Buy Lily AI if you are a B2C fashion, beauty or home brand chasing ROAS and AI-search visibility from an existing feed; buy Anglera if your catalog runs on supplier documents and your data model itself is the gap.
Both claim to enrich product data. This page is about where that claim stops.
With an enrichment-AI tool the question is what happens after the first pass: Lily Max tests consumer-facing language against a holdout on the surfaces it covers, while the record underneath still has to be completed against a supplier document and kept true as the catalog changes.
Mine every spec from every source.
Every value traced to a document you can open. The catalog is only as honest as what it was built from.
Aim the catalog at the buyer who actually buys.
Grounded data still loses if it answers questions nobody asked. Alignment is what turns specs into conversion.
Product data is a practice, not a project.
Markets move, suppliers reissue, buyers change what they ask for. A catalog that is right in March is wrong by August unless something is watching.
Scored against public documentation. Grouped by the three acts — so you can see which ones Lily AI leaves on your desk.
Feeds, PDPs, schema, product images; no supplier documents
PDFs, spec tables, drawings, manuals, images, sites
Applies own 20k-term library; no per-customer field proposals
Proposes fields your schema never had
20k+ term retail taxonomy; synonyms collapsed, human-curated
Normalizes and governs allowed values, versioned
Retail taxonomy classification; maps to Google, Meta, Amazon
Auto-classifies; channel and marketplace mapping
Review queue and lift evidence; no value-level source
Every value cites its source doc and page
Consumer language tuned per surface; B2C retail only
B2B specifier and B2C shopper enriched differently
Shopper searches, clickstream, surface performance feed back into enrichment
Reviews, search, competitor rails, social — fed back
Titles, descriptions, meta tags, Q&A, schema markup
Original copy per persona and channel
Computer vision reads images; outputs alt-text, not images
Generates studio-grade imagery for photoless SKUs
Always-on loop re-enriches and retests; human approves deploys
Re-enriches on its own after go-live
Scores completeness, correctness, compliance; gap detection, holdout lift tests
Scored against your standards; nothing publishes below bar
Publishes to GMC, Shopify, Algolia, Bloomreach, PIM connectors
Writes back to PIM, ERP, warehouse, commerce
No-code connectors; no public API docs or MCP found
API, webhooks, and MCP servers
Software agents enrich; customer sets goals, approves deploys
Anglera owns the work; review is a guardrail
“Buyer signals” is the emptiest phrase in this category, so here is the literal thing. Each of these is an observation from a live market, the gap it exposes, and the field that gets created as a result.
"Ordered my usual 10D and couldn't get my foot in. Exchanged for the 10EE and it's fine — the listing never says which width you're getting."
Width lives in the SKU suffix or in prose, so it never reaches the feed, the filter, or an agent asked for a wide 10.
shoe_widthUS width grades: 2A | B | D | 2E | 4E | 6E — one value per SKU, taken from the last, not the copy"fragrance free" returns eleven SKUs; "unscented" returns two; "no fragrance" returns nothing — and three of the eleven list parfum in the INCI.
Scent status is being read off marketing copy rather than the ingredient list, so shoppers and answer engines get a different catalog depending on which word they used.
fragrance_statusfragrance_free (no fragrance/parfum in INCI) | unscented (masking fragrance present) | lightly_scented | scented — resolved against the ingredient listTech pack: hand-tufted, 100% New Zealand wool, pile height 0.43 in, face weight 1,800 g/m². PDP: "low-pile wool rug, great under doors."
"Low pile" is a claim; 0.43 in is a number a shopper can hold against a door clearance. It exists upstream and stops at the tech pack.
pile_height_indecimal inches to two places, paired with construction: hand-knotted | hand-tufted | machine-woven | flatweave | shagLily Max starts from what you already have — Lily says it "works from your existing product feed and catalog data, the same titles, attributes, and descriptions your feed and PIM already hold." The loop detects gaps, generates enrichment in consumer language, and, in Lily's words, "measures every change against a control," scaling what wins. That question is which language performs. Then the catalog moves: a supplier swaps an outsole compound mid-season, a skincare line reformulates and drops an active, a rug is rewoven at a different pile height. The winning title reads exactly as it read in the test, and the surfaces still report lift. Nothing failed. Whether the record still matches the tech pack is a separate question, asked against the source document rather than against a holdout. Completing the record — one value per SKU, traceable to a spec — is what keeps tested language true after the test.
Values are normalized into a governed, versioned set of allowed values — so a filter works, and keeps working after the next import.
3/4 in0.75"3/4"19mm3/4 inchDN200.75 in (DN20)Six suppliers, six spellings, one physical size. Filters only work once they agree.
BlkblackBLACK MATTEMatte BlkRAL 9005Black — MatteFree text makes a colour filter useless. A governed value makes it a facet.
SS316316 StainlessStainless Steel 316A4 StainlessStainless Steel — 316 / A4Same alloy, four vocabularies, plus a trade name. Buyers search all of them.
Filling the fields you defined has an invisible ceiling: a catalog can hit 100% complete and still miss the attribute that loses the sale, because completeness is measured against a schema someone drew years ago. Schema Foundry reads competitor listings, buyer searches, review complaints and your supplier docs, and proposes the fields you never defined — which is where Lily AI stops.
How Schema Foundry works
Lily AI sells Lily Max, an agentic product intelligence engine that enriches retail catalogs so search engines, ad platforms and AI shopping agents can understand them. It reads a retailer's existing feed, PDPs, schema and product images, detects gaps, then generates consumer-language attributes, titles, descriptions and schema markup drawn from a 20k+ term retail taxonomy. Enrichments are A/B tested against a holdout and winners publish to Google Merchant Center, Meta, Amazon, onsite search and connected PIM/commerce systems. Founded 2015; named customers include Coach, J.Crew, UGG and Kate Spade.
Pricing: Undisclosed. The pricing page lists three quote-only tiers: Starter, Growth and Enterprise — all "book a demo for a quote". The FAQ states pricing "is based on your catalog size and the surfaces you choose, structured as credits," and is tailored to ad spend. A free 30-day trial on 500 products is offered.
Lily AI websiteConsumer fashion, beauty and home brands with a clean product feed whose goal is measurable Google/Meta ROAS and AI-search visibility, with lift proven by holdout testing.
We'd rather tell you here than in month three of an implementation.
Capability verdicts reviewed against Lily AI's public documentation on July 14, 2026. Vendors ship quickly — if something here is out of date, tell us and we'll correct it.
Bring one category and your supplier files. In 30 minutes you'll see it enriched — complete, structured, and consistent enough to launch on — plus the attributes your schema didn't have yet.
Book a demo