What I learned from consulting

What changed?

Sparker.ai now has a defined first offer: the AI Product Decision Sprint.

That is not proof that the market wants it. It is a hypothesis specific enough to test.

Why does it matter?

A consulting practice can become as assumption-heavy as the products it advises. It is easy to spend months polishing a service menu, publishing broad ideas, and interpreting attention as demand.

Sparker.ai should meet the standard it asks clients to use: define the decision, limit the investment, collect evidence, and stop when the evidence says stop.

This week’s Project

Hypothesis: product and technology leaders will pay for a short engagement that helps them decide whether and how to pursue one AI opportunity.

Experiment: publish the offer, speak with at least five qualified leaders, ask for a paid pilot, and record the objections.

Investment cap: 90 days and 20 prospect conversations.

The evidence

The offer and evaluation criteria exist. Market evidence does not, not yet.

That distinction matters. Preparation is an input. Buyer behavior is evidence.

The current decision

Researching. The next step is to test whether the problem feels urgent in the buyer’s own language and whether a clear decision is valuable enough to fund.

The practical meaning

When your organization says an AI initiative is “strategic,” ask what evidence would cause the team not to invest. If there is no acceptable answer, the initiative is not yet being managed as a decision.

The invitation

Update: I have recorded this as a failed project, closing September 30, 2026. What I learned is in the project record. If you are building your own decision system, email me with the hardest stopping rule you have had to set.