EN
DE
Why organisations are pulling too hard at Agentic AI
We live in a time when many things are happening simultaneously.
New tools, new models, new promises.
And suddenly everything is “agentic” – systems that are supposed to think, decide and act for themselves.
The idea is fascinating, no question.
But when I look at current developments, it reminds me of classic transformations that have failed at the same point before.
Head over heels – the old pattern in a new guise
The Handelsblatt recently reported on how transformations often fail because companies want too much too quickly.
They pull the rope too tight – and are surprised when it breaks.
That is exactly what is happening again – only this time with Agentic AI.
Instead of first understanding how people, processes and responsibility interact,
automation, orchestration and integration are implemented immediately.
The rope is under tension:
- technology pulls,
- governance holds back,
- and somewhere in between stands the human being.
When the pull becomes too strong, trust breaks – not the rope.
The balance between control and trust
The Dutch proverb ‘Niet te vast trekken, dan breekt het touwtje niet’ comes from seafaring.
It reminds us that leadership and control require sensitivity.
A rope keeps a ship safe, but only if it is allowed to withstand tension.
Applied to Agentic AI, this means:
If we control systems too tightly, we lose their creativity.
If we let them run too freely, we lose responsibility.
Transformation requires balance – and patience.
My conclusion
Agentic AI is not a sprint.
It is a new stage in the same journey that has already overwhelmed many organisations in their digital transformation.
Only those who learn to steer carefully now will benefit from this technology in the long term.
Or, to put it in old sailor’s language:
Keep the rope taut – but not too tight.
Otherwise, the ship won’t sail, it will break.
