The Bangalore-based company’s neural search will offer personalised recommendations by understanding users’ conversational and open-ended queries. The search feature is built using a Large Language Model (LLM), trained to understand terminologies related to dishes, recipes and restaurants, resulting in a more efficient search experience for consumers.
“Through a meticulous two-stage process, the model has been fine-tuned to respond accurately to relevant food-related queries, in real-time. These capabilities have been developed in-house, giving us greater control over the product, faster iteration time, and the flexibility to adapt to changing market trends,” Madhusudhan Rao, chief technology officer at the food delivery platform, said in its blog.
Neural search will also support voice-based queries as well as queries in specific Indian languages.
This, according to Rao, aims to make the platform more accessible to users speaking different languages. This feature will be in pilot by September and, based on the outcomes, it will be widely implemented.
ET had earlier reported on Zomato experimenting with generative AI to smoothen customers’ ordering experience.
Discover the stories of your interest
Other Swiggy platforms
Swiggy plans to extend its neural search feature to Swiggy Instamart, a grocery delivery platform. This aims to help users discover groceries in a more intuitive and efficient manner.
Similarly, generative AI will be used to transform Swiggy Dineout, the restaurant discovery platform, by helping users decide dining options based on their preferences, ambience, ratings and cost, among others.
“In our pursuit of enhancing customer service, we are collaborating with a third party to develop a GPT-4 powered chatbot. This bot will enable efficient and empathetic service to oft-asked customer queries, ensuring a seamless and delightful experience,” Rao said.
The platform is also piloting in-house tuned LLMs to allow restaurant partners to handle processes and questions related to onboarding, ratings and payouts leading to faster issue resolution and streamlining.
Swiggy intends to use neural search to make decision-making “a fun and effortless process” for its users.