The AI ROI Playbook

Calculate the cost of AI by the ROI it creates.

A practical guide for operators, commercial leaders and CFOs. Price the process you already have, separate what should be automated from what stays human, and run the numbers on capacity released, additional throughput and total system cost — including the cases where AI implementation is genuinely not worth the money.

What's Inside

A costing model, an automation decision map, the ROI equation and a 10-point investment test.

Software Price Is Not the Business Case

• Why comparing implementation to a subscription measures the ingredients, not the capacity
• Tool + human vs an AI-enabled operating model
• Where work executes and where humans govern
• Value comes from the process the stack can reliably execute

Price the Process You Already Have

• Direct labour + review and rework + delay + constrained throughput
• A fill-in table for research, execution, follow-up and admin
• How to calculate loaded hourly cost properly
• Use contribution margin, not headline revenue

Automate, AI-Assist or Keep Human

• Automate: data movement, status updates, follow-up, rule-based routing
• AI-assist: classification, drafting, personalisation, research summaries
• Human-owned: prioritisation, exceptions, negotiation, accountability
• Strong candidates vs tasks you should avoid automating

The AI ROI Equation

• Hours released × loaded cost + throughput + risk-adjusted opportunity value − total system cost
• An illustrative capacity map across six functions
• A worked example: $110,360 gross benefit, 3.4x benefit/cost
• The 5-question investment test scored out of 10

What you'll walk away with

• A defensible baseline cost for your current process
• A clear line between automation, AI assistance and human judgement
• An ROI equation your CFO will accept
• Six honest reasons to walk away from an implementation