AI productivity is up, but profit is still lagging
Eighty percent of employees say AI makes them more productive, but only 37% of organisations see any EBIT impact, according to Fortune. Nearly two in three organisations are not yet banking the value they feel. The gap is not technical. It is about rules, incentives and who is willing to change them.

AI productivity is up, but profit is still lagging
AI productivity is becoming visible inside companies, but profit is not keeping pace. According to Fortune, 80% of employees say AI makes them more productive, while only 37% of organisations see any EBIT, earnings before interest and taxes, impact. That gap matters because it means many companies are buying speed without converting it into financial results.
Key takeaways
- According to Fortune, 80% of employees say AI makes them more productive, but only 37% of organisations see any EBIT impact, so most of the value is still not banked.
- Fortune says nearly two in three organisations are not capturing the value they feel from AI, which points to a value capture problem rather than a model problem.
- Fortune cites Maersk's USD 3.5 billion operating profit last year as a reminder that operating model changes, not just tools, shape whether AI translates into profit.
- Fortune reports that 69% of executives use AI less than an hour a week, which weakens the case for asking employees to transform faster than leadership does.
Source: Fortune
Why is AI productivity not turning into profit?
The short answer is that faster task execution does not automatically change the business system around it. According to Fortune, the problem is not a lack of data, but the gap between knowing what should happen and being willing to act on it. If AI speeds up analysis, drafting or scheduling, the benefit disappears unless budgets, roles and decision rights also change.
That is why the piece argues that the constraint is often human, not technical. Employees can feel more effective, but if the organisation keeps the same approval layers, the same role definitions and the same reward structure, the extra output stays trapped inside the existing process.
Source: Fortune
What do the airline and pricing examples show?
They show that even a proven model can stall when people do not trust the recommendation or when the organisation keeps old limits in place. Fortune describes an aviation partner that built a solver to weigh crew, aircraft rotation, passengers and ground control together. It took 18 months to go from accepting half of its recommendations to almost all of them.
The same article describes an advisory client that implemented a dynamic pricing model and then surrounded it with so many constraints that it was effectively fenced in by the past. In both cases, the model could recommend the move, but the organisation still had to approve it.
Source: Fortune
Why do incentives decide whether AI pays off?
Because people respond to what the organisation rewards, not just to what the model recommends. Fortune uses Diageo as an example: when Ireland's rugby team won, money worked harder on TV and Guinness; when the sun stayed in Italy, Tanqueray's season ran longer. The math was clear, but every euro moved also changed who hit target, who earned a bonus and who lost ground.
That is the business lesson for AI. If the scorecard only rewards local efficiency, teams will protect their own metrics. If the scorecard rewards shared outcomes, leaders have a better chance of turning AI-generated recommendations into profit that shows up in EBIT.
Source: Fortune
What does this mean for leaders now?
It means the next bottleneck is likely to be leadership behaviour, not model quality. Fortune quotes a line from the piece that the constraint is us. The article also says 69% of executives use AI less than an hour a week, which makes it harder for senior teams to ask for faster adoption from everyone else.
For CEOs, CFOs and COOs, the practical implication is simple. If leaders want profit from AI, they need to change the rules around decisions, targets and accountability before they ask the front line to move faster.
Source: Fortune
What to do now
For most companies, the right response is not to buy more AI tools. It is to redesign how value gets captured from the tools already in use.
- Start with the profit outcome you want, not with the current process. Define the operating result first, then work backward to the decisions AI should change.
- Audit the hidden rules, including the ones people call common sense. Keep only the constraints that still make sense for the future state.
- Tie part of performance management to shared business outcomes, not only to local team metrics, so AI recommendations do not die in departmental silos.
- Protect time saved from being turned into an immediate headcount cut, because people will hide productivity gains if they think the benefit only funds layoffs.
- Ask senior leaders how often they personally use AI. If executives are not changing their own workflow, the rest of the organisation will notice.
Frequently asked questions
Why do employees feel AI is helping when profit does not improve?
Employees feel the benefit because AI can reduce time on analysis, drafting, scheduling and routine work. According to Fortune, 80% of employees say AI makes them more productive, but only 37% of organisations see EBIT impact. The missing link is capture. Companies often improve individual output without changing the incentives, approvals or role design needed to turn that output into profit.
What is the biggest barrier to capturing AI value?
According to Fortune, the biggest barrier is not data or model quality. It is the human and organisational constraint around acting on the insight. The article describes teams that know what is optimal but still hesitate because of hierarchy, legacy rules, safety concerns or fear of overruling long-held judgement. That is why AI value often stalls after the pilot stage.
What should CFOs and COOs track beyond productivity?
CFOs and COOs should track whether AI is changing EBIT, decision speed, approval cycles and accountability for shared outcomes. Fortune's examples suggest that productivity alone is not enough. The useful question is whether time saved is being reinvested into growth, margin or service quality, or whether it disappears into the same operating model with no financial payoff.


