Most AI agents don’t fail dramatically—they just fade away. They’re tried a few times and then abandoned. Others get used every day, but don’t really change how the team performs, or even add extra work. And only a minority of organizations say they’re seeing real efficiency gains from AI. Clearly, “we’re using AI” is not the same as “we’re getting value from it.”
This talk puts the spotlight on what most AI stories skip: focusing on what matters and measuring impact.
We’ll start by identifying real challenges, not cool ideas, and show how to design agents and reusable prompts that are tied to clear outcomes: time saved, improved day-to-day experience, better quality, reduced risk, or happier customers. I’ll introduce a simple, “flight plan”–style framework that any company can use to track the impact of their AI initiatives—across experiments, rollouts, and long-term adoption—so you can see what’s working and what isn’t.
You’ll see real-world agent examples from companies of different sizes: where they made a measurable difference, where they removed friction, and what teams changed in the way they work to unlock real value. You’ll walk away with a practical way to judge your own AI efforts: which problems are worth solving, what to measure, and how to keep only the agents that clearly improve outcomes for both the business and the people doing the work.