Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteTo measure who benefits from an AI investment, track both its total results and how those results—and its costs—are distributed. Set a baseline before deployment, define who the investment is meant to help, then assess productivity, quality, cost and service alongside worker experience, job quality, safety and transition effects. Keep monitoring after launch: an increase in output alone does not show who gained, who bore the costs or whether AI caused the change.
Why total return does not show who benefits
An AI investment can raise productivity or income in aggregate while producing very different outcomes for firms, workers, customers and communities. The OECD identifies skills, infrastructure, industry mix and integration into trade as factors that can shape how well countries, sectors and firms adopt AI and benefit from it. A single productivity or return figure therefore cannot establish that gains were broadly or fairly shared. OECD: Understanding the macroeconomic effects of artificial intelligence.
Outcomes also differ among workers. AI may automate some tasks and augment others, and the effects can vary by skills, experience, occupation, industry and disability. Measure job quality, safety and transition costs as well as output; otherwise, an apparent organizational gain may conceal losses or new burdens for particular groups. OECD report: The impact of Artificial Intelligence on productivity, distribution and growth.
A practical sequence for measuring an AI investment
1. Define the decision and the people affected
Specify which investment or deployment you are assessing, the decision the assessment should inform and the time horizon. Name intended beneficiaries and anyone who might bear costs: investors or owners, employees, customers, suppliers, the public or affected communities. State what success would mean for each rather than assuming that one group’s gains represent everyone’s.
#1 Best Overall
2. Set a baseline and intended outcomes
Before deployment, record relevant performance, costs, service levels, how tasks are allocated and worker outcomes. Choose a comparison period or group if feasible, and specify the outcomes the investment is intended to change. A baseline makes later comparisons more informative, but it does not by itself prove that AI caused any difference. The frameworks cited here support assessments tailored to a system and its purpose; they do not prescribe one experimental design for every investment.
3. Measure results at multiple levels
Pair organizational measures—such as productivity, quality, cost and service—with measures of worker and user experience, job quality, safety and access. Examine how gains and costs are distributed among affected groups. Disaggregate results by relevant worker or user characteristics when lawful and appropriate, and explain which groups are included in each measure.
Rank #2
4. Track automation and augmentation separately
Record which tasks change and how. For workers whose tasks are augmented, measure changes in capability, time and work experience; for tasks that are automated, consider reduced demand for particular work and the costs of transition. A total output measure can miss these different pathways and the people affected by them.
5. Monitor what happens after implementation
Check whether the intended benefits materialize once the system is in use. Track adoption, maintenance and risks alongside outcomes, and identify which groups actually experience the results. OECD guidance for governments recommends planning, implementing and monitoring AI investments to assess value for money, investment risks, timely deployment and realization of intended benefits. OECD: Governing with Artificial Intelligence.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →6. Report what the evidence can—and cannot—show
Separate observed changes from estimated causal effects, and measured performance from people’s reported perceptions. Describe the comparison used, its limits and any uncertainty. Do not treat a favorable survey response or a rise in productivity as proof that AI caused the result or that benefits were fairly shared.
Tailor the assessment to the system and its setting
There is no single measurement set that fits every AI investment. The OECD’s framework for classifying AI systems organizes relevant context around People & Planet, Economic Context, Data & Input, AI Model, and Task & Output. Those dimensions help explain why an assessment should reflect the system’s purpose, the people affected, its data and task, and the economic setting. OECD Framework for the Classification of AI systems.
NIST’s TEVV-Athlon framework describes a four-stage method for creating customized evaluations of AI performance and impacts. It is a framework for developing an assessment, not a universal score for deciding who benefits. NIST: TEVV-Athlon Framework for Evaluating AI Systems.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to compare two AI investments
Use the same time horizon and compare the same categories for each deployment. Report the underlying results by category rather than combining them into an unexplained weighted score.
| Comparison area | What to examine |
|---|---|
| Aggregate outcomes | Productivity, income, costs and service outcomes. |
| Distribution | Which firms, workers, customers and members of the public receive gains or bear costs. |
| Work and transition | Job quality, safety, displacement and transition effects, including differences between automation and augmentation. |
| Context and capacity | The system’s task, data and setting, and the capacity of the organization or sector to adopt and integrate it. |
| Realized benefits | Whether intended outcomes occurred after deployment, alongside adoption, maintenance and risks. |
These comparison areas synthesize OECD and NIST guidance. The sources do not prescribe a universal weighting scheme or a single causal return-on-investment formula that resolves distributional questions across all AI investments.
What worker surveys can tell you
In its 2023 publication Using AI in the workplace, the OECD reported that four in five workers surveyed said AI improved their performance at work, and three in five said it increased their enjoyment of work. These are reported perceptions, not causal estimates of investment returns or guarantees about every workforce. The cited publication information does not state the underlying survey fieldwork year. OECD: Using AI in the workplace.
Use fair-sharing goals as a measurement test
The OECD’s AI Principles call for responsible AI use at work that enhances worker safety, job quality and public services, fosters entrepreneurship and productivity, and aims for benefits to be broadly and fairly shared. For an investor or organization evaluating a deployment, that makes distribution a central question: identify who receives the gains, who takes on the risks or transition costs, and whether those outcomes match the stated purpose. OECD AI Principles.
Quick Recap
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.
Free tools Windows power users keep installed
One-click scans. No signup required.
