Enterprise AI investments—platform licenses, implementation, training, support—typically range $150,000 to $500,000+ annually. Boards expect return on investment. Most AI business cases justify investment through efficiency (cost savings) and revenue impact (dollars raised, programs enabled). Quantifying these requires systematic methodology.
The challenge: most AI benefits are intangible or indirect. Automation saves staff time, but salary expense doesn't drop (staff are reassigned, not eliminated). Grant research improvements might increase applications, which might increase grants won, which might enable more programs—but the causal chain is long. Quantifying requires both rigor and candor about assumptions.
The most defensible AI ROI calculations carefully estimate cost savings and revenue impact, document all assumptions, and present both conservative and optimistic scenarios. This transparency builds credibility with boards and funders.
Cost savings emerge through automation reducing labor. Quantifying requires understanding current costs and projected reductions.
List processes consuming staff time and estimate hours: grant research (5 hours per grant x 40 grants/year = 200 hours), proposal writing (8 hours per proposal x 40 = 320 hours), donor list maintenance (50 hours/year), proposal quality review (3 hours per proposal x 40 = 120 hours). Total: 690 hours annually.
For significant initiatives, conduct time-motion studies: observe actual work, measure how long tasks actually consume, validate assumptions. Time-motion accuracy matters—if you assume proposal writing takes 8 hours but actually takes 12, your savings calculations are wrong. For enterprise initiatives, invest time in accurate measurement.
Translate hours to dollars. A grants person earning $65,000 annually costs approximately $31/hour (salary / 2080 hours worked annually). But fully-loaded cost (including benefits, overhead) is typically 25-30% higher: $38-40/hour. Use fully-loaded cost for business cases.
If AI automates 300 hours annually at $40/hour, the cost savings is $12,000 annually. Across 5 grants staff, potential savings reach $60,000 annually.
AI typically reduces task time 20-60%, not 100%. Proposal writing might be 40% faster (AI drafts, staff refines). Grant research might be 50% faster (AI suggests opportunities, staff evaluates). Use realistic reduction percentages, not fantasy numbers. Conservative business cases underestimate AI's speed benefit.
For a process you're considering automating with AI: Estimate current hours (use time-motion study if significant). Calculate fully-loaded hourly cost. Estimate realistic time reduction (30-50% is reasonable). Calculate annual savings. This becomes the efficiency baseline of your business case.
Beyond direct time savings, AI improves quality, which has economic value.
If proposal quality improves (better alignment with funder requirements, fewer technical errors), acceptance rates increase. Quantify: Current acceptance rate 30%, post-AI 35% (5 percentage point improvement). If you submit 50 proposals currently (15 accepted), improvement means 2 additional grants. If average grant size $50,000, that's $100,000 additional revenue annually.
AI improving data quality reduces rework. If donor database quality improves, reports become more reliable, reducing time spent resolving inconsistencies. Estimate rework reduction: perhaps 20% fewer hours spent on data reconciliation.
AI identifying compliance issues before they become problems prevents expensive fixes. If AI catches grant requirement violations before submission, preventing rejected proposals, that's valuable. Quantify conservatively: perhaps 2-3 prevented rejection/resubmission cycles annually, saving 10-15 hours of rework.
Beyond cost savings, AI drives revenue by enabling growth nonprofits couldn't otherwise achieve.
AI-assisted grant research enables staff to identify and pursue more opportunities. If AI enables grants team to pursue 50 grants instead of 40 (without hiring additional staff), and current win rate is 30%, that's 5 additional grants won vs. previous 12. If average grant size $75,000, that's $375,000 additional revenue annually.
AI matching grants to programs improves fit, increasing acceptance rates. Calculate: if acceptance rate improves from 30% to 35% across all 50 pursued grants, that's 2.5 additional grants (at $75,000 each), $187,500 additional revenue annually.
AI identifying high-capacity donors who were undercultivated might increase giving. Estimate conservatively: identify 20 undercultivated major donors, increase their giving average $10,000 annually, $200,000 additional revenue.
ROI (Return on Investment) is calculated: (Benefits - Costs) / Costs x 100%.
Scenario: Cost savings $120,000 annually, revenue impact $300,000 annually. Total annual benefit: $420,000. First-year costs (implementation + licensing): $250,000. Ongoing annual costs: $100,000. Year one ROI: ($420,000 - $250,000) / $250,000 = 68% ROI in year one. Year two ROI: ($420,000 - $100,000) / $100,000 = 220% ROI in steady state.
How long until benefits exceed costs? Using above example: Year one costs $250,000, year one benefits $420,000. Payback occurs within year one. For initiatives where payback extends multiple years, calculate when cumulative benefits exceed cumulative costs.
Net Present Value (NPV) accounts for time value of money (a dollar today is worth more than a dollar next year). Internal Rate of Return (IRR) is the discount rate where NPV equals zero. For most nonprofits, simpler metrics (payback period, simple ROI) are sufficient. Boards appreciate straightforward numbers over complex financial engineering.
The future is uncertain. Sensitivity analysis tests how results change if assumptions vary.
Present three scenarios: Conservative (assumptions pessimistic—smaller time savings, lower revenue impact), Base (realistic assumptions), Optimistic (favorable scenario). Example: Conservative ROI 40%, Base 80%, Optimistic 120%. This range acknowledges uncertainty while showing likelihood of positive return across scenarios.
Identify which assumptions most affect ROI. If acceptance rate improvement drives large revenue impact, sensitivity on that assumption matters more than sensitivity on implementation cost. Vary high-impact assumptions across reasonable ranges.
The most valuable business case documents assumptions explicitly, making tradeoffs transparent.
State: "We assume AI reduces proposal writing time 40%, from current average 8 hours to 4.8 hours per proposal." This clarity allows others to evaluate whether assumption is reasonable. If others believe 40% reduction is optimistic, you can recalculate with different assumption.
Which assumptions, if wrong, change the conclusion? If ROI turns negative if acceptance rate doesn't improve at all, that's critical assumption worth investigating. If ROI remains positive even if time savings are only 20% (instead of assumed 40%), you have lower-risk case.
A national education nonprofit serving 100,000+ students deployed AI-assisted grant research and proposal writing. Year-one results:
Grants team of 6 people. AI reduced proposal writing time 45% (from 8 hours to 4.4 hours per proposal) and grant research time 50% (from 5 hours to 2.5 hours per opportunity). Combined, AI saved ~2,000 hours annually. At fully-loaded cost of $45/hour, cost savings: $90,000 annually.
Efficiency gains enabled team to pursue 60 grants (vs. previous 40). Win rate improved from 25% to 32% (better targeting). Year zero: 40 grants × 25% = 10 grants at average $300,000 each = $3,000,000. Year one: 60 grants × 32% = 19.2 grants = $5,760,000. Additional revenue: $2,760,000.
Year-one benefits: $90,000 cost savings + $2,760,000 revenue = $2,850,000. Year-one costs (implementation, licensing, training): $750,000. ROI: ($2,850,000 - $750,000) / $750,000 = 280% year-one ROI. Payback in month three. Ongoing annual costs $150,000, with benefits $2,850,000, creates 1,800% steady-state ROI.
This exceptional ROI required: sustained improvement in proposal quality (win rate didn't decline despite volume increase), staff redeployment (freed time used for new grant pursuit, not eliminated), and organizational discipline to pursue additional opportunities systematically. Less favorable scenarios with 20% time savings and unchanged win rate still showed 40% year-one ROI—positive across scenarios.
Quantifying AI impact requires careful measurement: identifying automatable work through time-motion studies, estimating realistic time reductions (not 100%), calculating fully-loaded labor costs, quantifying quality improvements and revenue impacts, and calculating ROI systematically. Sensitivity analysis acknowledging uncertainty strengthens credibility. Conservative business cases that remain positive across scenarios provide boards confidence in investment. When documented transparently with all assumptions explicit, AI business cases convince funders and leadership that investments deliver substantial organizational value.
Enroll in CAGP Level 4 to deepen your skills in organizational-scale AI implementation, measurement, and strategy.
Explore CAGP Levels