AI enablement
Treat AI like a brilliant freelancer with amnesia
It knows every trade and forgets you every time, so every job starts from day one.
hancrafted / Human-led. AI-accelerated.
- Consulting
- Coaching
- Engineering leadership
AI gave you a promotion
you never asked for.
You get to 10x your productivity — if you learn to plan, spec and review, …
Move toward Explore why to start clearing the workload; hover or focus it to clear it fully. Move away to let it grow again.
Available Episodes
Browse the interactive slide decks and visualised essays.
Featured
Sales
Why it's never too early to start with sales
Revenue arrives months after the costs that create it, so by the time you feel you need sales it is already too late: start the pipeline before you need the money.
Knowledge base
Maintaining markdown for AI
AI collapsed the cost of writing documentation and raised the cost of verifying it. The work moved — and the verifying half is nobody else's job.
How I work
Scroll to progress →
STEP 01 // People first
People first.
AI is an accelerator, not a solution. Plug it into a culture that isn't ready, and you'll get the good and the bad parts faster.
Worth reading: Conway's law (opens in a new tab). If your systems already mirror your org chart, AI will copy the cracks even faster.
STEP 02 // Reason backwards
Reason backwards.
Plan backwards from an outcome you can measure, and every step has to justify itself. Blockers surface early, not in week six. It works for projects and for AI: give it a destination, not a to-do list.
Heard of graph engineering? Does it feel like project management for AI to you too?
STEP 03 // Outcomes over methods
Outcomes over methods.
On a highway we agree on the destination, and nobody drives the wrong way. Which car you take is up to you.
STEP 04 // Obsolete by design
Obsolete by design.
I measure my work by how little you need me afterwards. Ownership stays with the people, not with me.
My favourite quote, Lao Tzu: “When the best leader's work is done, the people say, ‘We did it ourselves.’”
STEP 05 // Fast feedback
Fast feedback.
Short feedback loops let people and AI correct course before the iceberg. Transparency means knowing what's wrong while it's still cheap to fix.
P
People first
R
Reason backwards
O
Outcomes over methods
O
Obsolete by design
F
Fast feedback
000%
How I work
STEP 01 // People first
People first.
AI is an accelerator, not a solution. Plug it into a culture that isn't ready, and you'll get the good and the bad parts faster.
Worth reading: Conway's law (opens in a new tab). If your systems already mirror your org chart, AI will copy the cracks even faster.
STEP 02 // Reason backwards
Reason backwards.
Plan backwards from an outcome you can measure, and every step has to justify itself. Blockers surface early, not in week six. It works for projects and for AI: give it a destination, not a to-do list.
Heard of graph engineering? Does it feel like project management for AI to you too?
STEP 03 // Outcomes over methods
Outcomes over methods.
On a highway we agree on the destination, and nobody drives the wrong way. Which car you take is up to you.
STEP 04 // Obsolete by design
Obsolete by design.
I measure my work by how little you need me afterwards. Ownership stays with the people, not with me.
My favourite quote, Lao Tzu: “When the best leader's work is done, the people say, ‘We did it ourselves.’”
STEP 05 // Fast feedback
Fast feedback.
Short feedback loops let people and AI correct course before the iceberg. Transparency means knowing what's wrong while it's still cheap to fix.
About / Proof in practice
People × process × technologyChange you can point to.
Less dependency. Better decisions. Software people actually use. Three changes I helped teams deliver.
Read each case: before → intervention → result
Illustrations are schematics, not charts.RIB Software~5months to self-running ScrumExplore case
A team that doesn't wait for its manager.
Engineering ManagerBefore
One gatekeeper
Every decision queued behind one person.
My contribution
Shared decisions
Shared planning, capacity and peer review.
Result
Independent team
The team ran its own Scrum in ~5 months.
Scope & timingAbout five months of team development. “Runs itself” refers to day-to-day team routines, not an organisation without leadership.
See the case evidence
Refinement, estimates, assignments and reviews converged on one gatekeeper. The team had too little ownership to move independently.
I coached shared refinement and estimation, made capacity explicit and built peer review. Team members took responsibility for planning and their own Scrum rituals.
In ~5 months, the team was running its own Scrum rituals with little need for me to be in the room. Distributed ownership became daily practice.
My intervention was not another approval layer. I helped the team refine and estimate together, communicate its capacity and unblock work with other teams. Peer review gave colleagues a way to build agreement and carry decisions themselves.
Basis: personal accomplishment notes; Engineering Manager role and five-month transformation recorded in my CV.
Audi30minutes saved per meeting, on averageExplore case
Less meeting. More deciding.
Tech Lead at XL2Before
Legacy approval
Brief: rebuild the Excel approval flow.
My contribution
Model the process
Stakeholder sessions became a state model.
Result
Digital sign-off
A core-process demo, meetings 30 min shorter.
Scope & timing6–8 weeks refers to the working core-process demo. The initial release took 6 months. The 30-minute figure is the average meeting reduction, not a demo benchmark.
See the case evidence
The final approval before serial production depended on a legacy process. The starting request was to reproduce the existing Excel workflow.
With an architect colleague, I worked through stakeholder sessions, state diagrams and a relational data model. We built one core process end to end.
The core-process demo made the model tangible. The digitisation effort also shortened approval meetings by an average of 30 minutes.
State transitions gave stakeholders a model in their own domain language. The demo ran on pre-production with data import and export, side by side with the existing tool. It was a concrete basis for agreement on how the process should work.
Basis: personal accomplishment notes for the demo; CV for the six-month initial release and average meeting reduction.
Selfbits5factories across EuropeExplore case
Beyond the demo. Into production.
Founding EngineerBefore
Production data
Orders needed to become usable OEE data.
My contribution
Build and pilot
Web, mobile and TV software, built and run.
Result
Software in use
200+ users, 3 HEPA lines piloted in 9 months.
Scope & timing5 factories and 200+ users describe the total deployment in my CV. The nine-month milestone covers the 3-line, 2-factory pilot, not the full rollout.
See the case evidence
How can production orders become usable data about equipment effectiveness? The work needed to connect manufacturing processes with the people operating them.
I designed, implemented and operated manufacturing execution software with the team: web, mobile and TV interfaces for scheduling, master data and OEE.
The software served 200+ users across the total deployment. A defined pilot acquired order-based OEE data on 3 HEPA lines in 2 factories within 9 months.
Manufacturing execution software connects production planning with work on the factory floor. OEE means overall equipment effectiveness. Here, the pilot concerned acquiring that data per production order; no percentage improvement in OEE is claimed.
Basis: CV for deployment and pilot scope; professional profile for work on scheduling, OEE dashboards and master data.