How do we measure the return on an AI workflow — and what baseline is honest?

With 95% of enterprise GenAI organizations seeing no measurable return, your board demands proof: establish baselines before deployment and model workflows as mathematical optimization problems to show direct P&L impact.

2026-07-21 · Counsel verdict · AIssential

The question

We need to report the return on an AI-enabled workflow to our board. What should we measure, what baseline is defensible, how do we avoid crediting AI for gains it did not cause, and how long should the measurement run before we make a scale-or-stop decision?

Counsel's position

Define 3-5 financial outcomes, model the workflow as an optimization problem, and measure for 6-9 months against a pre-AI baseline.

Verdict

The verdict: Define 3-5 financial outcomes, model the workflow as an optimization problem, and measure for 6-9 months against a pre-AI baseline.

Effective AI measurement targets three to five financial outcomes and establishes baselines before deployment

Given your need to report ROI to the board, this framework defines the exact metrics to track and the cadence for reviewing them.

• Just What You Need to Determine AI ROI: The Big Book of AI Metrics

AI measurement requires modeling workflows as mathematical optimization problems targeting specific outputs

To avoid crediting AI for gains it did not cause, you must shift from tracking pilot activities to modeling the causal mechanisms of business outputs.

• Don’t Just Talk About AI. Measure Business Outputs. Here’s How.

AI agents fail when their cost per successful outcome exceeds the displaced human cost

When establishing what to measure, the fully loaded cost per successful outcome is the ultimate survival metric for your scale-or-stop decision.

• Your AI Agent Passed Every Eval. Finance Still Killed It.

High-returning AI teams redesign workflows and establish measurement infrastructure before selecting models

Given your need for a defensible baseline, you must build native measurement into the workflow before writing any AI code.

• The AI Bubble Is Deflating — What Gets Cut First and Why Engineers Should Care

95% of enterprise GenAI organizations see no measurable return on their investments

To avoid being in the failing majority, your board report must demonstrate direct P&L impact rather than soft pilot metrics.

• The $30B GenAI Divide Is Coming for Soft Approval Memos

Read another verdict

• Buy a tool for this process, or build around our own knowledge?
• Our documents are a mess. Clean them up before AI, or after?
• Our best people's know-how isn't written down — can AI even use it?
• Automate this workflow, or redesign it before we automate?
• Which process should we point AI at first?
• Our AI pilot works but nobody uses it — fix the workflow or kill it?
• Rent AI from a vendor, or run your own?
• Let non-developers ship AI-generated code?

Get Counsel for your own decisions →