Planning to order from GrabFood but unsure of what a dish contains or if it suits your taste?
That’s why accurate descriptions of menu items matter. They provide essential details about its ingredients, texture, and flavour profile. Plus, they can nudge consumers from browsing to ordering.
Yet, we found that more than half of the menu items listed on GrabFood didn’t have descriptions.
This is understandable because crafting these perfect little stories takes time and resources that not every restaurant has. They’ll include a simple title, like “beef burger” or “vegetarian fried rice’, but stop short of giving more detailed descriptions to make the dish stand out.
We decided to try using GenAI to change that. With AI-powered menu descriptions, we’re helping consumers get to a suitable choice faster while driving sales for merchant-partners. The AI model has since generated descriptions for some 2.8 million menu items.
This feature is currently available to merchants in Singapore, Philippines, Thailand, Myanmar, and Vietnam, supporting languages like English, Burmese, and Vietnamese.
The AI tool we developed uses large language models (LLMs) to generate succinct descriptions for menu items listed on GrabFood. The AI takes into consideration the merchant’s name, cuisine type, dish name, and even dish images. It then automatically generates short and snappy text.
When we first rolled out this feature to improve menu description coverage among our merchant-partners, we identified a list of restaurants that had little to no food descriptions. We then offered to automatically generate and upload AI-crafted descriptions for their menu items. Many restaurants were quick to seize the opportunity to enhance their menus.
Today, the feature is also integrated into Grab’s merchant app, which means restaurants can generate descriptions when they’re listing a new food item.
Merchant-partners can personalise these descriptions by providing the AI model with keywords and phrases they would like to include. But the AI can also write them from scratch. It will also recommend alternative descriptions for existing food items to help merchant-partners optimise their menus with minimal effort.
To ensure the accuracy of these AI-generated descriptions, we conducted trials with sample outputs in various languages across different markets. Our local operations teams evaluated these descriptions, providing scores that helped us tweak our AI models for each region. This process also allowed us to establish and maintain consistent quality standards.
Early experiments among merchant-partners in Singapore also yielded positive results. Most merchant-partners who utilised the AI-generated descriptions tool saw an improvement in menu to checkout conversion rates.
The work doesn’t stop here. We are taking major steps to further improve the quality of menus on GrabFood. We are planning to use LLM agents as quality reviewers in the process. These are LLM-based programmes that can be tasked with evaluating descriptions based on various benchmarks such as clarity and accuracy. It’s an exciting way to improve the quality of AI-generated text—also with AI.
We hope that the AI-generated food descriptions will enhance the food discovery experience for consumers on GrabFood, make life easier for merchant-partners and drive sales.
3 Media Close,
Singapore 138498
GrabFood delivery-partner, Thailand
GrabFood delivery-partner, Thailand
COVID-19 has dealt an unprecedented blow to the tourism industry, affecting the livelihoods of millions of workers. One of them was Komsan, an assistant chef in a luxury hotel based in the Srinakarin area.
As the number of tourists at the hotel plunged, he decided to sign up as a GrabFood delivery-partner to earn an alternative income. Soon after, the hotel ceased operations.
Komsan has viewed this change through an optimistic lens, calling it the perfect opportunity for him to embark on a fresh journey after his previous job. Aside from GrabFood deliveries, he now also picks up GrabExpress jobs. It can get tiring, having to shuttle between different locations, but Komsan finds it exciting. And mostly, he’s glad to get his income back on track.