Per-resolution pricing sounds reasonable on paper: you only pay when the AI actually solves something. No resolution, no charge. But the model has a perverse incentive baked in. The better your chatbot performs, the more you pay.
Teams that hit 60% deflection after a successful rollout often see their AI bill jump 30–50% in the same quarter finance was expecting savings. The trap is not a bug. It is how the model is designed.
38%
Avg. bill increase after rollout
$0.99–$2.50
Typical per-resolution fee
3,000+
Resolutions at 5k conv/mo
High
Annual overspend risk
How per-resolution pricing actually works
Vendors charge a flat fee each time the AI handles a conversation without escalating to a human agent. Intercom's Fin AI agent, for example, bills $0.99 per resolution on top of your base plan. At 3,000 AI-resolved conversations per month, that is nearly $3,000 in resolution fees alone.
| Monthly AI resolutions | At $0.99/resolution | At $1.50/resolution | At $2.50/resolution |
|---|---|---|---|
| 1,000 | $990 | $1,500 | $2,500 |
| 3,000 | $2,970 | $4,500 | $7,500 |
| 5,000 | $4,950 | $7,500 | $12,500 |
| 10,000 | $9,900 | $15,000 | $25,000 |
The success paradox
Improve your bot from 40% to 70% deflection and your resolution count nearly doubles. Your team saves headcount. Your vendor invoice grows in lockstep.
What counts as a resolution?
This is where the trap tightens. Every vendor defines resolution differently, and the definition often expands over time.
- Bot answers a FAQ and visitor leaves: resolution
- Bot collects a lead and hands off to sales: resolution
- Bot deflects to a help article link: resolution
- Bot handles 80% of the issue, human closes the last step: sometimes still a resolution
- Visitor returns within 24 hours on the same topic: may count as a new resolution
Switch to predictable AI pricing
Book a demo and we will model your cost at scale, with no per-resolution surprises.
Three signs you are in the trap
- Your AI bill grows every month even though traffic is flat
- Finance asks why chatbot savings are not showing up in the P&L
- You hesitate to improve the bot because each improvement costs more
- You cannot forecast next quarter's spend without guessing resolution rates
How flat-rate pricing changes the math
Flat platform pricing flips the incentive. When your bot deflects more conversations, your cost stays the same (or scales predictably with tier upgrades, not per-event charges). The ROI from better AI performance flows to your business, not your vendor.
NIVA uses flat platform pricing with usage-based tier upgrades. There are no per-resolution fees. When your real estate bot books 200 extra viewings or your insurance bot handles 500 more quote intakes, your bill does not spike per event.
Bottom line
Per-resolution pricing is a tax on success. Before you sign, model your bill at 70% deflection, not 30%. If the number scares you, that is the point. Choose a pricing model where better performance means better ROI for you, not higher revenue for your vendor.
Frequently asked questions
What is per-resolution pricing for AI chatbots?
Per-resolution pricing charges a fee each time the AI resolves a conversation without human escalation. Typical rates range from $0.99 to $2.50 per resolution, billed on top of the base platform fee.
Why does per-resolution pricing increase as my bot improves?
Higher deflection rates mean more conversations handled by AI, which means more billable resolutions. The pricing model scales with your success, not your subscription tier.
Does NIVA charge per resolution?
No. NIVA uses flat platform pricing. AI-handled conversations are included in your plan tier without per-resolution surcharges.
How do I estimate my real per-resolution cost?
Multiply your monthly conversation volume by your target deflection rate, then multiply by the per-resolution fee. Example: 5,000 conversations × 60% deflection × $0.99 = $2,970/month in resolution fees alone.
Switch to predictable AI pricing
Book a demo and we will model your cost at scale, with no per-resolution surprises.