The AI proposals that get rejected aren't the ones with weak technology — they're the ones with weak math. Partners don't vote against AI because they're technophobes. They vote against it because nobody showed them a cost-benefit analysis that accounts for the real costs, the real savings, and the realistic timeline to payback.

This template gives you the framework to build that analysis. It covers the four cost categories partners always ask about, the three revenue impact areas they don't think about, and the calculation methodology that turns 'we should invest in AI' into 'here's exactly what we'll spend, save, and earn — and by when.'


Direct Costs: Everything You'll Actually Spend

Partners distrust AI proposals that only show the license fee. They've been burned by 'hidden costs' before. List everything. Software licensing: annual subscription cost for the specific plan/tier you need. Get a written quote, not an estimate from a pricing page. For most legal AI tools, expect $30,000-$200,000 annually depending on firm size and tool category. Implementation: vendor professional services for setup, configuration, and data migration. Typically $10,000-$50,000 for mid-size firms. Some vendors include this in the license; most don't. Integration: costs to connect the AI tool with your existing systems (document management, practice management, billing). If IT handles this internally, calculate their time at fully-loaded rates. If you hire a consultant, get a fixed-fee quote. Typical range: $5,000-$25,000. Training: both the direct cost (vendor training sessions, typically $2,000-$10,000) and the indirect cost (attorney time spent in training — calculate hours x billing rate for the opportunity cost). Ongoing maintenance: annual support fees (often included in licensing), IT time for updates and troubleshooting, and the cost of one person spending 5-10 hours/month managing the tool. Year 1 total formula: Licensing + Implementation + Integration + Training + Maintenance. Year 2-3 total: Licensing + Maintenance (implementation and training costs don't recur). Always present the 3-year total cost of ownership, not just year 1. Partners think in multi-year investments.

Time Savings: Convert Hours Into Dollars

Time savings are the most cited AI benefit and the most poorly calculated. 'We'll save 2,000 hours' means nothing without context. Here's how to make the math compelling. Step 1 — Identify the specific tasks. Don't say 'contract review will be faster.' Say 'initial review of incoming third-party NDAs currently takes an average of 2.5 hours per contract. With AI-assisted review, comparable firms report an average of 45 minutes per contract.' Step 2 — Calculate annual volume. 'We processed 340 NDAs last year.' Step 3 — Calculate hours saved. 340 contracts x 1.75 hours saved = 595 hours saved annually. Step 4 — Convert to dollars using the RIGHT rate. This is where most analyses go wrong. Don't use the attorney's billing rate — use the fully-loaded internal cost rate (salary + benefits + overhead, divided by billable hours). For an associate at a mid-size firm, this is typically $100-$175/hour. For partners, $175-$300/hour. 595 hours x $125/hour = $74,375 in annual capacity created. Step 5 — Distinguish between capacity created and cost eliminated. Partners know you're not firing associates. Frame it as: 'This frees 595 associate hours annually — equivalent to 15% of one associate's capacity redirected to billable client work.' If those hours convert to additional billable work at standard rates, the revenue impact is 595 x $350 = $208,250 in billable capacity.

Revenue Impact: The Numbers That Convince Partners

Cost savings get attention. Revenue impact gets approval. Three revenue impact areas to calculate. Increased billable capacity: hours freed from non-billable or low-value tasks that become available for billable client work. Use a conservative conversion rate — assume 50% of freed hours actually convert to billable work. On the NDA example: 595 hours x 50% conversion x $350 billing rate = $104,125 in potential additional revenue. Faster turnaround wins more work: if AI lets you deliver work products 40% faster, quantify the competitive advantage. Survey your business development team: 'How many pitches last year did we lose partly because of capacity constraints or timeline concerns?' Even 2-3 additional matters won could represent $200,000-$500,000 in new revenue. Client retention from quality improvement: AI-assisted research and review catches errors that manual processes miss. First-pass quality improvements reduce client complaints, rework, and the silent attrition where clients leave without telling you why. Assign a conservative dollar value to retention improvement — even a 1% improvement in client retention on a $10M book of business represents $100,000 in preserved revenue. Total revenue impact formula: additional billable capacity + estimated new matter revenue + client retention value. Use conservative estimates. Partners respect restraint more than optimism.

Hidden Costs and Risk Factors: What Partners Will Ask About

Every partner meeting will have someone who asks 'what about the risks?' Have the answers ready. Adoption risk: what if attorneys don't use it? Mitigant: the 90-day pilot with defined adoption targets. If adoption is below 50% at day 60, you kill it. Cost of failure: the pilot cost ($25,000-$60,000), not the annual subscription. Security risk: what if there's a data breach or privilege waiver? Mitigant: completed security questionnaire, SOC 2 Type II verification, and contractual data protections. Quantify the risk: your cyber insurance covers $X. The vendor carries $Y in coverage. Residual risk is $Z. Quality risk: what if AI produces incorrect output that reaches a client? Mitigant: mandatory human verification policy. No AI output leaves the firm without attorney review. Quantify: the cost of one malpractice incident (deductible + reputational damage + time spent) versus the cost of the verification process. Opportunity cost: what happens if we don't invest? This is the number most analyses leave out. Calculate the cost of maintaining the status quo: continued outside counsel spend for work AI could handle, continued capacity constraints limiting billable hours, and the competitive disadvantage as peer firms adopt AI. The cost of inaction should be on the same page as the cost of investment.

The One-Page Summary: How to Present It

Everything above is the appendix. The actual presentation is one page. Section 1 — Investment (3 lines): Year 1 total cost. Year 2-3 annual cost. 3-year total cost of ownership. Section 2 — Returns (4 lines): Annual time savings in dollars. Annual revenue impact (billable capacity + new matters + retention). Annual risk reduction value. Total annual return. Section 3 — Payback (2 lines): Payback period in months. 3-year ROI percentage. Section 4 — Risk mitigation (3 lines): Maximum downside (pilot cost if killed). Security posture (SOC 2, contractual protections). Kill criteria (specific conditions under which you stop). Example: Investment: $185,000 (Year 1), $120,000 (Years 2-3), $425,000 (3-year TCO). Returns: $74,375 (time savings) + $104,125 (revenue impact) + $50,000 (risk reduction) = $228,500 annual return. Payback: 9.7 months. 3-year ROI: 61%. Maximum downside: $45,000 pilot cost. That's the slide. One page. Every number sourced from the detailed analysis behind it. Partners read the summary and vote. They reference the detail only if the summary is compelling.

The Bottom Line: The AI cost-benefit analysis that convinces partners has five elements: comprehensive direct costs (no hidden fees), time savings converted to dollars at the right rate (internal cost, not billing rate), revenue impact calculated conservatively, hidden costs and risk factors with quantified mitigants, and a one-page summary with payback period and 3-year ROI. Lead with the summary. Put the detail in the appendix. Use real numbers from your firm, not industry averages. The proposal that shows 'here's our specific math' beats 'here's what AI can do' every time.

AI-Assisted Research. This piece was researched and written with AI assistance, reviewed and edited by Manu Ayala. For deeper takes and the perspective behind the research, follow me on LinkedIn or email me directly.