Enterprise AI runs on context: A fireside chat between Nucleus Research and Strategy’s Chief Product Officer

Every competitor can buy the same AI models. None of them can buy your context.

AI agents rarely fall short because of the model. They fall short because they don't understand how your business defines its metrics, rules, and processes. In this session, Alex Wurm, Principal Analyst at Nucleus Research, shares new research on what makes AI agents consistent, drawn from interviews with Strategy Mosaic customers. Then he sits down with Saurabh Abhyankar, Executive VP & Chief Product Officer at Strategy, for a fireside chat on what they're seeing across the market and what data teams should do next.

What you'll learn:

• How the agent loop works, and why the model is only one part of what makes an agent reliable

• Why curated context outperforms accumulated context, and how competing metric definitions undermine agent accuracy

• What governed context delivered for Strategy Mosaic customers: up to 21% faster time to insight, 5–9% faster case resolution, and profit lift in target markets

• How a semantic layer reduces token and compute costs

• Where the semantic layer fits in your stack, and how BI teams can shift their focus to curating semantics

What we'll cover:

• The agentic loop: how LLMs, tools, and guardrails work together inside the harness

• Context engineering: why more documents, history, and memory can lower accuracy, and how ungoverned definitions produce three confident, conflicting answers

• Customer outcomes: Nucleus Research interviews with a multinational beverage company and a luxury automotive manufacturer

• The cost equation: how removing definition guesswork cuts wasted tokens and wrong analysis

• The road ahead: where the semantic layer sits in the stack, how it extends beyond analytics, and what it means for BI and data roles