$0.99 Per Resolution Sounds Cheap Until You Do the Math

A dollar per solved conversation. That's Intercom Fin's pitch. Less than a cup of coffee. Cheaper than a human agent handling the same ticket. On the surface, it's a no-brainer.

Then you get your first invoice.

Intercom Fin doesn't exist in a vacuum. You can't just pay $0.99 per resolution and call it a day. Fin requires an Intercom plan. And those plans have seat-based pricing that adds up fast, especially once your team grows past 2 or 3 people.

Let's run the actual numbers at different scales.

The Base Plan You Can't Skip

Fin is only available on Intercom's paid plans. The cheapest option is the Essential plan at $29/seat/month (billed annually). Monthly billing bumps that to $39/seat/month.

Every person on your team who needs access to the Intercom inbox is a paid seat. Support agents, managers who want to monitor conversations, anyone who responds to escalations. They all count.

So before Fin resolves a single conversation, you're already paying for seats.

200 Resolutions Per Month (Early Stage Startup)

A small SaaS product, maybe 500 users, getting around 300 support messages monthly. Fin handles about 200 of them automatically.

With Intercom (2 team members on Essential, annual billing):

• Seats: 2 x $29 = $58/month
• Fin resolutions: 200 x $0.99 = $198/month
• Total: $256/month

With Supp (pay-per-use, no seats):

• Resolutions with actions: 200 x $0.30 = $60/month
• Total: $60/month

Difference: $196/month. That's $2,352/year.

At this scale, the seat cost is almost as much as the AI cost. And you only have 2 people.

500 Resolutions Per Month (Growing Team)

Your product has traction. You're getting 700+ support messages a month. You've hired a third support person. Fin handles 500 conversations.

With Intercom (3 seats on Essential, annual billing):

• Seats: 3 x $29 = $87/month
• Fin resolutions: 500 x $0.99 = $495/month
• Total: $582/month

With Supp (no seats):

• Resolutions with actions: 500 x $0.30 = $150/month
• Total: $150/month

Difference: $432/month. That's $5,184/year. Enough to fund a contractor for a few months.

1,000 Resolutions Per Month (Scaling Up)

You're past product-market fit. Support volume is climbing. You've got 5 people who need inbox access, and Fin is handling 1,000 conversations monthly.

With Intercom (5 seats on Essential, annual billing):

• Seats: 5 x $29 = $145/month
• Fin resolutions: 1,000 x $0.99 = $990/month
• Total: $1,135/month

With Supp:

• Resolutions with actions: 1,000 x $0.30 = $300/month
• Total: $300/month

Difference: $835/month. That's $10,020/year.

2,000 Resolutions Per Month (Established Business)

Now things get interesting. You've got a real support operation. 8 team members, 3,000+ messages a month, Fin handling 2,000 of them.

With Intercom (8 seats on Essential, annual billing):

• Seats: 8 x $29 = $232/month
• Fin resolutions: 2,000 x $0.99 = $1,980/month
• Total: $2,212/month

With Supp:

• Resolutions with actions: 2,000 x $0.30 = $600/month
• Total: $600/month

Difference: $1,612/month. That's $19,344/year.

At this point you're saving enough to hire a full-time junior support agent.

What Counts as a "Resolution" Anyway?

This matters more than people think. Intercom counts a Fin resolution when the AI answers a customer's question and the customer doesn't need to talk to a human. If the customer asks a follow-up question in the same conversation and Fin handles that too, it's still one resolution. Good.

But if a customer comes back the next day with a related question, that's a new conversation and a new resolution charge. And if Fin tries to answer but the customer still asks for a human, you don't get charged for the resolution, but you do still pay the seat cost for the human who handles it.

Supp counts differently. A classification costs $0.20. If the AI takes an action (creates a ticket, processes a refund, sends a notification), that's $0.30. You're paying for work done, not conversations closed.

The Features You're Actually Paying For

Fin is genuinely good at what it does. It reads your help center articles, generates natural language responses, and handles multi-turn conversations. If a customer says "I can't log in" and then clarifies "I'm using Google SSO," Fin can follow that thread.

Supp works differently. It classifies the intent (account access issue, SSO login failure) and triggers a pre-configured action (send SSO troubleshooting steps, create a bug ticket if the issue persists). It's faster (100 to 200ms) but doesn't generate freeform responses.

The question is whether freeform conversation is worth 3.3x the per-resolution price plus seat fees.

For knowledge-heavy products where customers ask unique, varied questions, Fin's approach makes sense. The LLM can synthesize answers from multiple help articles in ways that a classification model can't.

For products where 80% of support messages fall into predictable categories (billing, password resets, refund requests, order status, bug reports), classification handles it just as well at a fraction of the cost.

The Annual Contract Catch

One thing that doesn't show up in the per-resolution math: Intercom's best pricing requires annual billing. If you want monthly billing, the seat price jumps from $29 to $39. At 5 seats, that's an extra $50/month just for the flexibility of paying monthly.

Supp has no contracts. No annual commitments. You pay for what you use this month. If your volume drops, your bill drops. If you want to stop, you stop.

Running Your Own Numbers

Here's the formula:

Intercom monthly cost = (seats x $29) + (resolutions x $0.99)

Supp monthly cost = resolutions x $0.30 (with actions) or resolutions x $0.20 (classification only)

Plug in your team size and expected monthly resolution volume. The crossover point where Intercom becomes worth the premium is when you genuinely need the LLM-powered conversation ability and have complex, unpredictable support questions that can't be handled with classification and pre-built responses.

For most teams under 2,000 resolutions/month, the math favors pay-per-use.