LLMs as Rankers, a Blueprint (AI for Ecommerce Search)

In “What it takes to Build Effective Search”, we discussed why Ecommerce search relevance is hard to get right without significant investment. In this article, we dig into a low-effort productizable solution to this problem.

When a user enters a search query on a modern Ecommerce website, two things happen under the hood. First, all products in the catalog that share the same words as the queries are retrieved — this is known as (keyword) retrieval. Second, these products are sorted to show the most relevant items at the very top — this is known as ranking.

Click here to read the rest of this article on the Branch AI blog