Source: Unite.AI
A retailer can have the right product in stock and still lose the sale because its catalog does not describe what the shopper wants. That problem becomes more consequential when an AI assistant is doing the searching: a persuasive answer is only useful if the system can find an item that actually meets the request.
Algolia has acquired Velou, a New York-based AI company specializing in product data enrichment and catalog intelligence, to strengthen that foundation. Announced on October 6, the deal brings product understanding closer to the search and recommendation systems that use it.
The acquisition announcement says financial terms were not disclosed and Velou’s six-person team is joining Algolia. Velou already serves more than 50 retail and premium brand enterprise customers. Algolia says it handles nearly two trillion searches annually across more than 18,000 customers, giving the acquired technology a much larger potential distribution channel.
Why better search starts with better product data
Search systems must connect a customer’s intent with the information available about an item. If a listing has a photograph and a brief title but omits important attributes, there is less reliable information for a system to match against a detailed request.
Consider a shopper seeking a compact desk lamp with an adjustable arm. A catalog that records only the brand and color may leave the search engine unable to filter for size or construction. A richer record can make those constraints explicit. This is an illustrative example of the data problem, rather than a claim about a particular Velou deployment.
That same issue carries into agentic commerce, where AI systems help consumers discover and select products. Better language understanding at the assistant level cannot, by itself, establish an undocumented product characteristic. Improving the underlying record makes the evidence available to more than one interface.
What Velou adds to Algolia
Velou uses multimodal AI to turn product descriptions and images into structured information. In its announcement on the company website, Algolia describes mapping that information to a curated retail taxonomy: a consistent vocabulary for categories, attributes and their values. The company says the enrichment runs as catalogs change and ties attributes to source evidence.
Velou’s website identifies Commerce-1 as its retail-specific model and describes product graphs that map relationships between items. Alongside generating missing attributes and standardized tags, the platform is designed to provide context for alternatives and compatible products, then push enriched records back into existing commerce systems.
A taxonomy and a product graph address different parts of the discovery task. A common vocabulary helps systems compare items whose suppliers use different descriptions. Relationships help connect those items. For an assistant asked to assemble a coordinated purchase, both the characteristics of each product and the connections between them can matter.
Algolia plans to place Velou’s product intelligence beneath search, recommendations, personalization, merchandising and Agent Studio. It says existing APIs, SDKs and integrations will remain in place. Its stated aim is to improve the information those products work with while avoiding a rebuild of retailers’ existing Algolia implementations.
The customer results behind the acquisition
The companies point to two retail examples. Get The Label added more than 190,000 product attributes over six months using Velou, with a reported 60% increase in site-search revenue. Everything5pounds increased its catalog attributes by more than 85%, alongside a reported 33% rise in search conversions.
These are company-reported customer outcomes, rather than a controlled comparison establishing how much of each improvement came from enrichment alone. They nevertheless make the commercial objective concrete: better product descriptions need to translate into shoppers finding suitable products and completing purchases.
For a retailer evaluating the technology, the useful measurements extend beyond the number of new fields. Does a specific query return an appropriate result? Do filters reflect accurate product characteristics? Can a merchandising team inspect and correct the evidence behind an attribute? Those questions connect data quality to customer experience.
A strategic purchase for the AI shopping stack
The acquisition gives Algolia a way to influence an input that shapes the effectiveness of its wider platform. Enrichment can support the search bar, recommendation engine and shopping agent from the same product record, rather than requiring separate catalog repair efforts for each experience.
Accuracy remains central. A product photo may reveal a visible design feature, but a retailer needs dependable evidence for specifications and care instructions. Attaching an unsupported attribute could produce an attractive recommendation that disappoints the buyer. Algolia’s emphasis on inspectable evidence is therefore a meaningful part of its integration promise.
The deal also illustrates a practical consideration in implementing agentic AI in a business: the systems need usable information about the objects they act on. In commerce, that means a catalog detailed enough to support the shopper’s actual requirements.
Algolia’s next test is whether bringing enrichment and retrieval under one roof improves discovery consistently across its customer base. The ambition is straightforward: make the right product easier to find, whether the request arrives through a search box or an AI assistant.
