In the next wave of Generative AI in the enterprise, the prompt is not the game, the business outcome is. And it’s with an Agent aligned to every business process, that the next level and business outcomes are going to be reached. Agentic AI aligned to business process requires a very different approach.

I’ve been doing customer work on this for a while now and building Agents is a different kind of game. With pressure on proving value fast, I think an agile yet simple approach is essential too. I’ve put a few slides together based on my recent experience and my general thinking on the topic.

NOTE: The thinking is still rough around the edges. Still working on an optimal solution design framework that can be used to guide and capture what needs to be built. I’ll share that separately.

Summary

• Precursors to Success: It’s good to have guardrails in place that guide you, i.e. what will help determine success. They include having engaged customers and people, focusing on business outcomes, ensuring technical capability, removing blockers, and adopting the right framework.
• Workshop Schedule: A one-day envisioning workshop includes sessions for capturing business scenarios through brainstorming, idea selection and deeper planning for the Copilot Agents. It emphasizes openness, exploration, and reviewing outcomes.
• Implementation Process: The one-day workshop will only get you so far. You need a process for building out something that can be executed in test mode and evaluated over a sufficient time period. Iterative development is key, with prototypes carried over into longer cycles if proven in early stages. Saying no is also important, e.g. include a go/no go stage gate step along the way.
• Mindset and Funnel Management: Contrasting project and product manager mindsets emphasise continuous evolution and learning in the latter. And then always be discovering with a four-step process for managing potential ideas: Discover, Prioritise, Build and Realize Value.
• Prioritisation: A simple prioritisation matrix for different scenarios is necessary using simple criteria for deciding what to focus on. Because you can’t boil the ocean.

Here are a few slides that go into a little more detail and include some visuals.