Building got cheap. Judgment did not.
I build and lead teams that ship production AI, and install the operating model that tells you whether it landed.
Vancouver, remote-friendlyFull-time or fractional
did the onboarding revamp land?
Activation up 14 points six weeks after release.
MeasuredTrial starts flat. Sample is still too thin to call.
BenchmarkedShipped nine weeks ago. Nobody has used it twice.
MeasuredA live read: one team, three launches, scored the same way. Toggle it to twelve teams and the pattern is the easiest thing on the page to see.
- Strategy & Bets
- Design
- Development
- Delivery
- Go-to-Market
- Data & Outcomes
- Operations
A builder, a point of view, and someone who has run the room.
The work moved to judgment. So did the way I measure it.
Agents removed the queue for competent output. What is scarce now is deciding what deserves to exist, and knowing whether what shipped actually changed anything.
The point of view, in fullOutput stopped separating good from busy.
Shipping totals climb for everyone now. They no longer tell better from busier.
Landed is a definition, agreed up front.
A behaviour, in named accounts, by a named date, set before the work starts.
Agents produce. One person owns every call.
Reading, drafting and checking run on a schedule. The calls keep a named owner.
- Writing the code
- Reading the tickets
- Drafting the options
- Checking the rules
Struck through because these stopped being where the cost sits, not because they stopped mattering.
Deciding what is worth building
When you can build almost anything this week, choosing becomes the expensive act.
Knowing whether it worked
Shipping is visible by default. Landing is only visible if somebody sets out to see it.
Hire the seat, or borrow the judgment.
Two ways in. The fractional side runs as eight fixed-purpose missions, and the diagnostic leads. Everything is scoped in a conversation.
A company filling a product, product-operations, or product-and-technology seat, from director to chief.
The seat filled by someone who has gone employee number one through post-Series B.
An exec team that needs senior product and technology judgment in the room every week.
A standing seat on the AI agenda, and it is reversible. One client at a time.
Scored with the same three words as everybody else.
Twenty years sit behind these entries. Each one is sourced outside this site, or anonymous and claiming no outcome.
Employee #1 and CPTO: the raise, the founding team, and a launch in about six months with every SOC 2 Type II and ISO 27001 control in place and the observation window open.
Dacard.ai applied research. The prototype runs against one company, a sample of one, so it claims watch.
AI modernisation for an established consultancy, one estimating desk first. Nothing delivered yet, so it claims nothing.
Bring the decision you are stuck on.
Thirty minutes, no deck. If I am not the right person for it, I will say so and point you somewhere better.