How It Works

Shopify put you on the AI shelf. Bot It works on whether you get picked.

Shopify now syndicates every eligible catalog to ChatGPT, Microsoft Copilot, Google AI Mode and Gemini, and Meta. Inclusion is automatic. Ranking is decided by your product data. Here is exactly what Bot It does with it.

The AI commerce stack, as of September 2026

Shopify Catalog is the feed. It syndicates your products to the AI surfaces and answers their searches through the Universal Commerce Protocol (UCP), which every Shopify store now serves at /.well-known/ucp and /api/ucp/mcp.

Shopify names the ranking factors plainly: product data quality, availability, price, shipping, and trust signals. When product data is missing, Shopify infers it with machine learning and marks those fields inferred, with what its own documentation calls "varying accuracy."

Bot It's job is to replace inferred data with explicit data, automatically, and to show you where you rank as a result.

Your product data
Category, attributes, barcodes, descriptions, images, variants, policies
Admin API
Bot It Discovery Engine
Audits every product, writes the missing data, logs and reverts, tracks rank
Bot It lives here
Shopify Catalog + UCP
AI shopping surfaces
ChatGPT, Copilot, Google AI Mode & Gemini, Meta, Perplexity, Claude agents, Shop

What Bot It does with your catalog

A background worker runs this loop for every store, every few hours on Pro and daily on Free. Nothing is written to your store without the automation level you chose.

1
Hydrate
Bot It reads every product through the Admin API: taxonomy category, category attributes, barcodes, description, images and alt text, variants and option names, availability, price, and your shipping and returns policies. Large catalogs sync in resumable pages, so a 10,000-product store never blocks.
2
Audit
Each product is scored out of 100 on 13 checks that map to the data AI surfaces filter and rank on. Catalog readiness is the largest dimension of your store-wide Agent Readiness Score. You see the failing products, the failing checks, and what each one costs you.
3
Propose
For every gap Bot It can close, an AI model reads the product and its photos once and proposes the missing pieces: image alt text, SEO title and description, category attributes, and an agent-facing narrative. Attribute proposals are constrained to Shopify's official taxonomy values for that category, so they land exactly as agents expect.
4
Apply
Your fire selector decides what happens next. Review: nothing is written until you approve. Semi Auto: low-risk fixes apply on their own; category and attribute changes wait for you. Full Auto: everything within the safety ceilings applies. Category changes never go above Semi Auto because they can affect tax. Barcodes are never invented.
5
Log and revert
Every write is recorded with its previous value in a changelog you can browse. One click reverts any change, or a whole batch. Bot It writes real Shopify data (standard metafields and metaobjects in the shopify namespace), so nothing depends on the app staying installed.
6
Broadcast and track
Changed products are pushed to the IndexNow layer that feeds Bing and Copilot (Pro). Every night the Rank Tracker runs your chosen queries against the same Shopify global catalog the AI surfaces shop from and records where your products place against named competitors.

The 13 checks

Each maps to a documented Shopify Catalog matching field, filter, or ranking factor. Background: how Shopify Catalog ranks products and how category attributes are set at scale.

Taxonomy category

Explicit and as specific as possible. It decides which attribute set applies and whether filtered agent queries can find the product at all.

Category attributes

Color, material, size, pattern, target gender, age group and the rest, set explicitly rather than inferred. These are the fields agents filter on.

Barcode / GTIN

The cross-retailer identity field. Bot It flags missing or invalid barcodes and tells you where to get real ones. It never generates them.

Description depth

Long, specific, question-answering copy. Shopify's own listing-quality meter scores word count.

Descriptive title

Literal and keyword-bearing. "Ocean Breeze" alone starves an agent; "Ocean Breeze linen wide-leg pants" does not.

Images and alt text

Image count and alt coverage. Catalog returns media with alt text, and multimodal search reads it.

Variants and options

Complete variants with meaningful option names. "Title / Default Title" is a fail.

Availability

In stock and published to the online store. Out-of-stock products drop out of agent results.

Price integrity

A real price above zero, with compare-at pricing for deals.

Shipping and returns

Complete policies. Shopify lists them as trust signals and returns them to agents alongside products.

Weight and dimensions

Needed for delivery estimates, which are a named ranking input.

Reviews

Rating and count from your review app. Ratings are a Catalog search filter.

Agent narrative

Fit, use case and materials in plain language, the context agents quote when they explain a recommendation.

Store-level gates come first

Perfect product data does nothing if the store itself is hidden. Bot It checks these on install and on every dashboard load. The manual version is in five gates to check before you touch product data.

Password page

A storefront password blocks Catalog eligibility and every agent fetch.

robots.txt blocks

Other SEO and security apps often add blanket disallows for AI crawlers without telling you. Bot It reads your live file and names the offending rules.

UCP manifest

Confirms /.well-known/ucp is reachable, on the current spec version, and lists payment handlers. Shopify serves this file for you; Bot It audits it.

agents.md and llms.txt

Confirms the agent-facing files resolve and are yours, not a stale template.

Agentic sitemap

Checks Shopify's sitemap_agentic_discovery.xml is live for the domain.

Agentic Storefronts

The per-channel toggles are not queryable by any API. Bot It tells you exactly where to verify them in your admin.

UCP and MCP, briefly

The two protocols under every agent interaction with a Shopify store, and where Bot It fits.

UCP is the commerce standard

The Universal Commerce Protocol was announced by Google and Shopify in January 2026 and is now governed with Amazon, Meta, Microsoft, Stripe and others. Version 2026-08-25 is live on every Shopify store. It defines discovery, catalog search, cart, checkout and order capabilities.

MCP is the transport

The Model Context Protocol is how agents call tools. Shopify exposes UCP through MCP servers on your domain: catalog, cart and checkout at /api/ucp/mcp, policies and FAQ at /api/mcp. Bot It audits and enriches what flows through them; it does not replace them.

Unsigned is normal

UCP allows signed manifests, but no platform signs merchant manifests today, Shopify included. Bot It reports this as information, not a defect, and will flag the day it changes.

Enrichment proxy (Pro)

Agents that want more than raw data can call the same UCP tools through mcp.botitai.com. Search results come back with your brand narrative, product narratives and review summary attached. Your agents.md advertises it automatically.

Frequently Asked Questions

No. Bot It writes product data through the Admin API and reads your storefront the way an agent would. Nothing runs in your theme for human shoppers. The optional agents.md and llms.txt templates are plain text files Shopify already serves.

The full catalog audit and score, the store-level gates check, brand voice and unlimited narratives, agent file management, and email support. Free shows you every gap. Pro closes them.

Only if you choose Semi Auto or Full Auto, and only within the safety ceilings. New installs start in Review mode. Every write is logged with its previous value and can be reverted in one click.

No, and you should be suspicious of any tool that offers to. Bot It flags missing or invalid barcodes and points you to GS1 or your supplier. Invented GTINs get products rejected across every retailer feed.

Everything that reads Shopify Catalog: ChatGPT, Microsoft Copilot, Google AI Mode and Gemini, Meta (since September 2026), Perplexity, the Shop app, and agents built on Claude or the Shopify UCP toolkit. ChatGPT is a discovery and referral channel; purchases complete on your store.

The Rank Tracker searches the Shopify global catalog (the same source the AI surfaces shop from) with your chosen queries and records where each product places, night after night. It is the closest available measurement of your position on the AI shelf, not a scrape of ChatGPT answers.

Install, approve the permissions, and the first audit starts within minutes. Small catalogs finish in one pass; large ones sync in the background and show progress on the dashboard.

See your catalog the way AI agents do.

The audit is free and takes minutes. Install from the Shopify App Store.

Install on Shopify →