Sell the Outcome, Not the Way to Get It

14 November 2026 · agents

Humans are not the right person to interact with an execution engine. A CEO is not the right person to interact with the most junior engineers. The CEO gives strategic direction, and the professionals do the work, validate each other’s work, and innovate.

That is how I think about what I am building. I am working on an engine — software that uses an AI agent to build things — and I want to sell it to people who use it to create outcomes for their business. What the agent builds is called an artifact: a finished piece of work. And I think we have been thinking about it wrong. We should sell the outcome, not the way to generate the outcome.

Right now we ask people to do a lot of steps before they get anything. First find the right harness — the setup the agent runs in — then change it until it works. Even then you are not sure it will create the right outcome. The outcome can be bad, and then you have to go back and fix everything, and update the artifact, until you feel it is done.

At any of those steps, people say they have seen enough. They would rather just use the tool they already pay for, the one they already have a subscription to. There has to be a very strong reason for them to use your product instead.

I tried Muse myself, and the whole motivation was that it was free: a free computer to run it on, and a million free tokens a week. Who doesn’t like free? It was a hook. And if it had launched with pricing, nobody would have used it, because we already had a lot of alternatives. Free gets you tried. It does not get you kept.

What I want with my product is to remove those friction points, so that value is generated by the product itself. I don’t want to just create something and expect users to create the value separately. The product should do that work.

Next time you pick up a tool, ask which one you are buying: the outcome, or the way to get it.

To hear about new posts, follow me on LinkedIn or subscribe to the RSS feed.