One of the questions I hear most often from executives is:
“How do we measure AI ROI?”
My answer is that it depends less on the AI than on the business process.
Some AI investments simply are not worth instrumenting. Others lend themselves to rigorous measurement. The difference is not the technology. It is the maturity of the underlying workflow.
Consider three common use cases: software development, IT service management, and proposal writing.
Software development
For software development, the answer is not to measure lines of code or even the amount of code generated with AI.
The real measure is how efficiently a team produces production-ready software that passes testing, code review, security review, and acceptance. The effort required to validate and correct AI-generated code is part of the workflow and must be included in the equation.
AI costs matter as well, so consumption needs to be attributed at the same level at which productivity is measured.
IT service management
In IT service management, it is not about tickets touched.
It is about tickets resolved correctly, quickly, and with minimal escalation.
Those efficiency gains may reduce support costs, but an even greater benefit often comes from reducing IT friction across the enterprise. Every minute AI gives back to employees is a minute they can spend doing something more valuable than waiting on a help desk.
If you are measuring only IT costs, you are probably understating the value AI creates.
Proposal writing
For proposal writing, it is not about pages drafted or the percentage of content generated by AI.
It is about producing compliant, compelling proposals that:
- Require minimal review and rework.
- Increase organizational throughput.
- Improve win rates.
- Contribute to backlog growth.
Measure what you understand
These examples have something important in common.
They are mature business processes with defined workflows, established quality gates, and measurable outcomes. AI is not evaluated in isolation; it is evaluated by how it improves the performance of the entire workflow.
And because these are mature processes, well-run organizations already understand their baseline performance without AI.
This leads to a broader point.
Many people tell me that AI ROI is difficult to measure. In many cases, I do not think this is really an AI problem at all.
AI does not create measurement problems. It exposes them.
If you cannot describe the workflow, identify where value is created, or establish today’s baseline, how can you credibly measure AI’s impact tomorrow?
The practical corollary is straightforward:
If demonstrating AI ROI is important to your organization, measurability should become one of your use-case selection criteria.
Do not start with the most exciting AI application. Start with the business processes you already understand well enough to recognize—and measure—real improvement.