|
PRODUCT
Product
Before we build, I ask which decision we are trying to improve —
for a user, an operator, or a reviewer who has to live with the outcome.
Feature volume is a weak scoreboard.
When AI enters the backlog, honesty about failure modes is part of the MVP.
Escalation paths are product requirements, not phase-two niceties.
|
|
| Before we build |
- Name the decision. What call gets better if this ships? If we cannot say, we are decorating.
- Name the owner. Every meaningful call has a person attached before the first ticket.
- Thin slices that teach. Small cuts that produce learning beat thick roadmaps that stall.
- Honest AI surfaces. Confidence, failure modes, and how to escalate — visible in the UI.
- Fewer, clearer bets. A backlog without owners is a wish list.
|
|
| Launch criteria I actually use |
- Success = decisions improved, not tokens spent or demos booked.
- Who can stop this? How do we reverse it?
- Builders, operators, and reviewers know who decides when something drifts.
- Writing and research feed the roadmap — not the other way around as theatre.
|
|
|
STICKY NOTE
“A product that cannot explain its decision rights will explain them in an incident review.”
Better to write them down while the coffee is still hot.
|
|
|
|
|
| Special Offer |
|
FREE
Intro call / coffee chat.
No hard sell. Just talk shop
— agents, governance, product.
|
|
|
|
|
|
|
|
|
|
|
|
|
N
Optimized for
Netscape Navigator
|
|
|
Mac
Platinum UI
c. 1996
|
|