Agile changed software teams. Lean-AI changes every team.
Since 2024, artificial intelligence has been producing software in departments that were never development departments. And in the departments that always were, delivery times have stayed exactly the same. Both have the same cause.
Artificial intelligence changes the work, not just the tools
Almost everyone is deploying AI. Almost nobody is redesigning the work around it. The money goes into licences and tools while the question of how people are meant to work with them stays open. That is measurably where the return fails.
88 per cent of organisations use artificial intelligence in at least one function. 6 per cent see a meaningful contribution to earnings from it.
of companies offer no AI training at all – while 35 per cent complain about missing AI skills.
have actually designed roles, processes and operating model around artificial intelligence.
The figures are self-reported by leaders.
Investing in technology and not in the people who work with it buys speed in a place where the problem never was.
This is a flow problem, not a tooling problem.
What Lean-AI is
Lean-AI is Lean, induced by artificial intelligence. Anyone who engages seriously with AI ends up engaging with lean practices: AI lowers the cost of doing and not the cost of deciding, moves the bottleneck into approval, review and alignment – and thereby exposes exactly the questions Lean has worked on for seventy years.
An organisation can only really use artificial intelligence once it rebuilds the way it works, not just its tools. Lean management supplies the toolbox for that rebuild, because for seventy years it has worked on exactly the problems AI is now exposing in every department.
Lean comes from manufacturing, not from software development, and rests on five principles:
- Value
- What counts is decided by the recipient, not by the producer.
- Value stream
- Make every step needed for delivery visible. Whatever adds no value is waste.
- Flow
- Keep the work moving instead of keeping people busy.
- Pull
- Produce nothing before the next step calls for it.
- Perfection
- Not an end state, but continuous improvement.
Utilisation versus flow. When producing costs almost nothing, a fully utilised organisation is no longer a sign of efficiency. It is the most reliable way to grow the queue.
HeptaMS arranges the major IT management frameworks along seven senses of an organisation: direction, leadership, flow, cadence, stability, collaboration, reach. The book describes that in full, the courses make it usable. To the book
Which of these describes your situation?
Agile for years – and now AI is changing the whole way you work.
For software teamsAI workflows and agents that have long been running on live systems.
For business unitsLimits, approvals, the way back – and who answers for it.
For leadersWhat people ask
What is Lean-AI?
Lean-AI is Lean, induced by artificial intelligence. The term says exactly that: engaging with AI means engaging with lean practices as well. AI lowers the cost of doing, not the cost of deciding; the bottleneck moves out of production and into approval, review and alignment – into the field Lean has worked on for seventy years.
Is Lean-AI another method next to Scrum, SAFe or ITIL?
No. Lean-AI describes the situation organisations are in. What to do about it is ordered by HeptaMS: along seven senses, across Scrum, SAFe, ITIL and the other frameworks.
Why does the organisation not deliver any earlier despite AI tools?
Because tools speed up individual activities, not the value stream. When a draft takes minutes instead of days, review, approval and alignment set the lead time. The volume of work requiring review goes up; review capacity does not go up with it. Without a change to the flow, the organisation merely produces half-finished work faster, in front of the same queue.
Why does every department become a development department?
Anyone who lets an agent loose on a real system is developing software, even if the org chart says “accounting”. That department inherits the problems software teams had to solve over twenty years: versions, rollback, orderly change, tests, limiting the blast radius.
Does this require IT knowledge?
Not to get started. The path for departments without an IT background begins with three free courses that need no prior knowledge and carry no exam requirement. The answers come from IT; understanding them does not presuppose it.