# Blog | AI Commerce Watch Source: https://aicommercewatch.com/blog/ [AI Product Descriptions That Invent Features: The Hallucination Tax on Returns](2026-11-09-ai-product-descriptions-invent-features-returns.md) 2026-11-09 - AI writers fill gaps in your product data with plausible specs from category patterns. A merchant found four of ten air fryer descriptions had the wrong wattage. Here is the fix. [AI-Generated Alt Text on Product Photos: Helpful or an Accessibility Liability](2026-11-09-ai-generated-alt-text-product-photos-liability.md) 2026-11-09 - Vision models can draft alt text for a thousand product photos in an afternoon, and they miss context-critical details 15 to 30 percent of the time. Here is where the liability sits. [Checkout Extensibility Killed the Old Tracking Snippet. Web Pixels Are the Replacement.](2026-11-08-checkout-extensibility-web-pixels.md) 2026-11-08 - Script tags with all or order-status scope stopped being creatable in February 2025, and they stop running on the order status page in August 2026. Web Pixels are the supported path. [Your Shopify Store Is Slow. LCP Tells You Exactly Where.](2026-11-08-shopify-lcp-where-store-slow.md) 2026-11-08 - Largest Contentful Paint is the metric that decides whether a Shopify store converts. Shopify's 2026 data puts a number on every 100 milliseconds. [The Chatbot Told Her Shipping Was Free. It Was Not.](2026-11-07-ai-chatbots-answer-shipping-questions-wrong.md) 2026-11-07 - AI chatbots on storefronts invent shipping times, free-shipping thresholds, and delivery promises the store never made. The hallucination shows up in the support queue, not in the chatbot's own metrics. [The AI Size Chart Measured a Different Garment](2026-11-07-ai-generated-size-charts-measure-wrong.md) 2026-11-07 - AI-generated size charts look complete and measure the wrong thing: a flat-lay photo, a different season's cut, or a size run the brand never sold. Returns follow the chart, not the garment. [AI Chatbots That Invent Discount Codes](2026-11-06-ai-chatbots-invent-discount-codes.md) 2026-11-06 - A chatbot that hallucinates a 20% code is not being helpful. It is writing a discount the store never offered, and the customer will expect the checkout to honour it. [AI Size Finders That Recommend the Wrong Size](2026-11-06-ai-size-finders-wrong-size-recommendations.md) 2026-11-06 - A size finder that recommends from the wrong data sends the customer the wrong size with confidence. The return costs the store more than the sale was worth. [AI Review Summaries That Contradict the Reviews Underneath Them](2026-11-05-ai-review-summaries-contradict-reviews.md) 2026-11-05 - Summarised review widgets invent consensus: 'runs large' over reviews saying true to size. Why summarisation drifts, and the sampling audit that catches it before shoppers do. [Personalised Price Displays: Auditing What Each Visitor Is Shown](2026-11-05-ai-personalised-pricing-displays-consistency-audit.md) 2026-11-05 - When an AI or rules engine varies prices, discounts, or bundles by visitor, nobody sees the full picture by testing in one browser. How to audit price presentation across segments without losing your mind. [AI Chatbots That Promise Delivery Dates They Cannot Know](2026-11-04-ai-chatbots-promise-delivery-dates.md) 2026-11-04 - A shopping assistant that answers when will it arrive with a confident date, drawn from nowhere, creates a promise the store then has to keep. Where that answer should come from, and when the bot should refuse. [When AI Search and Your Filters Disagree About Your Own Catalogue](2026-11-04-ai-search-filters-disagree-catalogue.md) 2026-11-04 - Shoppers filter for waterproof under 100 and the AI search box recommends something leather at twice the price. Two retrieval systems reading the same catalogue differently is a data problem, and it is measurable. [AI Product Descriptions That Invent Claims: An Audit Workflow](2026-11-03-ai-product-descriptions-invent-claims-audit.md) 2026-11-03 - Generated descriptions do not usually lie loudly. They add a material, a certification, or a benefit nobody supplied. A practical audit that catches invented claims before they publish. [AI Visual Search Returns: Accuracy, Bias, and the Wrong Product Problem](2026-11-03-ai-visual-search-accuracy-wrong-product.md) 2026-11-03 - A shopper photographs a chair and gets table lamps. Visual search fails in patterns worth measuring: colour drift, category confusion, and results that ignore stock and price. [When AI Search Returns Nothing: Synonyms, Slang, and Zero Results](2026-10-28-ai-search-zero-results-synonyms.md) 2026-10-28 - Shoppers search in their own words, not catalog words. Here is how AI search handles the gap, and what zero results pages reveal about your vocabulary mismatch. [AI Fit Recommendation Tools: How Accuracy Holds Up in Practice](2026-10-28-ai-fit-recommendation-tools-accuracy.md) 2026-10-28 - Fit recommenders promise fewer returns, but their accuracy depends on the data shoppers give them and the bodies they were trained on. Here is what the results actually show. [AI Visual Search: When a Photo Finds the Wrong Product](2026-10-27-ai-visual-search-accuracy-product-discovery.md) 2026-10-27 - Visual search promises a photo in, product out, but accuracy varies by category, lighting, and catalogue quality. How to judge whether it helps shoppers or misleads them. [AI Bundle Builders: Good Recommendations or Expensive Guesses](2026-10-27-ai-bundle-builders-recommendation-quality.md) 2026-10-27 - AI bundle builders promise higher basket size, but weak pairing logic creates returns and distrust. How to evaluate bundle quality before scaling it. ## [AI Agents Read Your Size Guides and Compatibility Data, So Write Them Clearly](2026-10-15-ai-agents-size-guides-compatibility-data.md) Oct 15, 2026 Shoppers now ask agents whether a part fits, whether a size runs true, whether two products work together. Agents answer from your guide pages, if those pages are readable at all. ## [When an AI Agent Asks Your Store a Question: Chat, Contact Data, and Answer Readiness](2026-10-16-ai-agents-ask-store-questions-chat.md) Oct 16, 2026 Agents increasingly arrive with a specific question a product page did not answer. Stores whose answers are findable get the sale. Stores whose chat traps the question do not. ## [What AI Agents Do When a Product Is Out of Stock](2026-10-17-ai-agents-out-of-stock-alternatives.md) Oct 17, 2026 When an agent finds your product unavailable, it does not wait. It substitutes, and often substitutes a competitor. How stock signals and alternative guidance shape what it recommends next. ## [Product Feed Quality Is Now an AI Shopping Problem](2026-10-18-product-feed-quality-ai-shopping.md) Oct 18, 2026 Agents shop from your feed data, not your photography. Thin attributes, vague titles, and stale stock in the feed quietly remove products from agent answers. ## [The Handoff Problem: When AI Chat Should Give Up Gracefully](2026-10-19-ai-chatbots-handoff-human-support.md) Oct 19, 2026 A chatbot that cannot say I do not know traps customers. Designing escalation, context passing, and honest limits for support AI. ## [AI Agents Reading Your Shipping Policy: Write One They Cannot Misread](2026-10-20-ai-agents-reading-shipping-policies.md) Oct 20, 2026 Delivery promises are where agent summaries hurt most. How to state shipping times, cutoffs, and exceptions so an assistant repeats them correctly. ## [AI Written Product Descriptions, Accuracy, and the Returns They Cause](2026-10-21-ai-product-descriptions-accuracy-returns.md) Oct 21, 2026 Generated descriptions speed up catalogue work and quietly invent details. How to keep AI copy accurate, and why inaccuracy shows up in the returns data. ## [How AI Search Summaries Describe Your Returns Policy](2026-10-22-ai-search-summaries-returns-policy.md) Oct 22, 2026 AI summaries now paraphrase returns policies for shoppers before they visit. When the paraphrase is wrong, the store lives with a promise it never made. ## [AI Visual Search and the Photos Shoppers Bring](2026-10-23-ai-visual-search-product-photos.md) Oct 22, 2026 Visual search lets shoppers upload a photo and find similar products. What it does well, where it misreads, and how stores should present its results. ## [AI Personalised Pricing: What Shoppers Can and Cannot See](2026-10-24-ai-personalised-pricing-transparency.md) Oct 22, 2026 Dynamic and personalised pricing systems adjust offers per visitor, which raises a transparency problem stores need to manage openly. A plain guide to the practice and its trust cost. ## [AI Visual Search in Stores: Accuracy Shoppers Can Check](2026-10-25-ai-visual-search-accuracy.md) Oct 25, 2026 Shop by photo tools promise to find products from a picture. Their misses are instructive: similar looking wrong items, filtered out stock, silent failures. ## [AI Chatbots and Returns Questions: Where They Help and Where They Stall](2026-10-26-ai-chatbots-returns-questions.md) Oct 25, 2026 Returns questions are among the most common chat topics in commerce. AI chatbots handle the simple ones well and mangle the edge cases. A clear eyed breakdown. ## [AI Size Recommendation Tools: How Accurate Are They Really?](2026-10-26-ai-size-recommendation-accuracy.md) Oct 25, 2026 Size recommendation tools promise fewer returns. What accuracy really means, where the tools guess wrong, and the questions to ask before trusting the numbers. ## [When AI Review Summaries Mislead Shoppers](2026-10-25-ai-review-summaries-that-mislead.md) Oct 25, 2026 AI summaries compress hundreds of reviews into confident paragraphs that can invent consensus and bury real complaints. What stores should watch. ## [When AI Chatbots Invent Store Policies](2026-10-24-ai-chatbots-invented-policies.md) Oct 22, 2026 Shopping chatbots sometimes state delivery times, returns windows, and discounts that do not exist. How stores can catch invented policy answers before customers rely on them. ## [When AI Recommends Substitutes for Out of Stock Items](2026-10-23-ai-recommendations-out-of-stock-substitutes.md) Oct 22, 2026 Recommendation engines increasingly suggest a substitute when the wanted item sells out. When that helps, when it annoys, and how to judge the swap. ## [When AI Chatbots Give Sizing and Fit Advice, Accuracy Gets Tested](2026-10-22-ai-chatbots-sizing-fit-advice-accuracy.md) Oct 22, 2026 Store chatbots now answer fit questions with total confidence. What happens to returns and trust when the confident answer is wrong, and how stores can audit it. ## [When AI Search Answers Send Shoppers to Out of Stock Products](2026-10-21-ai-search-answers-out-of-stock-products.md) Oct 21, 2026 AI answers increasingly name specific products. When the named product is gone, everyone loses. How stores can keep AI driven visits honest and useful. ## [When AI Recommendations Get Prices and Stock Wrong](2026-10-20-ai-recommendations-price-accuracy.md) Oct 20, 2026 An assistant that quotes a stale price creates a disappointed checkout. How stores should ground recommendations in live price and stock data. ## [When AI Search Summaries Rewrite Your Product Claims](2026-10-19-ai-search-summaries-product-claims.md) Oct 19, 2026 AI answers summarise your store in their own words. How claims drift, why it matters, and what clean source copy fixes. ## [How AI Agents Read Your Returns Policy, and What They Get Wrong](2026-10-18-ai-agents-returns-policies-windows.md) Oct 18, 2026 Agents quote your return window to shoppers before purchase. If the policy is vague, split across pages, or buried in an image, the quote will be too. ## [How AI Agents Handle Coupons, Discount Codes, and Price Rules](2026-10-17-ai-agents-coupons-discount-codes.md) Oct 17, 2026 Agents are bad at finding the discount box and worse at knowing which code applies. What stores should publish so an agent prices an order correctly instead of guessing. ## [How AI Shopping Agents Read Your Shipping and Delivery Estimates](2026-10-16-ai-agents-shipping-delivery-estimates.md) Oct 16, 2026 Agents compare delivery promises as carefully as prices, and vague shipping pages make them cautious. What stores should publish so an agent can quote a date with confidence. ## [How AI Agents Compare Product Bundles and Bundle Pricing](2026-10-15-ai-agents-product-bundles-value-comparison.md) Oct 15, 2026 Bundles are designed for human eyes: a hero price, a strikethrough total, a badge. AI agents read them differently, item by item, and stores with messy bundle data lose the comparison. ## [How AI Agents Read Product Reviews Before Recommending a Store](2026-10-14-ai-agents-read-product-reviews-before-recommending.md) Oct 14, 2026 Shopping agents scan review volume, recency, and complaint themes before they rank a store. What agents extract from your reviews and how to keep that signal strong. ## [How AI Shopping Agents Compare Shipping and Returns Policies Before Buying](2026-10-13-ai-agents-shipping-returns-policies.md) Oct 13, 2026 AI agents do not just compare prices. They read shipping speed, fees, and returns terms, and drop stores whose policies are vague or hard to find. ## [How AI Shopping Agents Handle Product Variants and Options](2026-10-12-ai-agents-product-variants-options.md) Oct 12, 2026 Size, color, bundle: variants are where agent purchases go wrong. What agents do with product options, and how merchants can make their catalogs agent-proof. ## [What AI Shopping Agents Do When the Product Is Out of Stock](2026-10-11-ai-agents-out-of-stock-behavior.md) Oct 11, 2026 An agent that cannot buy is an agent that improvises. How shopping agents handle stockouts, and what merchants should do about it. ## [What do AI shopping agents read first on a product page?](2026-10-10-what-ai-agents-read-product-page.md) Oct 10, 2026 Agents do not browse the way people do. They read structured data, availability signals, and policy pages first, and the design second. Here is the order that matters. ## [How should a store track which orders came from AI shopping agents?](2026-10-09-tracking-orders-ai-shopping-agents.md) Oct 9, 2026 Agent-driven orders look like normal checkout traffic in most dashboards. Tagging them at the session level is the only way to know how much revenue agents actually drive. ## [What happens when AI agents start comparison shopping for your customers?](2026-10-08-ai-agents-price-comparison-shopping.md) Oct 8, 2026 AI agents do not get tired, do not get distracted, and check every competitor before answering. When agents comparison shop at machine speed, the store with the cleanest structured data wins. ## [Should stores build agent-ready landing pages for AI shoppers?](2026-10-07-agent-ready-landing-pages.md) Oct 7, 2026 AI shopping agents read pages differently than people: they want structured facts, not lifestyle copy. A lightweight agent-ready page with clean product data can win the agent's recommendation without changing the human experience. ## [What should a store do when AI agents abandon carts?](2026-10-06-ai-agents-abandoned-carts.md) Oct 6, 2026 AI agents add to cart constantly and check out rarely. Agent-abandoned carts are a new analytics category: part research, part intent signal, and a test of how your checkout treats non-human shoppers. ## [How should a store structure product data so AI agents describe products accurately?](2026-10-05-product-data-ai-agent-descriptions.md) Oct 5, 2026 AI shopping agents describe your products using whatever structured data they can find, and they invent the rest. Clean titles, complete attributes, and honest availability signals are the difference between an accurate summary and a confident hallucination. ## [How do AI shopping agents handle discount codes and promotions?](2026-10-04-ai-agents-discount-codes-promotions.md) Oct 4, 2026 Agents find codes the way determined shoppers do: they read the banners, test the fields, and stack what stacks. Stores need a policy on agent-applied discounts before the agent checks out, not after. ## [How can a store tell AI shopping agent traffic apart from regular bot traffic?](2026-10-03-ai-agent-traffic-vs-bot-traffic.md) Oct 3, 2026 AI shopping agents and bad bots both look like automation in server logs, but they behave differently and need opposite responses. The distinction comes from declared identity, browsing patterns, and intent. ## [What do AI shopping agents do when a product is out of stock?](2026-10-02-ai-agents-out-of-stock-products.md) Oct 2, 2026 An out-of-stock product is a dead end for a human shopper and a decision point for an AI agent. What the agent does next depends on signals most stores never designed for machines. ## [What should a store's terms of service say about AI agent purchases?](2026-10-01-terms-of-service-ai-agent-purchases.md) Oct 1, 2026 If AI agents can buy from your store, your terms need to name who is responsible when they do. Silence on agent purchases means the store inherits every dispute the agent creates. ## [The Merchant's Guide to AI Crawlers: Who Is Reading Your Catalog and Why](ai-crawlers-storefront-guide.md) Sep 15, 2026 GPTBot, ClaudeBot, PerplexityBot and a growing cast of AI crawlers now read ecommerce sites at scale. Some send shoppers back; some just take. A field guide to telling them apart. ## [When AI Agents Start Buying: What Agentic Checkout Means for Your Store](ai-shopping-agents-checkout.md) Sep 15, 2026 AI assistants are moving from recommending products to purchasing them. Agentic checkout changes fraud screening, attribution, and bot policy all at once. * [AI Recommendations That Ignore Stock Status and Recommend Sold Out Products](2026-11-02-ai-recommendations-ignore-stock-status.md) 2026-11-02 Recommendation engines keep surfacing sold out variants because the feed knows the product, not the inventory. How to audit it, measure the damage, and set a guardrail. * [AI Search Synonyms, Misspellings, and the Zero Results Page](2026-11-02-ai-search-synonyms-zero-results.md) 2026-11-02 Shoppers search the words they know, not your product names. How AI search handles synonyms and typos, where it still returns nothing, and how to test with a real query list. * [AI Review Summaries Smooth Away the Patterns You Actually Need](2026-10-30-ai-review-summaries-hiding-useful-patterns.md) 2026-10-30 A summary that reports customers love the fit tells you nothing. The value in reviews is in the split: who it fit, who it failed, and why. Summaries average that away. AI assistants are moving from recommending products to purchasing them. Agentic checkout changes fraud screening, attribution, and bot policy all at once. + [When AI Search Answers and Your Product Feed Disagree](2026-10-31-ai-search-product-feed-structured-data-mismatch.md) 2026-10-31 AI shopping answers quote price and availability from feeds and structured data. When those disagree with the product page, the AI version is the one the shopper believes first. AI assistants are moving from recommending products to purchasing them. Agentic checkout changes fraud screening, attribution, and bot policy all at once. - [AI Size Recommendations: Measure Them Against Your Own Returns](2026-11-01-ai-size-recommendations-measuring-returns.md) 2026-11-01 Fit recommendation tools promise a lot and report their own success. How to test one against your returns data by reason code, size run, and category before renewing anything. AI assistants are moving from recommending products to purchasing them. Agentic checkout changes fraud screening, attribution, and bot policy all at once. * [Auditing the AI Chatbot Handoff to a Human](2026-11-01-ai-chatbot-handoff-human-support-audit.md) 2026-11-01 The handoff is the moment AI support audits skip: whether a stuck customer can actually reach a person, how long the bot argues first, and what context survives the transfer. AI assistants are moving from recommending products to purchasing them. Agentic checkout changes fraud screening, attribution, and bot policy all at once. + [AI-Translated Product Pages: When Sizes, Care, and Safety Wording Go Wrong](2026-10-31-ai-translated-product-pages-size-care-errors.md) 2026-10-31 Machine translation of a catalogue is fast and mostly fine, until a size converts wrong, a care instruction flips meaning, or a safety warning softens. Auditing translated pages where errors cost money. AI assistants are moving from recommending products to purchasing them. Agentic checkout changes fraud screening, attribution, and bot policy all at once. - [AI Product Descriptions Get the Specs Wrong: Auditing Generated Copy Before It Costs Returns](2026-10-30-ai-product-descriptions-spec-errors-audit.md) 2026-10-30 Generated descriptions invent dimensions, materials, and compatibility. The errors read fluently, which is why they survive review and turn into returns and support tickets. AI assistants are moving from recommending products to purchasing them. Agentic checkout changes fraud screening, attribution, and bot policy all at once. * [AI Recommendations on Out of Stock Pages: Helpful Detour or Dead End](2026-10-29-ai-recommendations-out-of-stock-pages.md) 2026-10-29 An out of stock page with a recommendation rail can save the session or waste it, depending on whether the alternatives are in stock, similar, and honest about why they were picked. * [AI Chat Answers That Invent Shipping Policies, and How to Audit Them](2026-10-29-ai-chat-invented-shipping-policies-audit.md) 2026-10-29 A support chatbot that states a delivery time the store never promised creates refunds, chargebacks, and screenshots. Here is an audit routine for the answers a chat tool gives on your policies.