When shoppers ask AI what to buy, is your store the answer?
Product research is moving into chat. "Best running shoes for flat feet," "gift for a coffee nerd" - AI answers with specific products and links. Viz shows you whether yours make the list, which assistants send shoppers who buy, and what the agents reading your catalog came to do, with tools they can use when they come to act.
Your store is being read, ranked, and recommended without you.
AI shopping advice is a black box
Assistants recommend products every day. You can't see which stores they pull from, whether yours is one of them, or which competitor owns the answer.
Scrapers love product pages
Price scrapers, training crawlers, and agents hammer catalogs harder than any other page type - unmeasured load, and your pricing data walks out the door.
"Direct" orders that aren't
A shopper who clicks through from a ChatGPT recommendation books as direct traffic. The channel driving new orders gets zero credit in your reporting.
Built for stores. Native to WooCommerce.
Run WooCommerce? Viz detects it and uses your product categories to suggest the shopping prompts worth tracking - no blank-page setup.
Track shopping prompts
Monitor the buying questions that matter for your catalog across the major AI engines - whether you're recommended, cited, and how you rank against competing stores. AI visibility →
Publish the answers shoppers ask for
Viz researches the buying questions your store should own, then delivers reviewed buying guides into Posts → Drafts - each one tracked for AI crawls, citations, and referred orders once you publish. AI Content Pipeline →
Credit every order to the assistant that sent it
Every WooCommerce order is credited to the assistant that sent the shopper, with value per assistant, captured on your server with no GA4 and no pixel. Landing Paths shows AI-referred shoppers one next step and reports the lift. Recommendations turn into revenue you can count, on every plan. AI Conversions →
Let the agent shop, let the shopper pay
On WooCommerce every page registers search_products, browse_products, get_product, get_cart, add_to_cart, update_cart and apply_coupon for the agent in a shopper's browser (WebMCP). It fills the shopper's own cart; the shopper checks out; the order is credited to "WebMCP agent". Nothing to configure. Agent tools →
Control the scrapers
See every bot hitting your product pages, then decide who stays. Block aggressive AI scrapers with enforced 403s while search engines keep indexing normally. AI Control →
Product data is exactly what machines read best - and take fastest.
A product page is the most machine-legible content on the web: a name, a price, specifications, availability, reviews. That is why catalogs attract more automated traffic than any other page type - price-comparison scrapers, training crawlers, retrieval bots assembling shopping answers, and increasingly, agents acting for a single shopper. Each reads the same page for a very different reason, and each deserves a different response from you.
A retrieval bot fetching your product page to answer "best espresso grinder under $300" can put you in front of a buyer at the exact moment of decision - that crawl is marketing. A competitor's price scraper hitting the same page hourly is extraction with zero upside. Treat them identically - block everything, or allow everything - and you either give away your pricing feed or opt out of AI shopping recommendations entirely.
This is why classification is the measurement that matters. Once you can tell scraper from shopper-assistant from search indexer on your own server logs, the policy almost writes itself: keep the bots that put products in front of buyers, throttle or block the ones that only take, and watch whether AI-referred sessions actually convert once they land.
Common questions, straight answers.
Do AI assistants actually send shoppers to stores?
Yes, and increasingly with intent. AI shopping referrals tend to land deep - on a specific product page rather than your homepage - because the assistant already narrowed the choice. Session counts are still small for most stores, but conversion behavior skews high-intent, which is exactly why the channel deserves attribution before it gets big.
Will blocking scrapers hurt my Google Shopping presence?
Not if you block precisely. Googlebot and the crawlers behind shopping surfaces are classified separately from third-party scrapers and AI training bots. Viz's per-bot controls let you refuse a price scraper without touching anything Google or Bing runs.
How does WooCommerce detection help?
Viz reads your product categories to suggest the shopping prompts worth tracking - the "best X for Y" questions a buyer of your catalog would ask an assistant. You start from a relevant list instead of a blank page, then edit it to match how your customers actually talk.
Can bots inflate my conversion-rate denominators?
They routinely do. Bot visits that slip into analytics as sessions depress conversion rate and pollute funnel metrics. Because Viz classifies at the server, you can see how much of your raw request volume is automated and read your human metrics with that context.
AI shopping agents are supposed to buy things soon. Should I prepare?
Prepare by measuring, and by handing them tools. Agent traffic - AI acting for a specific shopper - is one of Viz's eight categories, so when agents browse your catalog you see it as it happens, and what they came to do: check prices, check stock, try to buy. On WooCommerce every page registers WebMCP tools the agent in a shopper's browser can call: search and browse the catalog, read a product with its variations, view and fill the shopper's own cart, apply a coupon. The agent never checks out; the shopper does, and the order is credited to "WebMCP agent". For agents arriving from outside a browser, /ai-contact/ is the plain form. Agents completing purchases is still rare; the stores that are readable and reachable when it stops being rare get the first orders.