A story about promises, outdated work mentalities, and the age where companies stop building loyalty—and start building flexibility
She sits in a glass office building at one of the Netherlands’ most well-known banks, working on something that was sold to her as forward-looking.
The pitch was clear: an AI system would classify documents. She and a team of other workers would handle the edge cases—the documents the system couldn’t process. The exceptions. The nuance.
Human intelligence supporting artificial intelligence.
There was a deadline. There was urgency. There was even a sense that this was important work.
But within weeks, the reality looked very different.
The AI wasn’t classifying as much as expected. Entire batches fell through. Systems didn’t align. Workflows were adjusted on the fly. And instead of focusing on complex classification, her days became filled with something else: backtracking. Manual corrections. Working in older systems to fix what newer systems couldn’t yet handle.
The work became more repetitive. More procedural. Less aligned with what was originally described.
And the question started to form:
Is this really what I was hired for?
When promises don’t match the system
Her frustration is not abstract. It’s specific.
There was a promise—explicit or implied—about the nature of the work. About the level of thinking required. About the direction this could go.
Now that structure is unstable.
What she’s experiencing is not just a disappointing job. It’s a moving target.
And that’s the part that’s easy to underestimate: in AI-driven environments, the work itself is still being figured out.
- What the system can do changes
- What humans are needed for changes
- How productivity is measured changes
- What “valuable work” even means changes
So the role you enter is not the role that stays.
UK government guidance on AI automation emphasizes that systems should be “transparent, explainable, and aligned with human values,” but that guidance doesn’t fix the reality that most AI implementations are still in iterative development—and workers are the ones who absorb the instability.¹
A European Commission report on AI and employment notes that AI will boost overall employment in Europe, but the gains will be uneven—mainly benefiting high-skilled, prime-age workers, while low-skilled workers face higher risks without targeted support.²
Subscribe to continue reading
Subscribe to get access to the rest of this post and other subscriber-only content.
