How I work
Evidence in. Useful work out.
I want to understand the business well enough to make a useful decision, then stay close enough to the work to see whether it holds up. AI helps me research, build and check more of that work. It does not take ownership of the decision.
Start with the customer and business problem. Compare the useful options, including improving what already exists. Choose a proportionate next move. Build the thing its user needs. Review the evidence and the actual experience. Keep what works and change what does not.
One decision, followed through
One decision, followed through.
At Modus Create, one app’s name was a string of technologies and diagram formats. I renamed it Just Add+ Embed and built the positioning and supporting work around the customer problem. The reported portfolio results from that period are shown with their limits: no absolute baselines and no controlled single-asset attribution. The name was one decision inside a broader portfolio effort.
This is a retrospective account of actual career work, not a live company workflow or a measured business result. Read the Just Add+ case study.
The working system
Keeping the evidence connected to the work.
My working system keeps source context, decisions, execution and review connected. It helps reduce repeated setup and keeps the evidence available when the work changes. The examples here show the workflow; they do not claim a measured time saving or autonomous operation across a company.
A first useful move
Where I start.
I start by understanding the customer, the business goal, the team’s current work and the constraint. Then I bring back a specific recommendation and a bounded piece of work to test, with a clear owner and a way to judge it. This is an approach, not a guaranteed first-month result.
Next step