AI customer support has two pricing models. One charges per resolution. The other charges a flat monthly fee. The difference between them can mean $200/month or $10,000/month for the same volume of conversations.
This is not an abstract comparison. We will use real prices, real math, and real scenarios so you can calculate exactly what each model costs at your scale.
The Two Pricing Models
Per-Resolution Pricing
You pay every time the AI successfully resolves a customer question without human involvement. The price per resolution varies by platform:
- Intercom Fin: $0.99 per resolution (Intercom Pricing, 2025)
- Help Scout AI: $0.75 per resolution (Help Scout Pricing, 2025)
The appeal is clear. You only pay when AI works. If the AI resolves 10 conversations in a month, you pay $9.90 with Intercom. Sounds reasonable.
The problem appears at scale.
Flat-Rate Pricing
You pay a fixed monthly fee that includes a set number of AI answers. Whether you use 50 or 2,500, the bill stays the same.
- Starter: $49/month, includes 2,500 AI answers
- Pro: $99/month, includes 10,000 AI answers
- Scale: $249/month, includes 40,000 AI answers
Helpable is one example of flat-rate AI pricing. Plans include the knowledge base, live chat, and team members at no extra cost.
The appeal here is predictability. You know what you will pay next month regardless of how many questions your customers ask.
The Real Math: Cost at Every Scale
Here is what each model actually costs. No hidden fees. No "contact sales" asterisks. Just multiplication.
Cost Comparison Table
| Monthly Resolutions | Intercom Fin ($0.99) | Help Scout AI ($0.75) | Flat rate example |
|---|---|---|---|
| 500 | $495 | $375 | $49 (Starter) |
| 1,000 | $990 | $750 | $49 (Starter) |
| 2,000 | $1,980 | $1,500 | $49 (Starter) |
| 5,000 | $4,950 | $3,750 | $99 (Pro) |
| 10,000 | $9,900 | $7,500 | $99 (Pro) |
These numbers assume only AI resolution costs. They do not include the base platform subscription for Intercom ($74/agent/month) or Help Scout ($50/user/month). The actual total is higher.
500 Resolutions Per Month
This is a small SaaS with moderate support volume.
Intercom: $495/month in AI fees + $74/agent/month base. For a team of 3, that is $495 + $222 = $717/month.
Help Scout: $375/month in AI fees + $50/user/month base. For a team of 3, that is $375 + $150 = $525/month.
Flat rate: $49/month flat. 2,500 AI answers included. Team of 3 included. $49/month total.
The difference: $717 vs $49. That is 14.6x more expensive on Intercom for the same 500 AI resolutions.
2,000 Resolutions Per Month
This is a growing SaaS with an active customer base.
Intercom: $1,980 in AI fees + $370 base (5 agents) = $2,350/month.
Help Scout: $1,500 in AI fees + $250 base (5 users) = $1,750/month.
Flat rate: $49/month flat. Still within the 2,500 Starter limit. $49/month total.
At 2,000 resolutions, Intercom costs 48x more than a flat-rate plan.
10,000 Resolutions Per Month
This is a scaled SaaS with a large, active customer base.
Intercom: $9,900 in AI fees + $740 base (10 agents) = $10,640/month.
Help Scout: $7,500 in AI fees + $500 base (10 users) = $8,000/month.
Flat rate: $99/month (Pro plan, 10,000 AI answers included, 10 team members). $99/month total.
At 10,000 resolutions, you are paying $10,640/month with Intercom or $99/month on a flat-rate plan. The annual difference is $126,492.
The 12-Month Projection
The cost gap widens dramatically over a year. Here is the cumulative spend for a team that starts at 500 resolutions and grows to 2,000 over 12 months.
Intercom Fin (Per-Resolution at $0.99)
Month 1-3: 500 resolutions/month = $1,485 total. Month 4-6: 1,000 resolutions/month = $2,970 total. Month 7-9: 1,500 resolutions/month = $4,455 total. Month 10-12: 2,000 resolutions/month = $5,940 total. Annual AI cost: $14,850. Add platform fees for 5 agents ($74 x 5 x 12 = $4,440). Grand total: $19,290/year.
Help Scout AI (Per-Resolution at $0.75)
Same growth pattern. Month 1-3: $1,125. Month 4-6: $2,250. Month 7-9: $3,375. Month 10-12: $4,500. Annual AI cost: $11,250. Add platform fees ($50 x 5 x 12 = $3,000). Grand total: $14,250/year.
Flat Rate ($49/month)
Month 1-12: $49/month on Starter (covers up to 2,500). Annual cost: $588. Team members included. Knowledge base included. Live chat included. No add-ons. $588/year total.
The difference between Intercom and a flat-rate plan over 12 months is $18,702. That is the salary of a part-time support agent.
The Hidden Trap: Better AI Means Higher Bills
This is the part per-resolution vendors do not advertise.
When you first deploy AI, your knowledge base is small. The AI resolves maybe 200 conversations per month. Your per-resolution bill is low. You think "this is affordable."
Then you improve. You write more articles. You fill knowledge gaps. You refine your content. Your AI starts resolving 500, then 1,000, then 2,000 conversations per month.
Your AI is working better. Your customers are happier. And your bill goes up. You are penalized for improving.
With flat-rate pricing, the opposite happens. As your AI improves, your cost per resolution goes down. At $49/month with 2,500 AI answers, your cost per resolution is $0.02. Compare that to $0.99.
When Per-Resolution Makes Sense
Per-resolution pricing is not always the wrong choice. It makes sense in one specific scenario:
Very low volume, under 100 resolutions per month. If your AI resolves 50 conversations per month, Intercom costs $49.50. That is roughly the same as most flat-rate starter plans. At this volume, per-resolution is fine because the total is small.
But the moment you cross 100 resolutions per month, flat-rate pricing wins. And since the entire point of AI support is to resolve more conversations, you will cross that threshold quickly.
What Per-Resolution Vendors Say (And What They Mean)
"You only pay for results." This sounds fair. But "results" means "the AI answered a question." Whether the answer was good or bad, you pay. A resolution is logged when the conversation ends without escalation, not when the customer is satisfied.
"Start small, scale as you grow." True. But the scaling is linear. Double your resolutions, double your bill. There is no volume discount meaningful enough to change the math.
"Aligned incentives." The vendor's incentive is to maximize resolutions. Your incentive is to provide good support. These are not the same thing. An AI that aggressively resolves conversations without properly helping customers counts as "success" in the per-resolution model.
"Enterprise-grade AI." The word "enterprise" implies you need the most expensive option for quality. But AI quality depends on the language model and your knowledge base, not the pricing model. A flat-rate tool using the same underlying AI model produces the same quality answers. You are paying more for the billing structure, not the technology.
The Real-World Impact on Team Decisions
Per-resolution pricing creates a perverse incentive inside your team. When every AI resolution costs money, team leads start asking: "Should we let AI handle this, or route it to a human?" That question should never be about cost. It should be about customer experience.
With flat-rate pricing, the calculation disappears. Let AI handle everything it can. Route to humans only when the question requires human judgment. Your team focuses on support quality, not cost optimization.
Total Cost of Ownership
The pricing comparison above only covers AI resolution fees. The full picture includes:
Platform subscription. Intercom starts at $74/agent/month. Help Scout at $50/user/month. Flat-rate tools like Helpable include team members in the plan price.
Knowledge base. Most per-resolution tools require a separate knowledge base tool or include one at higher tiers. Flat-rate tools often include a public knowledge base on all plans.
Live chat. Some tools charge extra for live chat (human agents). Flat-rate plans often include live chat at no additional cost.
Analytics and reporting. Basic analytics are typically included. Advanced reporting is often a higher-tier feature.
When you add these up, the gap between per-resolution and flat-rate pricing widens further.
The Budget Predictability Factor
CFOs and finance teams need predictable costs. Per-resolution pricing makes budgeting nearly impossible. A product launch, a seasonal spike, or a viral moment can double your AI resolution count in a single month.
With flat-rate pricing, the number is fixed. You can plan quarterly, annually, and multi-year without worrying about surprise invoices. This matters especially for startups and small businesses that operate on tight budgets.
According to the 2024 SaaS Spend Management Report by Zylo, unpredictable vendor costs rank among the top three frustrations for finance teams. Per-resolution AI pricing directly contributes to this problem.
Switching Cost Reality
If you start with per-resolution and later want to switch, you face migration costs. Your conversation history, knowledge base setup, and widget customization need to move. The longer you wait, the more locked in you become.
Starting with flat-rate pricing avoids this trap entirely. You build on a pricing model that stays predictable whether you have 500 or 50,000 AI resolutions per month.
Making the Decision
Ask yourself these questions:
- How many AI resolutions do you expect per month? If under 100, per-resolution is fine. Over 100, run the math.
- Will your AI improve over time? If yes (and it should), your resolution count will grow. Your bill will grow with it under per-resolution pricing.
- Do you need cost predictability? If you report to a CFO or manage a budget, flat-rate pricing eliminates surprise bills.
- What is your total support budget? Compare the all-in cost: platform + AI + knowledge base + team seats. Not just the AI fee.
Frequently Asked Questions
What counts as a "resolution" in per-resolution pricing?
A resolution is typically logged when the AI handles a conversation from start to finish without escalating to a human agent. The exact definition varies by platform. Some count a resolution even if the customer was not satisfied, as long as they did not request a human.
Can I set a spending cap on per-resolution AI?
Some platforms allow limits. Intercom lets you set a monthly budget for Fin. Once the budget is reached, AI stops responding and all conversations go to human agents. This limits costs but also limits AI availability.
Is flat-rate AI pricing really unlimited?
Not unlimited, but generous. Flat-rate plans typically include a set number of AI answers per tier. For example, 2,500 on a starter plan, 10,000 on a mid-tier, and 40,000 on a high-volume plan. If you exceed the limit, you upgrade to the next tier. There are no surprise overage charges.
How do I estimate my monthly AI resolutions?
Look at your current support ticket volume. Estimate that AI can resolve 60-70% of repetitive questions. If you get 1,000 tickets per month and 700 are repetitive, expect roughly 400-500 AI resolutions initially, growing as your knowledge base improves.
Does cheaper AI mean worse AI?
No. The pricing model does not determine AI quality. The quality depends on the underlying language model, the RAG implementation, and your knowledge base content. A well-maintained knowledge base with a flat-rate AI tool produces better answers than a neglected knowledge base with an expensive per-resolution tool.