Why Is Coveo Targeting Unreadable Product Data?
Coveo, the Montreal-based "Experience AI" search and discovery vendor, acquired Anglera, a San Francisco AI company backed by Y Combinator, on October 29, 2026, folding Anglera's agentic enrichment engine into its Commerce offering. The deal responds to a concrete problem: buyers increasingly research purchases through AI assistants, but the product data those assistants need is often missing or buried in unstructured documents [1].
Coveo frames the acquisition as closing a loop rather than adding a bolt-on feature. The company's existing platform already interprets buyer intent through search and behavioral signals.
What it lacked was a systematic way to fix the underlying catalog when that data turned out to be incomplete. Anglera supplies that layer, reading manufacturer PDFs, spec sheets and supplier feeds the way a human analyst would, then standardizing the output against each customer's catalog schema [1].
Terms of the acquisition were not disclosed. Anglera's founders, Amay Aggarwal and Ray Iyer, are Stanford alumni who previously built catalog AI systems at Uber Eats. The deal brings Anglera's existing enterprise customer base, including multiple Fortune 500 companies, into Coveo's orbit [1].
Anglera's Engine Traces Every Attribute Back to Its Source
Anglera's core design choice is traceability: every enriched attribute keeps a reference to the document or source it came from, which is what lets an enterprise buyer trust an AI-generated catalog entry instead of treating it as a guess. That grounding was the reason Coveo's product team says it picked Anglera over an internal build or other enrichment vendors [1].
“Unlike solutions that rely primarily on images or web scraping, Anglera grounds AI-generated product information in multiple authoritative sources. This helps reduce guesswork and hallucinations, delivering the accuracy and trust enterprise customers need.”
The platform was built for catalogs with hundreds of technical attributes per SKU, the kind of complexity common in B2B distribution and industrial retail, where a wrong compatibility spec or a missing dimension carries the most cost [1].
Anglera's existing customers stay on the Anglera platform, now backed by Coveo's scale, while Coveo Commerce customers get access to the enrichment engine as an add-on, with deeper integration planned between the two systems [1].
How Big Is the Measurable Impact of Clean Product Data?
The business case behind the acquisition rests on one data point: a 30-day A/B test with an apparel retailer, comparing product pages with Anglera-enriched data against a control group [1].
That gap is the clearest evidence yet that catalog completeness, not just site design or ad spend, now directly shapes whether a product shows up in an AI assistant's answer.
The scale of the underlying problem is already visible industry-wide:
Digital Commerce 360 has already turned this into a scoring exercise. Its 2026 Top 1000 report added an "AI Commerce Rankings" layer measuring "bot friendliness": whether AI agents and agentic platforms can access and read a retailer's catalog data at all [3].
Coveo buying an enrichment company the same year a major retail research firm starts scoring catalogs on machine readability is not a coincidence; it is two signals of the same shift.
| Metric (30-day A/B test, apparel retailer) | Enriched vs. control |
|---|---|
| Organic landing-page traffic | +120% [1] |
| AI-referred visits | +270% [1] |
- NielsenIQ estimates missing product attributes can cut discoverability by up to 30% [1].
- Gartner found 54% of shoppers who used AI while shopping had to double-check what the tool told them, and 62% said the information wasted their time [2].
- An AI assistant that cannot verify a spec will often skip the product rather than guess, a lost sale a retailer may never see in its funnel analytics.
Is Product Data Enrichment Becoming Its Own Software Category?
Coveo's move is not an isolated bet. Gartner's own research has separately pushed digital commerce leaders to treat product data as something that must be actively optimized for AI agents, rather than a static catalog field, warning that consumers remain unwilling to hand AI full control of purchase decisions even as they lean on it heavily for research and comparison [2].
For commerce teams, the acquisition is a signal worth acting on, not just watching:
- Audit which SKUs are missing structured attributes (dimensions, materials, compatibility) that an AI agent would need to compare products confidently.
- Check whether product detail pages expose that data in a form a crawler or agent can parse, not only a human shopper.
- Treat unresolved or ambiguous attributes as lost revenue, not a data-hygiene backlog, since NielsenIQ ties incomplete data directly to a discoverability drop of up to 30% [1].
What Does This Acquisition Mean for AI-Powered Storefronts?
The Coveo-Anglera deal confirms that product data quality has moved from a back-office PIM concern to a frontline commerce metric, measured in AI-referred traffic rather than just search rankings. Retailers and B2B distributors with large, attribute-heavy catalogs face the most exposure.
Coveo's bet is that this becomes a distinct, monetizable layer of the commerce stack rather than a one-time cleanup project, and the AI-referred traffic lift from its own pilot is the number it will use to sell that case to its enterprise customers [1].
Sources
- [1]MarTech Series, Coveo Acquires Anglera to Make Product Data a Competitive Advantage in AI-Powered eCommerce — martechseries.com
- [2]Gartner, Gartner Survey Finds Consumers Want AI Shopping Help, But Not AI Purchase Decisions — gartner.com
- [3]Digital Commerce 360, Digital Commerce 360 and ReFiBuy Introduce First-of-Its-Kind AI Commerce Rankings, Expanding the 2026 Top 1000 With a New Benchmark for Retailer Readiness in AI Shopping — digitalcommerce360.com
