Session brief
Small Bets, Big Impact Building GenBI at a Fortune 100 – Asaf Bord, Northwestern Mutual
Asaf Bord , Northwestern Mutual
Overview
This talk discusses the development of GenBI, a Generative AI-powered Business Intelligence agent, at Northwestern Mutual. The core thesis is that by taking small, incremental bets and focusing on building trust through transparency and gradual delivery, large organizations can successfully innovate with GenAI despite inherent risk aversion. The approach emphasizes using real, messy data from the outset to bridge the gap between prototypes and production.
Who should watch
- AI Engineers and Product Managers exploring enterprise GenAI adoption.
- Builders seeking strategies for integrating AI into risk-averse industries.
- Those interested in practical approaches to data democratization and business intelligence.
- Individuals facing challenges in balancing innovation with stability in large organizations.
Key takeaways
- GenBI fuses Generative AI with Business Intelligence to enable users to answer data-driven business questions without relying on a dedicated BI team.
- A "crawl, walk, run" approach was used, starting with BI experts, then business managers, before considering executives, to build trust and gather feedback.
- Using actual, messy company data from the beginning, rather than synthesized data, was crucial for understanding real-world complexities and ensuring production readiness.
- Building trust with leadership involved a phased, incremental delivery process with clear deliverables and the ability to halt investment at any stage.
- Initial phases focused on research, understanding metadata, and multi-context semantic search, providing immediate business value and input for other initiatives.
- The architecture includes agents for metadata understanding, retrieving existing reports (RAG), generating SQL queries, and a final BI agent to synthesize answers.
- Automating report retrieval with the RAG agent freed up approximately 20% of the BI team's capacity, equivalent to two full-time roles.
- The project's incremental nature allowed for early productization of components like the RAG agent and metadata enrichment, demonstrating tangible value and de-risking further investment.
Notable quotes
*GenBI is basically an agent that helps people answer business questions with data like a business intelligence person would do in real life.*
*The reason that we're pursuing GenBI is really because of the data democratization that it can bring.*
*Stability is something that's very important for us because it's important for our clients. So how do we balance stability with innovation?*
Unofficial community note. Prefer the recording for nuance.