The Business Case for Applied AI: Why I Believe Organizations Should Start with Friction, Not Technology

Artificial Intelligence is part of almost every leadership conversation I’m having today.

Some leaders are evaluating Microsoft Copilot. Others are exploring enterprise AI platforms or looking at ways to introduce AI into their delivery organizations.

One observation has stood out.

Many of the leaders I speak with feel limited in what they can accomplish with Microsoft Copilot. Not because the technology lacks capability, but because they’re trying to solve business problems with a tool before they’ve identified where AI can create the greatest value.

The conversation often begins with questions like: Can Copilot help our teams? Which AI platform should we invest in? Should we build our own AI assistant? How do we govern AI across the organization?

Those are important questions. I just don’t believe they’re the first questions leaders should be asking.

Where is friction making delivery more expensive, less reliable, or less aligned with the outcomes your organization intended to achieve?

That question changes the discussion. Instead of focusing on technology, we begin talking about work.

Where are teams spending time on activities that add little business value? Where do decisions slow down because information is incomplete? Where does collaboration become expensive because work wasn’t sufficiently prepared before the meeting? Where does strategic intent become diluted as work moves from planning into execution?

Those questions are rarely about AI. They’re about how work gets done. And that’s exactly where I believe Artificial Intelligence creates its greatest value.

AI isn’t the investment. Improving the way an organization delivers technology is.

Technology is simply one of the tools that helps us get there.

The Wrong Conversation

Too many AI initiatives begin with platform selection. The discussion quickly becomes: Which platform should we standardize on? Which model performs best? Which licenses should we purchase?

Those conversations matter. But they don’t tell us whether AI is solving a meaningful business problem.

Before selecting technology, I believe organizations should identify where they are consuming expensive capacity, introducing unnecessary variability, losing visibility into delivery, or allowing strategic intent to drift as work moves into execution.

Once those friction points become visible, the next step is to decide how they will be measured before selecting the intervention. That is what turns an interesting AI use case into the beginnings of a business case.

Technology follows strategy. It should never replace it.

What Starting with Friction Looks Like

Consider something as ordinary as a product backlog.

Most organizations have established practices for reviewing whether backlog items are complete, properly structured, and ready for development. But that doesn’t necessarily tell leadership whether the work being prepared is still connected to the business objective that justified the investment.

A story can be perfectly written and still contribute to the wrong outcome. An epic can contain well-defined work and still represent only part of what an initiative was intended to deliver.

Over time, those small disconnects create something much more expensive than poor backlog hygiene. They create delivery drift.

That was one of the friction points I encountered while developing an AI-enabled backlog coaching capability.

The obvious application of AI was to evaluate story quality. But story quality wasn’t the most valuable problem to solve. The more consequential question was whether AI could help trace intent through the delivery chain and identify where alignment begins to break down.

More importantly, the problem could be measured before asking anyone to adopt the solution.

How long does it take someone to manually trace that alignment today? How frequently are structural or directional gaps discovered? When AI flags a potential misalignment, how often does that assessment match the judgment of an experienced practitioner?

And eventually, after someone acts on those findings, does the readiness of the backlog measurably improve?

Now we have the beginnings of a business case.

The value isn’t simply that AI can review backlog items faster. The value is creating earlier, measurable visibility into whether expensive delivery capacity is being directed toward the outcomes leadership actually funded.

That’s the distinction I mean when I say organizations should start with friction, not technology.

---

Ready to Find Where AI Can Create Value?

The AI Delivery Diagnostic helps technology leaders identify where delivery friction exists, where AI can have the greatest impact, and which opportunities should be prioritized first.

Take the AI Delivery Diagnostic