From Insight Projects to Insight Systems: Decision-Driven Agentic Workflows

Insights teams are moving fast toward agentic workflows. What started as pilots and experiments is turning into systems that run continuously, monitor signals, generate ideas, and push work forward without waiting for a brief or a human prompt.

Here is the problem. Once systems are always on, decisions are always being made. Most organizations have not designed for that reality. Judgment still lives in slide decks, hallway conversations, and last-minute reviews, while agents are left to operate on incomplete or implicit logic.

In this keynote, I will lay out why agentic systems represent a new operating model for insights, not just a faster way to run research. Using a simple Idea-to-Concept flow, I will show how swarms of specialized agents can work together to identify opportunities, develop concepts, and filter options, and why this only works when decision criteria are made explicit.

The core shift is this. Humans stop being reviewers of output and become designers of the decision system. Confidence thresholds, tradeoffs, escalation points, and guardrails are defined up front so agents know how to act, not just what to generate.

This session is for insights, analytics, and leaders who are past experimentation and now accountable for what AI systems actually do inside their organizations.