The Specs page
Publishing the eligibility and pricing extension contracts as a first-class reference surface. Drafts already live in the repo — this makes them public and citable, so the rules agents act on are out in the open.
Agent Reasoning Console
A visible console that explains whyan agent made each decision — why an item is visible or hidden, why a price changed, why a buyer is eligible or blocked. The Specs page is the contract layer that makes the Console's reasoning explainable.
Make your store visible to
AI shopping agents
RetailAgentOS helps small merchants publish pricing, eligibility, fulfilment, and promotion logic in a form AI agents can understand before they recommend, quote, or hand off to checkout.
Built on UCP. Works alongside your existing commerce system — not instead of it.
Why most small merchants are invisible to agents
UCP gives the rails — discovery, catalog, cart, checkout handoff. But agents still can't reason about a merchant's rules. That's the gap. These are the four ways it breaks.
AI shopping agents cannot recommend products they cannot interpret. Most small merchant catalogs have no machine-readable rules.
Without structured pricing context, agents default to list price — ignoring member rates, volume tiers, and active promotions.
Agents confirm shipping for items that are pickup-only, or recommend products unavailable in the buyer's region.
Eligibility, MOQ, and qualification checks happen at checkout — after the agent has already built the cart. Buyers abandon.
What RetailAgentOS does
Three steps from store rules to correct outcomes.
Publish your store rules
Pricing tiers, buyer qualifications, fulfilment zones, promotions — declared once in a structured, machine-readable form.
Let agents understand your store
Any AI shopping agent can read your declarations, evaluate buyer context, and reason about what is valid for this buyer at this moment.
Drive correct outcomes
The right product to the right buyer at the right price — with the right next action: recommendation, quote, callback, or checkout.
What it looks like for real merchants
Three different store types. Three different rule sets. One platform handling all of it correctly.
Sara's products now appear when AI shopping assistants look for personalised gifts. Declaring her catalog once made her visible to agents she never had to configure individually.
B&T's volume tiers and qualification gates are declared once. Agents always quote the right price to the right buyer — no manual corrections, no checkout surprises.
Fresh Corner's weekly sales and local delivery zones are visible to agents at browse time. Buyers only see what they can actually receive, at the price that actually applies.
Platform direction
Phase 1 proves the foundation. These are the capabilities being built next — turning the protocol layer into a full merchant-facing platform.
Store Setup
Configure pricing rules, buyer qualifications, fulfilment zones, and promotions — without needing to understand the underlying protocol.
Agent Understanding
See exactly how AI agents interpret your store rules. Inspect the reasoning behind every visibility, eligibility, and pricing decision.
Agent Actions
Quote, reserve, schedule a callback, route to WhatsApp, or hand off to checkout — all as agent-initiated actions driven by your declared rules.
Outcomes
Track what agents did on your behalf: completed handoffs, captured leads, qualification completions, and agent-assisted conversions.
See it working for merchants like yours
Walk through Sara, B&T, and Fresh Corner — see how each merchant's rules are declared and how the agent responds to different buyer contexts in real time.
Retail technology operator building the agentic commerce layer for small merchants.
The gap between UCP and what agents need isn't theoretical — it's diagnosed from years inside rules-enforcement systems where correctness has real consequences.