Resolution-Based AI Pricing Explained: The Hidden Cost of Per-Resolution Billing

Quick Answer:

Resolution-based AI pricing charges a fee every time a chatbot marks a conversation “resolved,” instead of billing a flat rate. Intercom’s Fin AI bills $0.99 per resolution, and an “assumed resolution,” where the customer simply stops replying, counts toward that total. Ollo Agent uses flat-rate plans from $49 to $299/month instead, with no per-resolution or per-seat charges.

A support team runs 1,000 conversations through Intercom’s Fin AI this month. The invoice adds $990 on top of seat fees, billed for conversations Fin marked “resolved.” Some of those are real: the customer got an answer and left satisfied. Others are what Intercom calls an “assumed resolution”: the customer stopped replying within 24 hours, and the system counted it as a win anyway.

That’s resolution-based pricing, the billing model now used in some form by Intercom, Freshworks, and Tidio. It sounds fair in a sales deck: pay only when the AI actually helps. In practice, the vendor decides what “helped” means, and that definition is where the bill grows.

We looked at how three major platforms define and price a resolution, ran the real-world math on each, and compared the result against a flat-rate plan. If you’re evaluating an AI chatbot for a service business right now, the difference between these two billing models can change your effective monthly cost by hundreds of dollars without a single line item looking dishonest on its own.

What Is Resolution-Based AI Pricing?

Resolution-based pricing charges a fee each time an AI chatbot marks a customer conversation as resolved, on top of a base subscription or per-seat fee. You pay for the platform itself, then pay again for every conversation the AI closes out.

Freshworks uses a close variant it calls “per session” pricing: it charges for every AI interaction, whether or not that interaction actually solves the customer’s problem. Tidio bills in blocks of AI conversations rather than individual resolutions, which shifts the risk in a different direction, since you buy a fixed batch upfront and hope enough of it resolves.

All three sit in contrast to flat-rate platforms, where the monthly price stays the same no matter how many conversations the AI handles or how many it resolves. The buyer’s risk moves in opposite directions depending on which model you pick: with usage-based pricing, a good month for your business (more inquiries) is also a more expensive month for your AI bill.

How Do Vendors Decide What Counts as a “Resolution”?

Each vendor sets its own definition of “resolved,” and two of the three define it in ways that favor the bill over the customer’s actual outcome. That distinction rarely appears on the pricing page itself.

Intercom’s Fin AI counts a conversation as resolved when the customer confirms the answer helped, or when the customer stops responding for 24 hours after Fin’s last message. The second case is the “assumed resolution”: a customer who gave up, got distracted, or left to call someone instead still generates a $0.99 charge. Seat fees run $29 to $132 per seat per month across Intercom’s Essential, Advanced, and Expert plans, before any Fin AI usage is added.

Freshworks’ Freddy AI Agent avoids the resolution question entirely by billing per session instead: every AI interaction costs money, solved or not. Freshdesk Classic’s email agent runs about $0.49 per session; Freshdesk Omni’s chat agent runs closer to $0.10 per session, on top of $19 to $119 per agent per month depending on plan and product line.

Tidio’s Lyro AI add-on starts at $39/month for 50 AI conversations, scaling to $139/month for 200 and $289/month for 500, layered on top of Tidio’s own $0 to $749/month base plans. Tidio’s top-tier Tidio+ plan includes a “Resolution Rate Guarantee,” an SLA that commits to only a 50% Lyro resolution rate. In other words, the vendor’s own contract assumes half the conversations you pay for won’t be resolved.

What Does Per-Resolution Billing Actually Cost at Scale?

The headline per-unit price is rarely what you pay per solved ticket, once you factor in how many interactions it actually takes to reach one real resolution.

*Freshworks publishes a 23–30% resolution rate for Freddy AI Agent. At a 25% rate, it takes roughly four AI sessions to close one ticket, which turns a $0.10 headline chat-session price into close to $0.40 per resolved conversation, and a $0.49 email-session price into close to $1.96 (analysis: eesel AI, 2026). Pricing as of August 2026; check each platform’s official page for current plans: Intercom pricing, Freshworks pricing, Tidio pricing.

Why Is the AI Chatbot Market Moving Toward Resolution-Based Pricing?

Usage-based billing lets a vendor’s revenue grow automatically as a customer’s chat volume grows, without renegotiating the contract. That’s the commercial logic behind the shift, and it also creates a built-in incentive problem.

A platform that gets paid more every time it marks a conversation “resolved” has a direct financial reason to define “resolved” loosely. Intercom’s assumed-resolution rule and Freshworks’ pay-per-attempt session model are two different ways that incentive shows up in a live pricing page today, not a hypothetical risk.

Buyers don’t see this at the moment they sign a contract, because a per-unit price like “$0.99 per resolution” reads as precise and fair. The gap only becomes visible on the third or fourth invoice, once the actual resolution rate and the assumed-resolution count start showing up as separate line items, if the vendor breaks them out at all.

Ollo Agent’s take: We expect the gap between flat-rate and resolution-based pricing to become the main battleground in this category over the next year, more than any feature list. Freshworks’ own published 23–30% resolution rate next to a roughly 4x real-cost multiplier is an early, public example of what that gap costs a business in practice, and we think more buyers will start asking for the real number before signing.

How Can You Estimate Your Real Monthly Cost Before You Sign Up?

Ask for the numbers below before comparing any two vendors, since the sticker price on a pricing page almost never matches the invoice three months in.

1. Get the vendor’s published resolution rate, not a marketing claim. Freshworks publishes 23–30%; ask any vendor that doesn’t publish a number why not.
2. Divide your expected monthly conversation volume by that resolution rate to find your real session or resolution count. At a 25% rate, 1,000 conversations means roughly 4,000 billable sessions.
3. Multiply that count by the per-unit price listed on the vendor’s current pricing page.
4. Add the base seat or platform fee for your team size, since usage charges sit on top of it, not instead of it.
5. Compare the total to a flat-rate plan at the same expected volume before signing anything.

Flat-Rate vs. Resolution-Based Pricing: Which Fits a Growing Service Business?

A flat-rate plan removes the volume math entirely: the price you see is the price you pay, whether the AI handles 300 conversations or 3,000. That predictability matters most for a small team that can’t absorb a surprise four-figure line item mid-quarter.

Ollo Agent’s plans work this way across text chat, website widgets, and messenger channels, with one unified inbox instead of a separate add-on meter for each channel. Voice is the one area where per-usage vendors often charge extra regardless of billing model: Intercom’s voice option is a paid add-on, while Ollo bundles a native voice agent into the Enterprise plan at no additional per-call charge.

Resolution-based pricing can still make sense for a large team with unpredictable, low-volume spikes and the staff to audit vendor billing monthly. For most SMBs handling a steady 300 to 3,000 conversations a month, a flat rate is easier to budget and harder to be surprised by.

The setup timeline is worth weighing alongside the billing model, since a vendor with a complex per-seat-plus-usage structure often has a longer contract and onboarding process to match. Ollo Agent’s plans go live within 24 hours of signup, which matters most for a business trying to fix a lead-loss problem this month, not next quarter.

Real Results: What Flat-Rate Automation Delivered for One Travel Agency in 30 Days

Flat pricing doesn’t mean a lower resolution rate. ALF, a travel agency running Ollo Agent across four messenger channels, resolved 65% of initial inquiries without human involvement, grew bookings +27% on the same traffic, and cut manager workload −35%, all on an unmetered plan.

Source: Ollo Agent internal data, ALF client, 2025–2026.

Start a free trial and see your real resolution rate in 24 hours → No developer setup, no per-resolution meter running in the background.

FAQ

What is resolution-based AI pricing?

It's a billing model where an AI chatbot vendor charges a fee each time it marks a customer conversation as "resolved," on top of a base subscription. Intercom's Fin AI is the clearest example, at $0.99 per resolution.

What counts as a "resolution" in AI chatbot pricing?

It depends on the vendor. Intercom counts it when a customer confirms an answer helped, or when the customer stops replying for 24 hours, an "assumed resolution" that still gets billed.

Does Ollo Agent charge per resolution?

No. Ollo Agent uses flat monthly plans from $49 to $299, with no per-resolution fees and no per-seat multiplier, regardless of conversation volume.

Why does Freshworks charge per session instead of per resolution?

Per-session billing charges for every AI interaction, solved or not, which avoids defining "resolved" at all. At Freshworks' own published 23–30% resolution rate, that can cost roughly four times the headline session price per actual resolution.

What is an "assumed resolution" in Intercom's Fin AI pricing?

It's a conversation counted as resolved when the customer simply stops responding for 24 hours after Fin's last message, even without confirming the answer helped. It still generates a $0.99 charge.

How do I calculate the real cost of a resolution-based AI plan?

Divide your expected monthly conversations by the vendor's published resolution rate to get your real billable session count, multiply by the per-unit price, then add the base seat fee.

Is flat-rate AI pricing more expensive than per-resolution pricing?

Not usually at typical SMB volume. A flat plan can cost less than a metered one once assumed resolutions or a low published resolution rate are factored into the real per-unit cost.

Can per-resolution pricing include conversations that weren't actually solved?

Yes. Intercom's assumed-resolution rule and Tidio's Tidio+ plan, which guarantees only a 50% Lyro resolution rate under its own SLA, both mean part of what you pay for isn't a solved customer problem.

Share this article