findable. Here is what that means for how you measure AI visibility.
Anthropic became the first AI provider to watermark its output for EU AI Act compliance. The watermark exists but no public detection tool can read it. Here is what the detection gap means for small businesses dealing with AI-generated misinformation.
The EU AI Act's transparency rules took effect August 2, 2026. If your marketing team uses AI to write content, this law may apply to you. Here is what Article 50 requires, who it affects, and what the human-review exception covers.
ChatGPT has 900 million weekly users. When they ask it for a recommendation in your industry, does your business come up? Here is a 10-minute diagnostic with five tests that catch the specific problems keeping your site out of ChatGPT answers.
France's main news publisher association filed an antitrust complaint alleging Google AI Overviews are costing publishers 33 to 38% of their traffic. Here is what the Arcom data means for brands measuring AI visibility.
You rank well organically, but does Google's AI Overview actually cite your site? Here is how to check your AI Overview visibility step by step, what to fix first, and how to tell the difference between being retrieved and being cited.
ChatGPT has 900 million weekly users. When they ask it for a recommendation in your industry, does your business show up? Here is how to check your ChatGPT search visibility, what breaks it, and what to fix first.
Ready-to-use FAQPage schema JSON-LD code you can copy and paste today. Includes single-question, multi-question, WordPress, Shopify, and React examples, plus validation and common errors.
A study of 216 Shopify stores found only 13% are fully readable by AI search engines. Meanwhile AI-referred shoppers convert 42% better than paid search traffic. Here is what is broken, what is causing it, and how to fix it.
Three AI visibility tools, three different scores. The inconsistency is not a bug, it is architecture. Here is why tools disagree, what decision-grade measurement means, and how to know which numbers you can trust.
91% of AI citations appear in only one search engine. If you track one engine, you are blind to most of your AI visibility. Here is what the data shows and what to do about it.
The biggest publishers in the world are blocking AI crawlers. Individual creators are doing the opposite. This divergence is reshaping who gets cited in AI answers, and brands need to understand it.
AI overviews from Google, Perplexity, and ChatGPT cite some pages and ignore others that rank just as well. The difference is not luck. Four signals determine whether your content gets cited, and they map directly to the IAB visibility framework published in August 2026.
Cloudflare replaced its single AI bot toggle with a three-tier system. The new defaults take effect September 15 and could accidentally block Googlebot. Here is what to check before the deadline.
You added FAQ schema, HowTo schema, and clean JSON-LD. Your AI citations did not move. The data shows schema is a hygiene factor, not a citation driver. Here is what actually gets you cited in AI search.
You rank in the top five, but Google's AI Overview cites someone else. Ranking got you retrieved. Content gets you cited. Here are the five reasons your page loses at extraction and what to fix first.
Google AI falsely declared a Japanese company bankrupt. More than half of retailers in a UK study had wrong information spread by AI chatbots. AI search can invent facts about your business that you never approved, never published, and cannot easily correct. Here is what causes AI hallucinations about brands, what the IAB calls Portrayal, and how to check whether AI engines are describing you accurately.
Publishers are blocking Google-Extended thinking they've opted out of AI Overviews. They haven't. Here is what each crawler actually controls, what blocking really does, and how to decide whether your site should opt out.
When publishers block AI crawlers, they change the pool of sources AI engines can draw from. That shift affects every measurement of AI visibility. Here is how Parceit accounts for it.
Perplexity cites some pages and ignores others that rank just as well. The difference is not luck. Perplexity runs a three-step pipeline: retrieve, extract, synthesize. Here is how each step works, what makes a page more likely to survive each gate, and how to check whether your content is making it through.
Semrush published an AI Visibility Index over 126M prompts, and Adobe launched Brand Visibility. The market is crowding up. But the IAB framework, published August 3, just leveled the field. Here is why methodology transparency, not funding, wins the AI visibility race.
GEO is the term Google uses and the term your customers search for. But most GEO advice is recycled SEO with new buzzwords. Here are the five fixes that actually move the needle for AI search visibility, with evidence from crawling 1,000 real websites.
Booking Holdings says AI mentions are up but referrals remain under 1% of room nights. That gap is the most important story in AI search right now. Here is why most AI visibility tools stop at the mention and never close the loop to revenue, and what the IAB framework says about it.
Your robots.txt says AI crawlers are allowed. But are they actually crawling your pages? Server access logs tell you exactly which pages GPTBot, ClaudeBot, and PerplexityBot are reading and which they are skipping. Here is how to check, what the common problems look like, and what to fix.
The IAB released Measuring Visibility in the AI Era, the first industry-standardized framework for tracking brand visibility in AI search. The 4 P's: Presence, Prominence, Portrayal, and Persuasion. Here is what the framework says, how it maps to the signals that actually determine AI visibility, and what brands should do with it.
Google killed the FAQ rich result in 2023 and the SEO industry gave up on FAQ schema. That was a mistake. ChatGPT, Perplexity, Google AI Overviews, and Claude all read FAQPage schema directly. Here is what actually happened, where things stand now, and whether you should implement it.
Google Search Console shows zero clicks for AI search. It reports rankings that do not match what real users see. Here is why GSC is blind to AI search, and the five checks that tell you the truth about whether ChatGPT, Perplexity, and Google AI Overviews can actually find your business.
Parceit, GrackerAI, Knowatoa, Otterly, Semrush, Cloudflare, Adobe, and Somantra all claim to help with AI search visibility. We compared what each one actually does, what it costs, and whether it tells you what is wrong or actually fixes it. Honest, fair, and current as of August 2026.
When a customer asks ChatGPT or Perplexity for a restaurant recommendation, most independent restaurants do not appear in the answer. It is not because the food is bad. It is because their websites are unreadable to AI crawlers. Here is what is going wrong and what to do about it.
Structured data is the fastest path to getting cited by ChatGPT, Perplexity, and Google AI Overviews. Here is which schema types matter, how to add them, and the mistakes that make them useless. Current as of August 2026.
Most businesses have no idea whether they appear in AI search results. Here is how to check if ChatGPT, Perplexity, and Google AI Overviews can find your website, and what to fix if they cannot.
Your robots.txt file decides whether AI search engines can read your site. If it blocks GPTBot, PerplexityBot, or ClaudeBot, you will not appear in ChatGPT, Perplexity, or Google AI Overviews. Here is how to check, allow, and verify AI crawler access.
We analyzed the robots.txt files of the top 1,000 websites. 25.6% block at least one critical AI search crawler. GPTBot, ClaudeBot, and Google-Extended are the most blocked. Here is the full data.
Google killed the FAQ rich result in 2023, but AI search engines like ChatGPT and Perplexity still read FAQPage schema directly from your HTML. Here is how to add it, where to put it, how to test it, and the mistakes that break it. Current as of July 2026.
Perplexity AI is the answer engine that actually sends you traffic. Unlike ChatGPT, it puts clickable source links directly inline with every claim. This guide covers the signals that determine whether Perplexity retrieves, reads, and cites your content. Here is what works.
OpenAI struck a licensing deal with Yelp in July 2026. The terms give ChatGPT direct access to Yelp's structured review data: ratings, business hours, categories, and verified customer reviews. When someone asks ChatGPT for a restaurant recommendation now, the answer may come from a paid data pipeline, not from crawling your website.
FAQ sections improve performance in AI search by giving language models exactly what they need to generate cited answers: short, self-contained blocks of text that directly address specific questions. When those sections are marked up with FAQPage schema, they become the easiest content on your page for an AI search engine to extract, trust, and cite.
The llms.txt file is a plain text file that websites use to give AI crawlers context about what the site is and which pages matter most. It sits at the root of a domain, the same place you would find robots.txt or sitemap.xml. If you want to know whether a site has one, and what it says, checking takes about five seconds.
Google AI Overviews sit at the top of search results for a growing share of queries. They summarize content from multiple sources, cite those sources inline, and push the traditional blue links below the fold. If your site is not in the overview, it might as well be on page two.
A new meta-study of 45 GEO papers finds the famous "40% visibility gain" is a lab artifact, not a business outcome. Here is what the evidence actually says about ranking in AI search, and why core retrieval signals beat prompt-tracking every time.
Is AEO really different from SEO? We analyzed real websites and their signals. The evidence shows exactly where they overlap, where they diverge, and what you actually need to change. No buzzwords.
Everyone publishes generic "AI visibility" lists. We took a different approach: we analyzed what makes ChatGPT cite one site and skip another. Here is the playbook.
llms.txt is an emerging standard for giving AI crawlers context about your website. This guide explains what it is, how to create one, and whether you actually need it, with real examples from sites we have crawled.
Parceit crawled 18 major restaurant chains and 49 independent restaurants and cafes. We measured exactly what ChatGPT, Perplexity, and Google AI see when they visit these sites. Only 1 out of 67 had FAQPage schema. McDonald's scored zero.
When Google killed FAQ rich results in 2023, the SEO industry abandoned FAQPage schema. That was a mistake. AI search engines like ChatGPT and Perplexity rely on FAQ schema for citations. Here is the full picture of what happened, who benefits now, and why independents have a window that is closing.
AEO (Answer Engine Optimization) is the practice of structuring your website so AI search engines like ChatGPT, Perplexity, and Google AI Overviews can find, understand, and cite it. Here is exactly how it works, what it costs, and how to do it yourself.
Answer Engine Optimization and Generative Engine Optimization are not separate disciplines. They are SEO with structured data on top. Here is the technical evidence from crawling real websites.