Per-resolution charges for every deflected ticket. Per-investigation charges only for tickets requiring active AI investigation. For teams where 60-70% of volume is tier-1 deflection and 20-30% needs investigation, per-investigation pricing can cost 2-4× less per month because the billable event count is substantially lower than resolution count.
What per-resolution charges for
Per-resolution pricing bills for every conversation the AI closes without sending it to a human. Intercom Fin publishes the clearest reference point here at $0.99 per resolution, and Intercom reports a 56% average containment rate. That gives support leaders a direct way to estimate monthly spend from ticket volume alone.
On 1,000 tickets per month, 56% containment produces 560 billable events. At $0.99 each, that lands at $554.40 per month. At 5,000 tickets, the same containment rate creates 2,800 billable resolutions, or $2,772 per month. The cost line rises directly with ticket volume and how often the bot closes the conversation.
What counts as a billable resolution? Any conversation Fin answers and closes without transferring to a human agent. That includes FAQ replies, documentation lookups, and simple deflections where the user gets what they need inside the bot flow. What does not count is a conversation the bot attempts but cannot finish, then routes to a human. This model is built for deflection-first support operations.
What per-investigation charges for
Per-investigation pricing uses a different denominator. The billable event is an investigation, not a deflection. An investigation happens when AI actively queries production systems to diagnose a ticket: checking account state, tracing billing behavior, reviewing API usage patterns, or narrowing down probable root cause.
That matters because not every ticket needs that level of work. Tier-1 support volume such as FAQ answers, docs guidance, shipping updates, or simple status lookups usually does not require investigation. In B2B technical support, investigation rates often sit in the 20-40% range depending on product complexity, integration depth, and how much of the queue comes from technical users.
On 1,000 tickets per month, a 25% investigation rate produces 250 billable investigation events. That is far lower than the 560 billable events in the per-resolution example above. The structural difference is simple: per-resolution charges for closed conversations, while per-investigation charges for tickets that require system-aware diagnostic work.
The math head-to-head
Take a 1,000-ticket monthly queue for a B2B SaaS support team. Assume a mixed workload with 600 tier-1 tickets and 400 tickets that could require deeper analysis. Under per-resolution pricing, the cost depends on how many conversations the bot closes. Under per-investigation pricing, the cost depends on how many tickets trigger investigation work.
| Metric | Per-resolution ($0.99/Fin) | Per-investigation (Altor) |
|---|---|---|
| Billable event definition | Resolved without human transfer | Investigation completed |
| Billable events at 56% containment | 560 resolutions | — |
| Billable events at 25% investigation rate | — | 250 investigations |
| Monthly cost ratio | 560 × $0.99 = $554.40 | 250 × [contact for quote] |
| Cost driver | Volume × containment rate | Ticket complexity × investigation rate |
| Scales with | Every deflection | Investigation complexity |
Altor pricing is usage-based per investigation and not published, so the comparison point is event count first. If your queue produces far fewer investigation events than resolution events, the billing denominator already points to the likely lower-cost model. The remaining question is whether the per-event rate closes or preserves that gap.
Which costs less: depends on your investigation rate
The breakeven depends on three inputs: your investigation rate, the per-investigation rate, and your containment rate on deflectable tickets. Start with event counts before you compare contracts. Monthly tickets multiplied by containment rate gives per-resolution events. Monthly tickets multiplied by investigation rate gives per-investigation events.
A useful rule of thumb is this: if your investigation rate is under 30% and your containment rate is above 50%, per-investigation billing generates fewer billable events. That does not automatically make it cheaper, but it gives the pricing model a structural edge. Teams should compare quote to quote using their own queue mix rather than headline rates alone.
Fin vs Altor: not substitutes
Intercom Fin is optimized for deflection. Its job is to keep humans out of the loop on answerable tickets by using help center content, approved knowledge, and conversation flows. That makes it strong for tier-1 support volume.
Altor is optimized for investigation. Its job is to help teams resolve tickets that require querying production systems, pulling account state, and narrowing down what actually broke. That maps better to tier-2 and tier-3 support work.
- Fin: tier-1 tickets such as FAQ, docs, and simple lookup work → deflect.
- Altor: tier-2/3 tickets such as billing error, API failure, and account state issues → investigate.
For mixed queues, many teams will not treat them as direct replacements at all. They solve different layers of the support stack and can run in parallel.
Compare event counts before you compare vendors
Use your own queue mix to see whether you are buying deflection volume or investigation volume.
FAQ
Is per-resolution or per-investigation pricing cheaper?
It depends on your investigation rate and resolution rate. At 56% containment, 1,000 tickets generates 560 billable events under per-resolution. At 25% investigation rate, the same 1,000 tickets generates 250 billable events. Per-investigation wins if the per-event cost is comparable — get a quote against your actual investigation count.
How do you calculate whether per-resolution or per-investigation saves money?
Step 1: multiply monthly tickets by containment rate to get per-resolution event count. Step 2: multiply monthly tickets by investigation rate to get per-investigation event count. Step 3: multiply each by the respective per-event cost. The lower total wins. Use the AI Support ROI Calculator at /ai-cost-optimization/ai-support-roi-calculator/ to model your numbers.
What is a typical investigation rate for B2B support teams?
For B2B technical support with complex products such as APIs, billing systems, and integrations, investigation rates of 20-40% are common. Simple consumer support teams may see 5-15%. Investigation rate is the key variable — teams with high investigation rates benefit most from per-investigation pricing.
Does Intercom Fin handle technical investigation tickets?
Intercom Fin works within Intercom's knowledge base and conversation model. It does not query external production systems such as billing databases, API logs, or infrastructure monitoring. Tickets requiring root-cause investigation across multiple systems are typically routed to human agents by Fin.