Situation · a knowledge process is stuck

People keep joining. And it is not getting faster.

A process that runs on reading, judging, comparing and preparing does not scale with more people. It grows in handovers, in checks, and in dependence on the few colleagues who really understand it. Whoever goes on holiday takes the lead time with them.

Lead times are climbing and we cannot find anyone who can do this work.

Two people know how it really works. Everyone else asks them.

Every check we add makes it slower, not better.

01 · Divide, not speed up

Where does it stall?

Speeding up one step moves the queue. What helps is deciding again what a human must do because it takes judgement, and what a system can do because it is rule work. Three questions that need answering first.

Question 01

Where does the judgement actually sit?

A magnifying glass with a red rim on a sheet of paper.

And where do we call something judgement when it is really a rule. That distinction decides what stays with people and what does not.

Question 02

What happens when it goes wrong?

A red alarm bell on a bracket, its clapper in motion.

Who notices, how fast, and what the damage is. That determines how much checking the new process needs.

Question 03

Is this allowed on our own files?

A filing cabinet with three drawers and a red padlock on the top one.

Permissions, privacy, and where it runs. On a clean test set you prove nothing, so this has to be clear up front.

When this does and does not fit.

Works well when

The work consists of reading, judging, comparing and preparing, and does not scale with more people.
Two or three colleagues carry the knowledge; everyone else asks them.
You want to prove on one piece that a different division works, before rebuilding the whole process.

This is not

It is not an investigation into what is going on in the first place; if that is still unclear, the Scan is the first step.
It is not about retyping between screens that already exist, because then The Connector is faster and cheaper.
It is not a replacement of judgement: what takes judgement stays with people, and establishing that is the work.

02 · Redivide, on one piece

How do you approach it?

Not automate, but redivide. Before you rebuild a whole process, you want to prove on one bounded piece that the new division works, and that is the size of a ten-day Sprint.

A drafting compass with a red point on a folded drawing.

Outcome 01

The new division

Which actions a system carries out, which it prepares, and where a human signs off. With the exceptions included, because that is where the real work sits.

A small gearbox on a test block, with a lit red lamp.

Outcome 02

Proof on the hardest part

Working code on the part it usually breaks on, on your own files. Not on a clean test set.

A Y-shaped pipe with one branch shut off by a red valve.

Outcome 03

A decision with a path

Build or no-build, and if it is build: what it costs in time, people and money, and what it returns in lead time or capacity.

Is a process like that stuck in your organisation?

A half-hour call costs nothing. If it turns out the work is mostly retyping, or the question is still too broad, I will say so.

Book a call

03 · What is running

What can you check?

What is below runs, and you can check it yourself before we talk money.

A round gauge on a stand, with one red needle.

Cranely proves the other side

A scorecard tool where judgement deliberately stays with people. Not everything should be delegated, and establishing that is the work of a Sprint.

To check · cranely.app

A conveyor belt with three boxes, the red one just rolling off.

I run my own delivery pipeline

Agents pick up tasks from my backlog, open their own branch and deliver the work as a pull request. I specify, review and merge. An agent cannot skip or talk its way past the quality gates: test suite, code style, front-end build and a separate review round.

How the pipeline works

To check · This site was built that way. An agent never merges by itself.

I know what production demands

Building software since 1999, fifteen of them independent, through Progress Software in enterprise environments for organisations including WHO and Van Lanschot. That is what you need at the moment a demo has to become a business-critical system.

About Daniel

To check · The full career is on the About page.

04 · Three products, from small to large

Which step fits?

Which step it becomes depends on how clear it already is which piece of the process comes first. Prices and scope are on the page of the product itself.

A pair of binoculars on a wooden tripod, with a red band around the barrel.

The Scan

Not yet clear which piece?

Then the Scan is the smaller step: two working days to establish where the judgement sits and which part has to be proven.

See the Scan

A small device under a glass bell jar, with a red tag on a thread.

The Sprint

Is the piece clear?

Then the Sprint proves in ten working days whether the new division works: on your own files, with working code on the hardest part and a decision.

See The Sprint

A connector of two pipes with a red valve on top.

The Connector

Is the work stuck inside one system?

If it is mostly looking up and retyping between screens that already exist, The Connector is faster and cheaper.

See The Connector

05 · What people want to know

Any questions?

The questions that always come, with short answers. If yours is not among them, a half-hour call is the fastest route.

Do you work on our real files?

Preferably yes, because on a clean test set you prove nothing. If that is not possible we work with a representative copy and I state explicitly what stays unproven as a result.

What if the judgement has to stay with people after all?

Then that is the outcome, and a valuable one: you know which part is rule work and which part is not, and you build only the first.

Do our people lose their work?

They lose the looking up, comparing and retyping. The judgement stays with them, and that is where the time goes.

Which models and vendors do you use?

Whatever fits the question, including locally running models when files cannot leave the country. I build so you are not locked into one vendor.

Who figures it out

Starts by listening to the people who do the work.

I have seen too many stuck processes get sped up instead of redivided: one more tool, one more check, and the queue only moves along. What helps starts with the people who do the work. They know where the judgement sits and where it is really a rule, and only then can a system take over the rule work.

Daniel Plomp · AI Architect & Builder

Which of your processes is stuck?

Describe in two sentences where the time goes. You will hear whether a Sprint is the right tool, and if it is not, I will say that too.