The biggest mistake in GenAI projects?
Agility and traditional data warehousing have long been incompatible. But modern tools and methods are changing the game.
Torsten Glunde
• Feb 25, 2026
We expect from a bot what we would never expect from a specialist.
One that can do everything.
**Strategy. Code. Texts. Data. Law. **
Preferably quickly. And ideally at a low cost.
The reality in a team is different:
- Specialists beat generalists. Always.
And it’s the same with AI.
A super bot that is supposed to do everything is usually:
- slow
- expensive (hello token costs 👋)
- difficult to maintain or debug
The most exciting development for 2025/2026?
Specialization.
- Many small agents.
- Each really good at one thing.
- Like a strong team of specialists.
The real challenge is not the agent.
It’s collaboration.
Because 50 specialists without orchestration are not a team, but chaos.
How do you deal with specialization vs. all-rounders?