Bloomberg launches AskB, an AI agent that builds investment screens and research reports using natural language, drawing on consolidated financial data as its competitive advantage.

Bloomberg unveiled AskB at its London offices, an AI agent feature that lets users employ natural language to navigate the terminal, construct investment screens, and generate full research reports with financial modeling and bull and bear cases. The system combines models from Bloomberg and frontier AI companies like Anthropic.

Bloomberg's chief technology officer Shawn Edwards cited data consolidation as the critical differentiator: AskB pulls from Bloomberg News, sell-side research from over 800 providers, market data, and alternative datasets (anonymized credit card transactions, foot traffic from cellphone pings, satellite imagery, app usage data). Before Sweetgreen's fourth-quarter 2025 earnings call, alternative data signaled the chain would miss consensus forecasts, which it ultimately did.

Internal human evaluations by domain specialists—bond covenant experts, equity analysts, market structure experts, and journalists—are essential to building trustworthy AI agents, Edwards said, describing this emphasis as one of Bloomberg's biggest cultural shifts in the past two years. AskB routes queries to multiple models (commercial frontier models, open-weight models, and internal models) to balance cost and performance.

Customers pay approximately $30,000-per-user annual subscription price. Edwards noted that while some customers attempted to ingest alternative data and run analyses independently, most found it harder than expected and chose to focus on their own trading strategies instead.