Why Can't Anyone Answer Questions About the Business? — Garrett Galow, WorkOS

2026-06-11 · Source: AI Engineer · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Intermediate, long

Summary

WorkOS developed an internal tool called Studio, an LLM-powered workspace designed to help non-technical employees answer complex business questions and build custom dashboards. Studio integrates with various data sources like Snowflake, Linear, and Notion, using an agent (Lane Graph) tied to an LLM (Opus) to parse natural language queries, understand schemas, and generate SQL. It can then create reusable JavaScript-based "widgets" that encapsulate UI, APIs, and queries, providing live, interactive data. This system addresses the common problem of rigid dashboards and the need for engineers to manually answer one-off data requests, enabling self-serve analytics for teams like marketing and support.

Key takeaway

For Directors of AI/ML struggling with rigid dashboards and manual data requests, WorkOS's Studio demonstrates a powerful approach to democratize business intelligence. By enabling self-serve query generation and reusable widget creation via an LLM agent, you can significantly reduce engineering overhead and empower non-technical teams to gain immediate, validated insights from diverse data sources. Consider adopting similar internal tools to boost productivity and data accessibility.

Key insights

WorkOS's Studio uses an LLM-powered agent to answer internal business questions and build reusable data widgets from diverse data sources.

Principles

• Sequence agent actions with pre-flight checks and context injection.
• Layer prompts with defaults, org rules, and distrust product knowledge.
• Validate generated queries and data returns before deployment.

Method

Studio takes natural language questions via Slack or dashboard, parses them, and an LLM-powered Lane Graph agent queries integrated data sources (Snowflake, Linear, Notion) to generate answers or reusable JavaScript widgets.

In practice

• Analyze content effectiveness for new team sign-ups.
• Enable support teams to self-serve customer session data.

Topics

• WorkOS Studio
• LLM Agents
• Business Intelligence
• Data Integration
• Internal Tools
• Snowflake
• Data Governance

Best for: AI Engineer, Data Scientist, Director of AI/ML

Related on AIssential

• AI Builders: Building a text-to-SQL agent — Text-to-SQL agents democratize data access by translating natural language queries into structured database commands.
• A new era for software testing — LLMs can significantly enhance software QA and testing by automating complex, traditionally manual processes without compromising quality.
• d0: Vercel's data analysis agent — Vercel's DZ agent shows 100x growth in PR merges, indicating a fundamental workflow shift.
• Clio Rolls Out Agents For Work and Vincent — Agentic AI in legal tech automates multi-step workflows, shifting from task execution to outcome-based delegation.
• How Bloomberg is Using Agentic AI for Complex Workflows — Bloomberg's agentic AI integrates market data with proprietary firm knowledge to enhance financial analysis and investment workflows.
• The Business Engineer Agent OS — The business engineer agent harness integrates multiple LLMs for practical application within a CLI.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.