What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
AI can improve performance on some tasks and save workers time, but current evidence does not support a universal company-savings percentage or a count of jobs that AI will eliminate. The clearest results come from specific experiments and employer surveys; time saved, higher output, lower costs, and changes in staffing are different outcomes.
How to tell whether AI is improving productivity
Productivity means producing more or better work for a given amount of input. It is not the same as finishing a task faster: saved hours may be used for other work, absorbed by checking AI output, or simply not converted into additional output. Study results also depend on the task, the workers, and how the tool is incorporated into a workflow.
| Evidence | What was measured | Result and scope |
|---|---|---|
| Professional-writing experiment, Shakked Noy and Whitney Zhang, 2023 | Completion time and output quality | In a preregistered experiment with 453 college-educated professionals doing midlevel writing tasks, ChatGPT reduced average completion time by 40% and increased output quality by 18%. These results apply to the experiment’s tasks, not every job or company. |
| Customer-support study, Erik Brynjolfsson, Danielle Li, and Lindsey R. Raymond, 2023; published in the Quarterly Journal of Economics in 2025 | Customer issues resolved per hour | Across 5,179 agents, a generative-AI assistant was associated with 14% more issues resolved per hour on average. The increase was 34% for novice and lower-skilled workers, while the measured effect was minimal for experienced, highly skilled agents. |
| Field experiment, Eleanor W. Dillon, Sonia Jaffe, Nicole Immorlica, and Christopher T. Stanton, NBER, 2025, revised November 2025 | Email time and the quantity and composition of work | Across 66 firms and 7,137 knowledge workers in a six-month trial, tool users in the second half spent two fewer hours on email each week. Researchers did not detect a change in the quantity or composition of tasks from individual-level access. |
The results show why one productivity percentage should not be applied across an entire workforce. The first two studies measured outcomes on bounded tasks or in a particular service operation; the field experiment found less email time without a detected shift in the amount or mix of work. The OECD’s review of generative AI and productivity likewise finds that gains are especially pronounced for well-defined tasks with clear objectives. Effective use also depends on workers’ understanding, trust, skills, and an organization’s ability to integrate the tools.
Does AI actually save companies money?
There is no single, well-supported net-savings percentage for companies in the evidence summarized here. AI may help with writing, summarizing, editing, translation, coding, marketing content, sales, supply-chain management, or customer service. Whether those efficiencies reduce costs depends on what the organization does with the time saved and what it spends on AI tools, checking outputs, training, security, and workflow changes. A faster task is not automatically a lower total operating cost.
#1 Best Overall
The OECD’s 2025 representative survey of more than 5,000 small and medium-sized enterprises (SMEs) in Austria, Canada, Germany, Ireland, Japan, Korea, and the United Kingdom offers bounded, self-reported evidence. It was conducted in late 2024, and its findings should not be treated as audited savings or as results for all businesses.
- Among AI-using SMEs, 65% said generative AI improved employee performance.
- One third of SME users said it reduced employee workload.
- 14% of users said reliance on external contractors had fallen.
Those reports indicate possible operational benefits, but the survey does not establish a common percentage of net cost savings or show that every firm saved money. It also does not isolate savings after adoption and operating costs.
Rank #2
Will AI take people’s jobs?
Exposure estimates are about the potential for AI to affect work tasks, not a count of jobs already lost or a forecast of future layoffs. The International Labour Organization’s (ILO) 2025 global index estimates that one in four workers worldwide is in an occupation with some generative-AI exposure; 3.3% of global employment is in the highest exposure category. Clerical occupations remain especially exposed, and exposure varies by gender and income level.
The ILO’s conclusion is that job transformation is more likely than full redundancy because most occupations combine tasks, many of which still require human input. Its June 2026 synthesis of experiments, firm data, platform studies, and surveys finds that large-scale displacement remains limited in the evidence it reviewed. That describes observed evidence to date; it does not prove that future job losses will not occur.
The OECD’s late-2024 survey of SMEs in seven countries gives a separate, employer-reported view of staffing. Across the surveyed SMEs, 83% said AI had no effect on staff need, 9% said need decreased, and 6% said it increased. The reported responses do not sum to 100%, and they should not be treated as a complete distribution beyond the figures given. Staffing need is also distinct from the mix of skills a business wants: 20% said generative AI increased the need for highly skilled workers, while 9% said it decreased such needs.
The ILO’s June 2026 synthesis also notes that reported time savings have not yet translated into higher measured output, earnings, or employment in the evidence it reviewed. It flags concerns involving inequality, opportunities for younger workers, worker autonomy, coordination, and job quality. These findings make it important to distinguish task changes from changes in employment or pay.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What this means for workers and household finances
For an individual, an occupation-level exposure figure cannot tell you whether your particular job will disappear or whether your income will change. A more useful question is which parts of your role are routine and clearly specified, which depend on judgment or interaction, and whether your employer expects AI to change the way those tasks are done. The studies above also show that outcomes can differ by experience: a tool may help newer workers more than experienced specialists in a particular setting.
- Look for task changes, not just job titles. Ask which parts of your work are being automated, assisted, or newly checked, and how success will be measured.
- Check whether time savings change the work. A reduction in email or task time does not by itself establish higher output, better pay, or lower risk of staff reductions.
- Notice where human contribution remains important. The ILO’s exposure measure is not a redundancy forecast, and occupations contain a mix of tasks.
- Evaluate employer claims in context. Ask whether a claimed efficiency is measured output, self-reported time saved, reduced workload, contractor use, or an audited net cost reduction. Those measures are not interchangeable.
For personal financial planning, the available evidence supports neither assuming that AI will soon eliminate a specific job nor relying on an employer’s task-level productivity claim as proof that wages or employment will rise. Treat workplace changes as a reason to understand how your duties and skill requirements are evolving, rather than as a precise forecast of your future income.
Recommended Free Tools
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
