This calculator estimates monthly LLM API cost, cost per investigation, labor savings from faster ticket resolution, net monthly savings, and a side-by-side comparison of per-seat vs per-resolution vs per-investigation pricing. Enter your ticket volume, model, and agent cost — results update live. Submit your email to unlock the full savings breakdown and pricing comparison.
| Pricing model | Billable events | Est. monthly cost | Cost driver |
|---|---|---|---|
| Per seat (10 agents × $50) | N/A | $500/month | Agent headcount — fixed |
| Per resolution ($0.99 · Intercom Fin) | 600 | $594/month | 60% of 1,000 tickets |
| Per investigation (Altor — usage-based) | 300 | Contact for quote | 30% of 1,000 tickets |
How the calculator works
The calculator uses three formulas:
At default inputs — 1,000 tickets, GPT-4o mini, 2,000 input / 500 output tokens:
API cost = (2000/1,000,000 × $0.15) + (500/1,000,000 × $0.60) = $0.0003 + $0.0003 = $0.0006/ticket × 1,000 = $0.60/month
Labor savings = 300 investigations × (15/60) × $45 = 300 × 0.25 × $45 = $3,375/month
Net savings = $3,375 − $0.60 = $3,374.40/month
The pricing model comparison table calculates per-resolution cost as: resolution count × $0.99 (Intercom Fin's published rate). Per-investigation shows the investigation count — the number Altor's usage-based pricing would apply to.
What the results tell you
If net monthly savings is positive: the labor time recovered from faster investigations exceeds the API cost at current inputs. If net savings is negative: either the minutes-saved input is too low (under 5 minutes doesn't justify API overhead) or the model + token combination is expensive relative to your agent hourly rate. Try GPT-4o mini or reduce token counts.
The pricing model comparison reveals where vendor markup lives. At default inputs, the raw API cost is $0.60/month. Per-resolution billing at the same volume adds $594/month. That gap — $593.40 — is what vendor pricing adds on top of the underlying API cost.
Bring your actual ticket volume, investigation rate, and escalation patterns. We'll show you the cost breakdown against Altor's per-investigation pricing in 30 minutes.
Frequently asked questions
How is monthly AI API cost calculated?
Monthly API cost = (input tokens ÷ 1,000,000 × input $/MTok + output tokens ÷ 1,000,000 × output $/MTok) × monthly tickets. At GPT-4o mini (2,000 in / 500 out, 1,000 tickets): ($0.0003 + $0.0003) × 1,000 = $0.60/month. GPT-4o at the same volume costs $10/month.
How are labor savings calculated?
Labor savings = investigations × minutes saved ÷ 60 × agent hourly rate. At defaults (300 investigations/month, 15 minutes saved, $45/hr): 300 × 0.25 × $45 = $3,375/month. Set "minutes saved" to your actual pre-AI investigation time for an accurate estimate — not an optimistic one.
Why is net savings sometimes negative?
Net savings = labor savings − API cost. If API cost exceeds labor savings, net savings is negative. This happens when investigation time savings are very short (under 5 minutes) or when an expensive model runs on high token volumes with a low agent rate. Switch to GPT-4o mini or increase the minutes-saved input to match real pre-AI investigation time.
Does this calculator include vendor pricing?
The API cost shows raw LLM API costs only. The pricing model comparison table shows per-seat ($500 for 10 agents at $50 each), per-resolution ($0.99 × resolution count — Intercom Fin's published rate), and per-investigation (Altor — usage-based, contact for quote). The gap between API cost and vendor cost is the vendor's operating margin and service cost.