Source: Fortune
Summary
Bloomberg’s new “AskB” feature, which uses natural language to navigate the terminal’s features and functions, is a major rethink of the terminal in Bloomberg’s history. AskB acts as an agent, building investment screens and producing full research reports, including sophisticated financial modeling and bull and bear cases for a particular stock. Shawn Edwards, Bloomberg’s chief technology officer, shares lessons on how Bloomberg built AskB, including the importance of data, building robust evaluations, using multiple models to contain costs, and proactive agent-to-agent workflows.
Our Reading
The numbers tell one story. Bloomberg’s AskB feature is a response to the threat from AI-native startups. The company is using AI models to ingest and analyze complex data sets, making accessing features more intuitive. Edwards says customers can’t replicate this by using AI models themselves due to the high cost of data ingestion and validation. Bloomberg’s approach is to build robust evaluations, use multiple models to contain costs, and create proactive agent-to-agent workflows. The company is taking the potential threat from AI-native startups seriously.
Original observation: The terminal’s future is not just about accessing data, but about navigating the complexity of that data.
Author: Evan Null
