CUSTOMER STORIES
Thomson Reuters Improves Customer Support with an AI Data Analyst Agent Built on Snowflake
With Snowflake and Cortex Analyst, Thomson Reuters built a secure AI data analyst agent that helps customer support teams automate case analysis and unearth insights faster.
Industry
Advertising, Media & EntertainmentLocation
Toronto, CanadaStory highlights
- Building a reliable foundation for AI: With a single source of truth on Snowflake, Thomson Reuters powers its AI initiatives with both structured and unstructured data that the company knows is accurate and secure.
- Delivering a superior customer experience with AI agents and Cortex Analyst: Using Snowflake Cortex Analyst and its APIs, Thomson Reuters created a secure AI data analyst agent that helps customer support teams automate case analysis and deliver faster insights.
- Expanding efficiencies across the enterprise: After seeing strong adoption and results with the customer support team, Thomson Reuters plans to roll out customized AI agents across finance, marketing, HR and other business units.
Video Transcript
Rittika Jindal: "The AI strategy is incomplete without data strategy. So we need to ensure that we get the data right, leveraging Snowflake to build our AI strategy. My name is Rittika Jindal. I'm a principal engineer with Thomson Reuters.
For over 100 years, Thomson Reuters has been a trusted global information provider company. Our main mission is to build content automation, workflows, and help professionals to know today and navigate tomorrow. Back in 2021, Thomson Reuters started working with Snowflake to build an enterprise data platform. The idea was to build a one single source of truth for our entire structured and unstructured data.
For you to build any AI system, you need to have the data right, and that's where we leverage Snowflake's LLM function to preprocess the data so that our AI system has got the right data. Now, we use Cortex Analyst to analyze the data, and we leverage Cortex Analyst APIs to integrate into our agent, which helps us in generating the accurate sequel and deliver the insights. You want to ensure your agent is secured, right? So that's where we leverage Snowflake's RBAC roles and policies to secure the data.
So our customer support team is using the AI data analyst agent. Before this agent, our customer support team would read each and every case. They would try to read the description. So there were humans reading out these descriptions to understand our customers' pain points. But now with the agent leveraging Snowflake's Cortex capabilities, we are able to automate this whole solution. Now our customer support team is able to deliver insights at a much faster pace.
Well, we have heard really good reviews from our customer support team about this agent, and we are really excited to launch it to other business units. We are also excited about Cortex Agent, which has the capability to navigate between different tools, so it helps us navigate the structured and unstructured data.
Our plan is to roll out the AI Data Analyst Agent to various businesses across Thomson Reuters. So the idea is to build customized agent for each business unit. For example, you have an agent for a finance team, you have an agent for a marketing team, and you have an agent for HR team. We have been working together with Snowflake for over four to five years and the partnership has been amazing. We really look forward to build more solutions and leveraging Snowflake's new capabilities."
