Key takeaways
- Just because something can be automated doesn't mean it should be.
- 90/10 rule: AI does the legwork, the human does the final review and decision.
- Output isn't outcome: without human review, quality assurance is missing.
Just because you can automate something doesn't mean you should. That's the most important guideline when AI arrives in a company. Technically almost everything could be fully automated today: emails, offers, posts. The question isn't whether it's possible, but where the human has to stay.
The 90/10 rule
AI is great at the legwork: research, drafting, delivering variants, gathering ideas. That's easily 90% of the effort, and exactly what it should take over.
But the last 10% belong to the human: the final review, the personal note, the nuance, the strategic decision. That's exactly where it's decided whether an AI output becomes a good result.
Output isn't outcome
A fully automated pipeline produces output. Whether that becomes a reliable result is checked by no one but a human. Cut that step and you save a few minutes while risking quality, context and reputation.
That's why we always build automation with a human in the loop: AI does the legwork, the human owns the result. That's not distrust of the technology, it's quality assurance.
