AI ROI Worksheet Template
A practical worksheet for calculating the return on investment of an AI initiative. Guides you through cost identification, benefit quantification, sensitivity analysis, and payback period calculation. Designed for finance teams and project owners who need to justify AI spend.
Overview
What's included
Cost Inventory
Cost Inventory
Project name: Assessment period: years Currency: £
One-Off Costs
| Category | Item | Cost | Notes |
|---|---|---|---|
| Technology | AI platform / licence setup | £ | |
| Technology | Infrastructure provisioning | £ | |
| Development | Custom development / integration | £ | |
| Development | Data preparation & cleaning | £ | |
| People | Hiring (new roles) | £ | |
| People | External consultancy | £ | |
| Change | Training programme | £ | |
| Change | Process redesign | £ | |
| Other | Contingency ( %) | £ | |
| Total one-off costs | £___ |
Recurring Annual Costs
| Category | Item | Year 1 | Year 2 | Year 3 |
|---|---|---|---|---|
| Technology | SaaS / API fees | £ | £ | £ |
| Technology | Cloud compute / storage | £ | £ | £ |
| People | AI team salaries (portion) | £ | £ | £ |
| People | Ongoing training | £ | £ | £ |
| Operations | Monitoring & maintenance | £ | £ | £ |
| Operations | Support / helpdesk | £ | £ | £ |
| Total recurring | £___ | £___ | £___ |
Total Cost of Ownership
| Year 1 | Year 2 | Year 3 | Total | |
|---|---|---|---|---|
| One-off costs | £ | — | — | £ |
| Recurring costs | £ | £ | £ | £ |
| Total | £___ | £___ | £___ | £___ |
Benefit Quantification
Benefit Quantification
Direct Financial Benefits
For each benefit, estimate the annual value and assign a confidence level.
| # | Benefit | Calculation Method | Annual Value | Confidence |
|---|---|---|---|---|
| 1 | Labour savings | hours/week x £ /hr x 52 weeks | £ | High / Med / Low |
| 2 | Error reduction | errors/month x £ /error x 12 months | £ | High / Med / Low |
| 3 | Revenue uplift | % increase x £ baseline revenue | £ | High / Med / Low |
| 4 | Faster processing | days saved x £ /day opportunity cost | £ | High / Med / Low |
| 5 | Customer retention | % churn reduction x £ avg customer value | £ | High / Med / Low |
| Total direct benefits | £___ |
Confidence-Adjusted Benefits
| Confidence Level | Weight | Applied To |
|---|---|---|
| High | 90% | Benefits you can calculate from existing data |
| Medium | 60% | Benefits based on reasonable assumptions |
| Low | 30% | Benefits that are possible but uncertain |
Confidence-adjusted annual benefit: £
Indirect Benefits (Not Quantified)
- Improved customer experience and brand perception
- Better employee satisfaction and retention
- Enhanced data-driven culture
- Competitive positioning
- Future AI capability building
ROI Calculation & Sensitivity Analysis
ROI Calculation
Summary Table
| Year 1 | Year 2 | Year 3 | |
|---|---|---|---|
| Total costs | £ | £ | £ |
| Total benefits (confidence-adjusted) | £ | £ | £ |
| Net benefit | £___ | £___ | £___ |
| Cumulative net benefit | £ | £ | £ |
Key Metrics
- Simple ROI (Year 1): (Net benefit / Total cost) x 100 = %
- 3-Year ROI: (3-year net benefit / 3-year total cost) x 100 = %
- Payback period: months
- NPV (discount rate: ___%): £
Sensitivity Analysis
What happens if our assumptions change?
| Scenario | Benefit Adjustment | Cost Adjustment | 3-Year ROI | Payback |
|---|---|---|---|---|
| Best case | +20% | -10% | % | months |
| Expected case | 0% | 0% | % | months |
| Worst case | -30% | +20% | % | months |
Break-Even Analysis
The project breaks even if we achieve at least % of the expected benefits. This requires a minimum annual benefit of £ against total annual costs of £ .
Recommendation
Based on the analysis above, the recommendation is:
- Proceed — Strong positive ROI even in worst case
- Proceed with caution — Positive ROI in expected case, negative in worst case
- Reassess — Marginal ROI; consider reducing scope or costs
- Do not proceed — Negative ROI in expected and worst case scenarios
Instructions
How to use this template
List all costs comprehensively
Include every cost category, even small ones. Understating costs is the most common reason ROI projections miss the mark.
Quantify benefits with evidence
For each benefit, document the calculation method and data source. Use existing operational metrics as the basis wherever possible.
Apply confidence weightings
Be honest about which benefits are certain and which are speculative. Confidence-adjusted figures are more credible than optimistic projections.
Run the sensitivity analysis
Test what happens if costs increase by 20% and benefits decrease by 30%. If the project still makes sense, it is robust.
Present the one-page summary
Use the ROI summary to communicate the key figures: total investment, expected return, payback period, and confidence level.
Watch Out
Common mistakes to avoid
FAQ
Frequently asked questions
Use your organisation's weighted average cost of capital (WACC) or hurdle rate. If you do not have one, 8-12% is a common range for technology investments.
Start by measuring the current process manually for 2-4 weeks to establish a baseline. If that is not possible, use industry benchmarks or analogous projects as a starting point, and label the confidence as 'Low'.
Yes. Include the cost of proofs of concept, failed experiments, and team learning time. These are legitimate project costs that should be captured.
Most organisations target 100-300% ROI over 3 years (2-4x return). However, early-stage AI projects may have lower financial ROI but high strategic value in building capability.
Model benefits as increasing year-over-year as the AI system improves with more data and user adoption. A common pattern is 50% of full benefit in Year 1, 80% in Year 2, and 100% in Year 3.
Need a custom AI template?
Our team can build tailored templates for your specific business needs. Book a free strategy call.