In this episode of the AI Agent & Copilot Podcast, Shawn Dorward, COO Innovia Consulting, explains why moving from prompts to outcomes means giving Microsoft Cowork a clear objective, boundaries, and definition of success so it can determine the work plan.

The Podcast is hosted by Dynamic Communities, the home of Summit NA!

Click here for the full video or the audio file from Spotify.

Key Takeaways

• The shift is from prompts to outcomes. Cowork works best when you stop thinking in terms of one prompt at a time and start giving AI an assignment. Instead of asking it to complete a single task, you define the business outcome you are trying to achieve and give it room to analyze the situation, connect the dots, and produce the work needed to get there. That is the shift I talk about in Moving from Prompts to Outcomes with Microsoft Cowork.
• An output is not an outcome. Asking AI to build a presentation, write a report, or analyze a spreadsheet is still just asking for an output. The outcome is the reason behind the work. What decision are we trying to make? What problem are we trying to solve? What should be different when the work is finished? When Cowork understands that context, it can do much more than simply create the thing you asked for.
• Good delegation doesn’t mean describing every step. One of the most important skills with Cowork is learning how much direction to give. You want to be clear about what needs to be accomplished, why it matters, what authority AI has, and what you expect back—but you don’t necessarily need to tell it exactly how to do every part of the job. If we prescribe every step, we’re not really delegating. We’re just turning a long prompt into a checklist. That’s also the idea behind Delegating Work to AI: Stop Prompting, Start Asking.
• Start treating AI more like a coworker. This is probably the easiest way to understand the difference. If I delegate something to a capable coworker, I don’t normally explain every mouse click they need to make. I explain what I’m trying to accomplish, make sure they understand the assignment, review their approach when necessary, and give feedback along the way. Cowork starts bringing that same model to AI. You define the outcome, let it propose or execute the approach, and then guide it as the work develops.
• The feedback loop matters more than the perfect prompt. As AI gets better at actually doing the work, our job increasingly becomes providing clear direction and good judgment. The skill isn’t going to be writing the world’s longest or most complicated prompt. It’s understanding what work needs to be done, why it matters, what good looks like, and when to step in with feedback. The clearer we are about those things, the more room we can give AI to figure out the path.

The big takeaway for me: The future of working with AI isn’t about becoming better at telling it every step to take. It’s about becoming better at delegating the right work, defining the right outcome, and judging whether the result actually gets us there.

Click below to listen to the Spotify audio version

Thanks for reading!

Shawn Dorward

Microsoft MVP, Business Applications

LifeHacks365.com | LinkedIn | Twitter