Measuring AI ROI: how enterprise leaders evaluate impact

In 2026, we are seeing widespread AI experimentation but most leaders still struggle with proving business value.

At the recent Section The AI:ROI Conference, leaders from OpenAI, Salesforce, Cisco, and IBM discussed how organizations are moving beyond pilots to measure the real value of AI agents and copilots.

The consensus: AI ROI requires new measurement frameworks, new governance models, and a different mindset about productivity.

What emerged from the conference discussions is that measuring AI ROI requires a structured operating model.

1. Start with workflows and baseline metrics

Many AI initiatives begin with tools, but the more effective approach is to start with specific workflows.

Not every business process is a good candidate for AI automation. The workflows delivering the strongest results tend to share several characteristics:

• High volume
• Repeatable steps
• Measurable outcomes
• Limited reliance on complex human judgment
  
  

Processes that depend heavily on negotiation, contextual decision-making, or sensitive judgment typically require hybrid approaches where AI assists rather than replaces human input.

Organizations seeing measurable impact first identify processes that are repetitive, measurable, and operationally structured. Examples include reporting, internal research, documentation, or workflow coordination.

Once a workflow is identified, the next step is to establish baseline metrics:

• How long the task takes today
• How many manual steps are involved
• How frequently the task occurs
• What the operational cost per task is

Only after this baseline is established can organizations evaluate whether AI meaningfully improves performance. Without this step, most AI projects remain difficult to evaluate and rarely scale beyond pilot programs.

2. Measure productivity gains, not just cost savings

The more mature companies speaking at the conference are not just  focused on cost reduction and job replacement but on productivity amplification and using AI for competitive advantage.

The result is not necessarily fewer employees. Instead, teams gain more time for higher-value activities such as analysis, decision-making, and strategy. Organizations that evaluate AI purely through cost-cutting risk missing the larger opportunity.

“The only way to win is to find a way to pivot into understanding that your company can do so much more than what you're doing today.”
Edo Segal, CTO, Napster

3. ROI compounds as AI systems scale

The economics of AI change dramatically as adoption expands.

An isolated automation may produce modest gains. But when multiple workflows are automated across an organization, the productivity improvements begin to compound.

Over time, companies may operate with large ecosystems of AI assistants embedded in different processes, from operations and customer support to product development and internal reporting.

“Once AI starts executing hours of operational work every week, the conversation changes from efficiency to what teams can now build that they couldn't before.” DJ Sampath, SVP of AI Software and Platform, Cisco

‍

At that point, AI becomes less of a tool and more of an operational layer that augments how the organization functions.

‍

In other words, AI success depends less on the sophistication of the model and more on the discipline of how it is deployed.

‍