Conversational Search in Flipkart

Submitted Sep 11, 2023

Abstract

With Conversational Commerce becoming more popular, the way people search for products on a Conversational (Chat) Interface is very different from the way they search on a traditional “Search Text Box” based interface. This talk is around the differences, the challenges they pose and the solutionn to those challenges based on our learnings in Flipkart.

Broad Outline

Conversational Search is different from Traditional Search for the follwoing reasons.

Long, descriptive and subjective queries - Eg. “Show me best camera phones with good battery life”. This is very different from traditional search queries which are keyword based. Eg. “5g phones”
Suffer from Articulation gap - Eg. "Show me phones for daily use for my father. Here, we will need to understand what features of a phone are required for daily use and for elders. These queries are not performed on traditional search methods and its not natural.
Lots of Comparison queries - Eg. “is vivo v15 better than oppo k10”.
Required high relevance - Since lesser results are are shown in a chat based search, the results needs to be extremely relevant.

We used LLMs for solving for articulation gap and query re-write. We also used different labelling techniques such as using LLMs to label the results rather than relying on metrics such as CTR to get high quality labelled results to train our models.

We will cover all of the above in the talk.