80% of Workers Say AI Boosts Productivity. Only 37% of Firms Bank It
80% of employees say AI makes them more productive, but only 37% of firms see EBIT impact. A Fortune commentary pins the gap on bonuses, hierarchies and fear.
By Nathan Brooks
5 min read

What's News
- 80% of employees say AI makes them more productive; only 37% of organisations see any EBIT impact, per McKinsey data.
- An airline AI solver took 18 months to go from half its recommendations accepted to almost all.
- Maersk earned $3.5 billion in operating profit last year after its integrated logistics pivot.
- 69% of executives use AI less than one hour per week, per Stanford research cited in the piece.
Eighty percent of employees say AI makes them more productive, yet only 37% of organisations see any EBIT impact from it, according to McKinsey data cited in a Fortune commentary by Louisa Loran, a former Google strategic business transformation lead. For nearly two in three companies, none of the felt value has been banked. The gap, Loran argues, sits in people, not in systems.
The core problem isn't missing data, she writes. Knowing what needs to happen and being willing to act on it are two very different things. An Irish rugby win, a long Italian summer and an aircraft waiting at the gate all illustrate the same tension: the data knew what was optimal long before the organisation was willing to move.
What's actually blocking the value?
Many of the constraints holding AI value back aren't rules at all. They are past decisions that have hardened into rules — and people have the power to change them.
Loran draws on a large aviation customer she worked with while leading a global industry team and later Google's strategic business transformation unit. The airline had a tenured EVP as sponsor, millions in multi-year capital and the best teams on both sides. Together they built a solver weighing crew, aircraft rotation, passengers and ground control at once. Only legislation, capacity and safety went in as hard constraints; everything else went in as input. Every time someone said "I just know," the team went back to what had driven that decision, and coded it in only if it held true for the future.
The solver's value was proven. It still took 18 months to move from accepting half its recommendations to accepting almost all. The controllers, carrying responsibility for passengers' lives, set the pace — every recommendation they accepted questioned the judgement they had built a career on. The airline became one of the most profitable in the world, but expanding the tool across the network stalled on decision hierarchies and fear of an AI model overruling decades of service.
You can automate the analysis and still preserve the corporate politics. Loran recounts an advisory client who described, with real pride, a newly implemented dynamic pricing model — and then listed all the constraints bolted onto it. In her words: "They had bought a model to find the future and fenced it in with the past."
Why incentive plans kill good recommendations
The model can recommend the move; only the bonus plan can approve it. More than a decade ago at Diageo, Loran's team combined sell-out, sentiment, media spend and weather data by brand and country. When Ireland's rugby team won, money worked harder on TV and Guinness than where the plan was first built. When the sun stayed in Italy, Tanqueray's season ran longer. It was just math.
But every euro moved changed which country manager would hit their target, who would gain their bonus, and who would have to tell their family that this year's summer vacation was at stake. The tool remained a lagging indicator of what could have been.
That is what sits between the 80% and the 37%. People use AI to do tasks faster, then get asked to do much more with the time it frees. When AI is the reason given for cutting headcount, showing how much it saves you can feel like drafting your own redundancy, Loran writes. Companies are shaving tasks off old roles, and no one is paid to imagine the new ones.
The incumbent's dilemma, inside every employee
Every employee is now an incumbent, facing the classic incumbent's choice: protect the past or own what comes next. At Maersk ten years ago, Loran lived this at corporate scale. Margins were thin and the value lay in customers' wider supply chains, but a company that owned ships had to prove it would choose the carrier best for the customer, not for its own fleet. In 2016, the industry wasn't profitable; integrated logistics paid off through the pandemic, and last year Maersk earned $3.5 billion in operating profit on $54 billion in revenue.
Often the future is already in the work people are proudest of. After a session Loran led, an executive assistant told her she used AI a lot — more than her boss knew — and worried she'd be replaced. Her proudest contribution wasn't a task: she had aligned a leadership team of strong personalities behind one common goal, catching their fears before they were voiced. If an AI model were ever trained to do that, Loran notes, she would be the one to train it. What stood between her and that future was not technology. It was knowing her own value.
Multiply her by a workforce, and you have the gap between the 80% and the 37%. The economy organises around its next constraint. Until now it was something you could raise, build or buy. This time, it is us.
Four moves that shift the human barrier
In her advisory work, Loran says four moves consistently help:
- Start from the ideal ambition, not from today. An outcome is too small; it invites people to optimise the present. An ambition lets them imagine where they stand in it.
- Audit the constraints, including the ones in our heads. It takes curiosity to ask what drove each rule and whether it still fits, and honesty to name the narratives we cling to.
- Steer the system, not just the budget. Tie part of every target to the shared result. Pay tells people how they're coping this year; ownership tells them whether they take part in the next ten.
- Live the change before you ask for it. If every hour saved becomes a headcount cut, people will hide it.
The leadership example matters most: Stanford research cited in the piece finds 69% of executives use AI less than an hour a week — a transformation leaders ask others to inhabit while negotiating how much of themselves they'll change.
Momentum doesn't start with the plan, Loran writes. It starts with the first honest move — and organisations move quickly once their leader does.
Source: fortune.com
More from Nathan Brooks
Show full bio
News editor covering regulation, policy and their impact on entrepreneurship, startups and venture funding at Business Bearings.
2,232 articles