Make products easier to find, compare, and buy by completing the catalog details that search engines and AI assistants need.
ReFiBuy Agentic Commerce Optimization
Products stay eligible, competitive, and actionable when an AI shopping agent does the researching, comparing, and buying.
A closed loop that scores, enriches, distributes, and monitors every SKU for the AI shopping surfaces where agents now pick products.
What the agents do
Scores every SKU against seven readiness dimensions, drafts sourced enrichment, and runs shopper prompts to find and fix gaps.
- Built by
- ReFiBuy, the team behind ChannelAdvisor
- Contributed by
- ReFiBuy (Agentic commerce optimization platform)
- Industry
- Apparel & Footwear, Retail
- Audience
- B2B, B2C
01
The problem
Why this exists
Product feeds were built for filter navigation: price, color, size, and a product identifier. An AI shopping agent asks a different kind of question. A shopper planning a rocky 50k trail race wants to know about terrain, cushioning, and foot support, and a feed that carries only the standard fields cannot answer any of it. When the engine cannot understand the product, the product is not in the conversation.
Read more
Commerce teams spent two decades stripping product data down to what a search index and a filter menu needed. Agentic shopping rewards the opposite: enough content and context for an agent to reason about why this product, for this shopper, now. Readiness is not a publishing problem. It is a data governance problem, and it sits upstream of the answer engine and the search result.
02
Outcomes
What changes when this is in place
- Visibility, win rate, top-three placement, and content accuracy are measured per SKU on the AI digital shelf, not estimated.
- Monitoring runs the full catalog against roughly 100 live prompts per product, and every gap links to a specific fix.
- Steve Madden credits the platform with accelerating its understanding of how GenAI reads its catalog and of its readiness for AI shopping experiences.
Read more
- Evaluate and Monitor are available today, with Amazon's Alexa for Shopping monitored in production.
03
How it works
Architecture
The platform runs one loop over the catalog a brand or retailer already has. Six stages repeat rather than rewrite once. Ingest reads the existing catalog from the systems of record in whatever format it arrives, so nothing is re-platformed. Evaluate scores every SKU from 0 to 100 against seven weighted readiness dimensions, from identity and discoverability through conversational readiness, visual assets, and trust signals, and attaches a specific fix to each shortfall instead of a grade.
Read more
Enrich writes the agentic layer over the standard fields: attributes, titles, descriptions, and question-and-answer content, each with its source, drawn from the brand's legal, go-to-market, and policy material the agents are trained on. The commerce team approves every change before it moves. Distribute renders approved data in the format each shopping surface expects, and Sync writes the approved changes back to the system of record so the catalog and the engines do not drift apart.
Monitor closes the loop. It runs shopper-like prompts across engines continuously, roughly 100 live prompts per product, and scores each product on visibility, win rate, top-three placement, and content accuracy. Every gap links to the exact fix, which re-enters at Ingest and makes the next pass sharper. Agents do the scoring, the drafting, and the gap analysis; the commerce team keeps approval, and the systems of record keep the master data.
Tech stack
Salsify and Syndigo product dataShopifyAmazon SP-API listing ingestionChatGPT OpenAI product feedGoogle Merchant Center and UCPSchema.org JSON-LDCSV, JSON, and SFTP feeds
Resources
Next step
Request an intro
ReFiBuy is behind this solution. Ask about getting your catalog ready for agentic shopping, and the MACH Alliance will put you in touch with the person who can answer.
Not ready to talk to anyone? MACH Insider is the Alliance mailing list. New solutions, program openings, and the research behind them, once a month.
Everyone credited on this solution
- ReFiBuyAgentic commerce optimization platform
Every organization named here has a page listing the rest of its work on the Hub.
Related solutions
100 product descriptions in 30 minutes, against 100 a day by hand.
Matalan Product Description Generation
Product recommendations arrive with a reason a shopper can read, not only a score.
IPSY Customer Preference Knowledge Graph
AI Exchange Hackathon
Verification
Contributor submission
The detail on this page comes from the contributor's own account of their own work. The MACH Alliance reviews a submission before publishing it and does not independently verify the figures in it.
What this page does not establish
- No customer-measured uplift in visibility, conversion, or revenue is published.
- Enrichment and activation are design-partner extensions, not generally available stages of the loop.
Where this came from
Solution detail is drawn from ReFiBuy's submission to the hub: its solution brief, product brief, and platform overview.
Status: Staged rolloutAdded Oct 30, 2026A listing is an editorial decision, not a MACH Alliance certification.
Built something worth showing?
Put your agentic solution in front of the enterprise leaders who use the hub to decide what to build next. Submissions are open to MACH Alliance Members, Collaborators, or Ambassadors.