Every client asks the same question: “How do I know if this will pay for itself?” Here’s the math I use for every AI automation proposal.
The Formula
ROI = (Time Saved × Hourly Cost × 12) ÷ Project Cost
If a chatbot saves your support team 20 hours/week, and your support agents cost $20/hour including benefits, that’s $20,800/year in savings. If the chatbot costs $8,000 to build and $200/month to run, your first-year ROI is 2.1x. It pays for itself in under 6 months.
Real Project Examples
Customer Support Chatbot
- Time saved: 25 hours/week (answering FAQ, routing tickets)
- Agent cost: $22/hour
- Annual savings: $28,600
- Build cost: $6,000–12,000
- Monthly API cost: $100–300
- ROI: 2–4x first year
Data Pipeline Automation
- Time saved: 15 hours/week (manual data entry, report generation)
- Analyst cost: $35/hour
- Annual savings: $27,300
- Build cost: $8,000–15,000
- Monthly hosting: $50–150
- ROI: 1.5–3x first year
Lead Qualification Bot
- Revenue impact: 30% more qualified leads reaching sales team
- Sales team time saved: 10 hours/week on unqualified calls
- Annual impact: depends on deal size. For a business closing $5K deals, even 2 extra conversions/month = $120K/year
- Build cost: $5,000–10,000
What to Watch Out For
- API costs scale with usage — estimate worst-case, not average
- Maintenance is real: plan for 10–15% of build cost annually
- The biggest ROI comes from processes done frequently, not occasionally
Rule of thumb: if a task takes more than 10 hours/week and follows a pattern, it’s almost certainly worth automating. If it’s under 2 hours/week, the ROI rarely justifies the build.