The ROI of an AI agent on WhatsApp is calculated using three inputs: (1) monthly conversation volume, (2) human agent cost per conversation handled, and (3) the self-resolution rate achieved by the AI agent. Subtracting the platform cost plus the target cost from these figures gives you the net savings. For typical SMEs, the AI agent pays for itself in 30-90 days and delivers a 3-10x return in the first year.
ROI Formula
Annual ROI = (Savings in human hours + Increase in revenue) - Annual platform cost
Components of savings
Savings in human hours:
Annual_savings = Monthly_conversations × 12 × Auto-resolution_rate × Cost_per_human_conversation
Increased income:
Annual_increase = Monthly_conversations × 12 × Conversion_rate × Average_ticket × AI_lift
Typical inputs
| Variable | Small SME | medium-sized SME | Mid-market |
|---|---|---|---|
| Conversations/month | 500 | 3.000 | 10.000 |
| Cost per human conversation | $5 USD | $5 USD | $5 USD |
| Auto-resolution rate (well-configured AI) | 65% | 70% | 75% |
| Current conversion rate | 5% | 8% | 12% |
| Average ticket | $200 USD | $300 USD | $500 USD |
| AI-powered lift in conversion | +30% | +35% | +40% |
Calculation: Small SME
Annual savings:
- Conversations/year: 500 × 12 = 6,000.
- Self-solved: 6,000 × 0.65 = 3,900.
- Cost avoided: 3,900 × $5 = $19,500 USD/year.
Annual income increase:
- Current conversions: 6,000 × 0.05 = 300/year.
- Conversions with AI: 6,000 × (0.05 × 1.30) = 390/year.
- Additional sales: 90 × $200 = $18,000 USD/year.
Platform cost:
- Aurora IA $179 USD/month × 12 = $2.148 USD/year.
- Target rates: ~$300 USD/year.
- Total: $2.448 USD/year.
ROI: ($19.500 + $18.000) – $2.448 = $35.052 USD net savings Year 1 (15× return).
Calculation: Medium-sized SME
Annual savings:
- Conversations/year: 36,000.
- Self-solved: 25,200.
- Cost avoided: $126,000 USD/year.
Annual income increase:
- Current conversions: 2,880.
- With AI: 3,888.
- Additional sales: 1,008 × $300 = $302,400 USD/year.
Platform cost:
- Aurora IA Plus + 5 users = $400 USD/month × 12 = $4,800.
- Goal: ~$1,500 USD/year.
- Total: $6,300 USD/year.
ROI: ($126.000 + $302.400) – $6.300 = $422.100 USD net savings Year 1 (67× return).
Calculation: mid-market
Annual savings:
- Conversations/year: 120,000.
- Self-solved: 90,000.
- Cost avoided: $450,000 USD/year.
Annual income increase:
- Current conversions: 14,400.
- With AI: 20,160.
- Additional sales: 5,760 × $500 = $2.880.000 USD/year.
Platform cost:
- Aurora IA Plus + 50 users = ~$1,000 USD/month × 12 = $12,000.
- Goal: ~$8,000 USD/year.
- Total: $20,000 USD/year.
ROI: ($450.000 + $2.880.000) – $20.000 = $3,310,000 USD net savings Year 1 (165× return).
Variables that affect ROI
Self-resolution rate
- Incorrectly configured: 30-40%.
- Properly configured: 60-70%.
- Excellent with full RAG: 70-80%.
Every 10 points of auto-resolution changes the ROI dramatically.
Current conversion rate
Businesses with low conversion rates (1-3%) have much more upside; businesses with high conversion rates (15%+) have less additional lift.
Average ticket
Higher-ticket prices mean higher absolute ROI. B2B with high-ticket prices = higher ROI than B2C with low-ticket prices.
Payback time
- Small SME: 30-60 days.
- Medium-sized SME: 7-15 days.
- Mid-market: 1-5 days.
ROI Summary Table
| Company | Investment Year 1 | Net Savings Year 1 | ROI multiplier | Payback |
|---|---|---|---|---|
| Small SME (500 conv/month) | $2.448 USD | $35.052 USD | 15× | 30-60 days |
| Medium-sized SME (3,000 conversions/month) | $6.300 USD | $422.100 USD | 67× | 7-15 days |
| Mid-market (10,000 converts/month) | $20,000 USD | $3,310,000 USD | 165× | 1-5 days |
Errors in calculating ROI
- Ignore the lift on conversion. Saving human hours is only half the story.
- Underestimate self-resolution rate. Platforms with real AI (GPT-5 + RAG) reach 70%+; flow builders ~30-40%.
- Do not include hidden costs. Conversational markup, AI add-ons, and ongoing engineering can destroy ROI.
- Compare with the current non-optimized situation. The benchmark should be "what would happen if the human agent used better processes".
Why Aurora Inbox maximizes ROI
Aurora Inbox combines a high auto-resolution rate (GPT-5 + native RAG + catalog + scheduling) + predictable cost with no markup + native integrations. For Latin American SMEs, it's the combination that delivers the highest ROI.
Frequently Asked Questions
When should I expect to see ROI?
Typically 30-90 days for SMEs. Mid-market 7-30 days.
What is the average ROI of an AI agent?
3-15x return in the first year for typical SMEs. Mid-market 50-150x return.
How do I measure the actual self-resolution rate?
Aurora Inbox displays it on the dashboard. Closed conversations without human escalation / total conversations.
What happens if the auto-resolution is low?
Refine the RAG (more documents), the prompt (better personality), and the tools (more things the agent can do). 70%+ is achievable with a good setup.
Does my specific vertical have a different ROI?
Yes. High-ticket verticals (B2B, real estate, automotive) have a higher ROI in absolute value.
Does the calculator apply to B2B cases?
Yes, by adjusting for average ticket size (much higher) and volume (typically lower). Same principle.
