Skip to content

Discovery

Discovery is Bot It’s catalog engine. It scans every product in your store, scores it on the data points AI shopping agents use to rank results, and — on Pro — fixes the gaps with AI, on your terms.

Why this matters: Shopify syndicates eligible stores to AI surfaces like ChatGPT, Copilot, and Gemini automatically. When your product data is incomplete, the missing pieces get inferred by AI with varying accuracy. Products with explicit, complete data are the ones agents can filter, trust, and recommend.

Every product is scored 0–100 (graded A–F) across checks that mirror the agentic ranking factors:

CheckWhat it looks at
Taxonomy categoryIs a category set, and is it as specific as possible?
Category attributesAre attributes (color, material, size type…) explicitly set instead of AI-guessed?
Description depthIs the description detailed enough for agents to answer buyer questions?
GTIN / barcodesDo variants carry barcodes? Missing GTINs are a common cause of AI feed indexing failures.
Images & alt textEnough images, and does each one have descriptive alt text?
Variants & optionsAre options clearly named (Color, Size) so agents can filter?
Availability & priceIn stock and priced — out-of-stock products drop out of agent results.
SEO fieldsSearch-engine title and description set.
Agent narrativeBot It narrative present for the MCP enrichment layer.
PublishedLive on the Online Store with a URL agents can cite.

Run a scan from Discovery → Overview. Large catalogs continue scanning in the background — the page updates automatically.

Above product data sit the on/off switches that can hide your whole store from agents. The gates check verifies, from the outside, what agents actually see:

  • Storefront reachable (no password page)
  • robots.txt doesn’t block AI crawlers (a common side effect of other SEO/security apps)
  • /agents.md and /llms.txt serving
  • UCP discovery profile present
  • Agentic discovery sitemap serving

Generate fixes with AI runs a vision pass over your products: Bot It looks at your product images and text, then writes what’s missing — image alt text, SEO title and description, detected attribute values, and an agent narrative. Fixes never invent facts: no fabricated barcodes, no made-up specs.

You control how fixes apply, globally and per field type:

ModeBehavior
Full ReviewEvery fix waits in the Fix Queue for your approval
Semi AutoFills for empty fields apply automatically; anything replacing existing content waits for review
Full AutoEverything applies automatically

Sensitive fields are capped for safety: taxonomy category changes never exceed Semi Auto, and barcodes and category attributes are always recommend-only.

Each write Bot It makes is recorded in Discovery → History with its previous value. One click reverts any change — in any mode. Search the history and filter by date range.

Fixes waiting for review appear in Discovery → Fix Queue with a before/after view. Approve or reject individually, or approve everything at once. Recommend-only items (like detected attribute values) include the suggestion for you to apply in your Shopify admin.

Discovery → Products lists your catalog sorted worst-first, with each product’s grade and its top issues. Review & fix opens the full scorecard, lets you run an AI pass on just that product, and deep-links to the product in your Shopify admin.

AI generation uses a monthly token budget included with Pro. Track usage on the Discovery Overview. Image content rarely changes, so the first full-catalog pass is the big one — after that, usage drops to a trickle.