Where AI pays.
Value concentrates. Benchmark against the function and the workflow—not a market headline that has nothing to do with the work you are changing.
Most potential value sits close to high-volume work.
Customer operations, marketing and sales, software engineering, and R&D hold roughly three quarters of generative AI’s estimated annual value potential.
The useful benchmark is the function, then the workflow inside it.
Customer operations
Resolution, routing, summaries and knowledge retrieval.
Marketing and sales
Research, proposals, campaigns and next-best actions.
Software engineering
Code, tests, review, documentation and incidents.
Research and development
Literature, simulation, analysis and design generation.
A fast-growing market does not make your use case valuable.
Software and information services, banking, and retail led 2024 AI spend. That says where capital is moving. It does not prove a specific workflow will return more than it costs.
The jobs number describes redesign—not AI alone.
The World Economic Forum projects 170 million jobs created and 92 million displaced by 2030. Technology is one of several forces in that forecast.
Measure what changed in your workflow before claiming what changed in your workforce.
A net 78 million jobs could emerge across technology, demographics and the green transition.
AI alone will create or remove those jobs inside your company.
A business result has to exist outside the AI itself.
Usage, demos and enthusiasm can all rise without producing a result the company can defend.
Name the result before claiming the value.
Shows that people interacted with the technology.
Shows that something meaningful actually changed.
Sources and method
- McKinsey, The economic potential of generative AI
- IDC, worldwide AI spending industry outlook
- World Economic Forum, Future of Jobs Report 2025