You may have been mis-sold

AI is going to reduce our work, isn't it?

Well, No. Not if you take the research published in Harvard Business Review this week.

They found that AI doesn't reduce the work, AI tools are expanding workloads, adding time managing, reviewing and correcting the work of the AI, sometimes more than the benefit of outsourcing in the first place.

This feels very familiar for those who grew up in the world of shared services and outsourcing.

Clarity and the ability to codify the work, so that you minimise reviews, corrections and oversight is essential if you want to benefit.

Sloppy requests = Sloppy work.

Unfortunately, like any outsourcing project, the headcount is removed based on the plan, not the reality of the work left and those who remain are overloaded trying to sort out the mess.

Their research showed that AI adoption correlates with longer task lists, tighter deadlines, and heightened performance expectations.

The work doesn't reduce, there is just more of it and the overhead of managing and overseeing AI has not been taken into account.

The study also highlighted how different populations see the success of AI. At a senior level they see more outputs, more graphs, more reports, more insights, and it seems like a resounding success.

Below the waterline you have teams trying to keep up, adapt, correct and fix the AI generated work and increased expectations. They are drowning.

Not only that, this isn't the work they signed up for. They aren't supervisors, or managers, they are workers. This gives them little skill or satisfaction.

Link to a great article: https://www.webpronews.com/the-great-ai-paradox-why-artificial-intelligence-is-making-workers-busier-not-freer
Link to HBR Research (Paywall): https://hbr.org/2026/02/ai-doesnt-reduce-work-it-intensifies-it

BESCI AI OPINION

As someone who grew up in the heady days of outsourcing, this feels strangely familiar.

There is solid research done by Lisanne Bainbridge in the 1980s where she called this the "ironies of automation."

Automation is great at routine, well-specified operations, but struggles where judgment, context-sensitivity, and adaptive reasoning are required.

If you can't codify it, or the path isn't clear you become dependent on humans to solve the inconsistencies.

Ganna Pogrebna, a UK based researcher, talks about Behavioural Exhaust, that AI is responding to what it has been trained on; the outputs, not how the humans got there. The relationships, the trade-offs, the trust that was built.

It is also a reminder of garbage in=garbage out.

Many, who move into a supervisory role, feel grief at leaving behind the job they felt competent and skilled at to oversee others.

With the move for humans to oversee AI doing the work they would have done, there is going to be a collective grief until competency is reestablished.

Will it last forever. No. AI will get better teaching, prompting and will be more able to solve the inconsistencies it faces. Humans will work out how to supervise effectively. The pattern repeats.