How AI is driving faster data-driven decisions at Grab

When a Grabber asks a question, Jarvis translates it into a SQL query to fetch the relevant data. 

Here’s what goes on under the hood:

• Jarvis analyses past queries, data schemas, and business definitions to understand the questions’ intent.
• It builds a tailored prompt that combines business metrics, table definitions, and examples.
• Using this prompt, Jarvis generates an accurate SQL query and retrieves the data from our data lake, a large repository that stores our raw data

The result? Grabbers get the exact information they need—without writing a single line of code.

Of course, Grabbers can retrieve only the type of data their role gives them access to. Jarvis understands the user’s data access privileges and denies queries that are out of scope.

2. Insights generator: from data to understanding

Jarvis doesn’t stop at delivering data. It helps teams interpret the numbers and uncover the “why” behind trends.

Whether it’s diagnosing a revenue drop or identifying growth opportunities. Jarvis reveals what drives trends, anomalies, and gives explanations, turning numbers into narratives that teams can act on. 

Real results across Grab

Today, Jarvis has already transformed how teams work with data. 

Our Merchant AI Assistant is powered by Jarvis and allows merchant-partners to check their sales performance and get tailored recommendations. Questions like “How can I increase my sales next week?” are answered in seconds, helping merchants optimise their strategies.