Image the scene. You’re at your desk, deadline looming, and also you resolve to let AI deal with the primary draft. You kind a immediate. The result’s flawed. You strive once more. Nonetheless not proper. Ten minutes later you’re no nearer, and the clock is ticking. Finally you shut Copilot and do it the previous method, the way in which that takes longer, however not less than you realize it really works.
If that sounds acquainted, you’re not alone. And in line with Forrester VP and Principal Analyst JP Gownder, it’s changing into one of many fundamental productiveness issues within the fashionable office.
“I need you to place your self into the place, I do know I’ve personally, of: I strive a immediate and it fails. I strive one other immediate and it fails,” he informed UC At present. “At that second, I’ve a call to make. Both I can hold messing round with Copilot with no precise assure that I’m going to get it to do what I need, or I may give up and do it the previous method. What we’re seeing is plenty of abandonment behaviour, as a result of persons are both losing time and by no means getting a solution, or just abandoning the instrument. And after they abandon the instrument, they fall off the training curve completely.”
Watch the complete interview: Why AI Literacy Is Hurting Productiveness: Forrester’s JP Gownder
The numbers behind the issue
Gownder’s feedback come alongside Forrester’s second AIQ report. AIQ stands for Synthetic Intelligence Quotient, a measure of worker readiness to succeed with AI instruments at work. The findings make uncomfortable studying for any organisation that has invested closely in enterprise AI.
Regardless of greater than 80% of firms having deployed not less than some AI instruments, simply 16% of staff throughout the US, UK, Germany, France and Australia achieved a excessive AIQ rating in 2025, up from 12% in 2024. Gownder is obvious that the tempo of progress is nowhere close to matching the tempo of deployment.
Solely 51% of organisations prepare non-technical employees on generative AI in any respect. Simply 23% train immediate engineering. And solely 37% of staff really feel assured adapting to AI-driven methods of working, a determine that has barely shifted 12 months on 12 months. As UC At present has beforehand reported, almost half of all AI licences go unused, costing giant enterprises a mean of $80.6 million yearly — and the AIQ knowledge helps clarify why.
“For many staff, the fee to that particular person of utilizing a instrument like Copilot or Gemini is usually larger than the time financial savings they obtain on the opposite finish,” Gownder explains. “As a result of they’re studying by doing, and that studying is sluggish, painful, and taking place with out almost sufficient help.”
A brand new downside: AI slop
Past the abandonment cycle, Gownder identifies a second productiveness drain rising in workplaces. He calls it AI slop.
“Work slop, AI slop that folks ship round at work is changing into an enormous downside,” he says. “Folks don’t need to learn it, in order that they don’t learn it. It’s all these folks producing all this content material that’s filling folks’s inboxes after which they don’t learn it. That’s damaging productiveness proper there.”
The image is one among expertise creating new inefficiencies as quick because it guarantees to take away previous ones, not as a result of the instruments are dangerous, however as a result of the folks utilizing them haven’t been given what they should use them effectively.
The duty hole
That is the place organisations are essentially getting it flawed. There’s a widespread assumption in enterprise AI rollout that the instruments will largely communicate for themselves, that staff will discover, experiment, and naturally enhance. Forrester’s analysis suggests in any other case, and the results are falling on the workforce.
“Staff are usually not accountable for buying these abilities on their very own,” Gownder says. “You because the employer are accountable for cultivating a studying and engagement atmosphere that may equip them with the abilities, understanding and ethics they should succeed. That is your duty as a pacesetter. It isn’t one thing you simply push all the way down to the staff and say, good luck.”
The answer, he argues, just isn’t extra on-line coaching modules. Organisations have to rethink how they help AI adoption, constructing steady, hands-on, peer-based studying that places the worker somewhat than the expertise on the centre. Forrester’s analysis discovered that social studying is not less than twice as efficient as formal coaching in the case of elevating AIQ in follow.
“This looks as if a really techno-focused train,” he says. “It’s a human-focused train. We have to make investments extra in folks as we roll out AI, not much less.”
For the organisations nonetheless ready to see a return on their AI funding, that could be a very powerful line in the entire report.
