RedHub AIRedHub AI
Tool Gate · How it works

How the gate works

How the Agentic Commerce Product Feed Readiness gate works

This tool grades a single product listing across 7 machine-readability and trust signals AI shopping agents check before recommending it. Each signal is scored 0–1; the verdict is the weakest one — then a gate that can only make it worse.

The 7 signals

01Machine-readable feed
02Identifier completenesstop trigger
03Attribute & answer depth
04Variant grouping
05Price & inventory freshness
06Feed-vs-page consistencytop trigger
07Trust & seller block
AGENT-READY
every signal marked 2 - fully clean
GAPS
weakest signal = 1, with no gate signal tripped
UNREADABLE
any signal = 0, or the identifier-completeness/feed-vs-page-consistency gate fires (feed-page contradiction at 1-or-below, or missing identifiers at 0)
Agent-distrust gate · worsen-only

Identifier completeness at 0, or feed-vs-page consistency at 1-or-below, forces UNREADABLE even when every other signal reads clean - an agent distrusts a self-contradicting or unidentifiable item rather than just finding it thin.

UNREADABLE

Why: Six of seven signals read fully clean. But the feed and the live product page disagree - feed-vs-page consistency marks 1, tripping its gate threshold. On the raw minimum this would read GAPS, but the contradiction gate overrides it: an agent that catches a feed lying about price or stock distrusts the whole listing, not just this field. weakest signal 1/1 · 6 of 7 signals clean.

Fix firstFeed-vs-page consistency - resolve every contradiction so the feed, the live page, and your schema agree on price, stock, title, and ids.

A clean sheet didn’t save it. The verdict is the weakest signal — and a top trigger is dispositive on its own — so it names the one thing to fix first. Grades feed and listing readiness for AI shopping agents from marks you set - not a guarantee of placement, ranking, or purchase by any specific agent or platform.