Building With AI Is Not the Same as Building About AI

There's a pattern I keep seeing in pitches, product launches, and client briefs: "We're adding AI to our platform." It's usually followed by a chatbot, a summarization feature, or some form of auto-generation bolted onto an existing product.

That's building about AI. It's the feature-level thinking that treats AI as a checkbox.

Building with AI is something different entirely.

The feature trap

Adding a chatbot to your SaaS product is not an AI strategy. It's a feature. And like most features, it follows a predictable curve: initial excitement, modest adoption, eventual irrelevance.

The problem with feature-level AI is that it doesn't change how people work. It adds a new button to an interface that still operates the same way it did before. The user's workflow is unchanged. Their mental model is unchanged. You've just given them one more thing to click.

Workflow-level thinking

The products that are actually winning with AI are the ones that rethink the workflow from scratch. They ask: if we had this capability from day one, would we have built the product the same way?

The answer is almost always no.

Consider how this plays out in practice. A project management tool that adds AI summarization is building about AI. A project management tool that eliminates status updates entirely because the AI already knows what everyone is working on — that's building with AI.

Why this matters for service companies

At Kilowott, we see this distinction constantly. Clients come to us wanting to "add AI" to their product. Our job is often to push back and ask: what problem are you actually solving? And is a chatbot really the answer?

Usually it's not. Usually the real opportunity is somewhere upstream — in how their users gather information, make decisions, or coordinate with each other. The AI isn't the product. It's the infrastructure that makes a better product possible.

The question to ask

Before adding AI to anything, ask this: does this change how someone works, or does it just add a step?

If it just adds a step — even a clever one — you're building about AI. And that's a race to the bottom, because everyone else is adding the same feature.

If it changes how someone works — if it removes steps, collapses decisions, or eliminates friction — you're building with AI. And that's where the real value is.