AI support pricing: what the four models actually cost

Q: What are the four AI support pricing models?

Per seat bills each licensed agent. Per conversation bills every bot attempt, including failures. Per resolution bills only when the AI closes the ticket without a human. Flat platform fee is a fixed monthly or annual commitment. Each model behaves differently as your automation rate climbs, which is why unit rates alone are a weak comparison.

Q: Is per-resolution pricing cheaper than per-seat?

Not automatically. Per-resolution aligns cost with outcomes, but the invoice rises as the AI succeeds more often. Per-seat looks stable until automation lets you stop hiring, at which point seat cost stops falling while ticket volume keeps rising. Model both against your volume curve and your automation target.

Q: Where can I see IrisAgent plan cards?

See live plan cards on the IrisAgent pricing page at /pricing/. Free, Standard from $500/month, and Enterprise custom are the public bands. Book a demo to match usage-based or resolution-based structure to your spike pattern.

Updated September 2026

Compare per-seat, per-conversation, per-resolution, and flat platform pricing across AI support vendors. See how IrisAgent overlays the helpdesk you already run. Plan cards stay on /pricing/.

Book a Demo

IrisAgent AI support pricing and automation overlay

The four AI support pricing models

Buyers searching for AI chatbot pricing usually get a sticker. The packaging decides the bill. Here are the four models in plain language.

1. Per seat

You pay a monthly fee for each licensed agent, often with an AI or copilot add-on on top. Inherited from helpdesk licensing. Best when the primary use case is agent assist and headcount is stable. Weak when you buy automation to absorb growth without hiring, because the seat bill does not fall when deflection rises.

2. Per conversation

You pay each time the bot opens a conversation, whether or not it solves anything. Weakest incentive alignment in the category: failures cost the same as resolutions. Watch how a conversation is defined, including 24-hour inactivity resets that rebill the same issue.

3. Per resolution

You pay only when the AI closes a ticket end to end without a human. Publicly reported rates often sit between about $0.50 and $6.00. Best aligned with outcomes on paper, most variable in practice. Insist on a written resolution definition and a reopen window.

4. Flat platform fee

A fixed monthly or annual commitment, sometimes with an implementation fee. Predictable, and high automation does not raise the bill inside the tier. The catch is the floor: quiet months cost the same as peak ones, and smaller teams can overpay.

Worked cost bands at 2,000 / 10,000 / 50,000 tickets per month

Illustrative only, as of September 2026. Assumes roughly 50% of tickets are AI-resolved where a per-resolution rate applies. Competitor unit rates are publicly reported or as cited in IrisAgent comparison and blog pages. IrisAgent shows public plan bands only (no invented Iris $/resolution list price). Seat and add-on costs are excluded unless noted.

Vendor / model	2,000 tickets	10,000 tickets	50,000 tickets	Source note
IrisAgent Flexible usage or resolution; Standard from $500/mo	From $500/mo Standard band	From $500/mo; structure scales with plan	Enterprise custom common at this volume	Live /pricing/: Free $0, Standard from $500/mo, Enterprise custom
Fin (Intercom) ~$0.99 / resolution + seats	~$990 resolution fees (1,000 AI res.)	~$4,950 resolution fees (5,000 AI res.)	~$24,750 resolution fees (25,000 AI res.)	Publicly reported Fin rate; seats $29 to $132 and Copilot add-on extra. See IrisAgent vs Intercom.
Ada ~$3.50 / resolution (illustrative)	~$3,500 (1,000 AI res.)	~$17,500 (5,000 AI res.)	~$87,500 (25,000 AI res.)	Illustrative band used in IrisAgent blogs from publicly reported figures; Ada is often quote-only. See IrisAgent vs Ada.
Zendesk AI Suite-tied	Quote from Zendesk	Quote from Zendesk	Quote from Zendesk	Typically seats + Advanced AI add-ons + automated resolution metering. Dollar rates: unknown / confirm with Zendesk. See IrisAgent vs Zendesk AI.
Gorgias Ecommerce-oriented	Ticket-tier / plan pricing	Ticket-tier / plan pricing	Higher tiers / custom	Ecommerce helpdesk packaging; confirm current public plan tiers on Gorgias. Not a general SaaS support overlay.

Choose a model if…

Choose per seat if agent assist is the main job and headcount is stable.
Choose per conversation if failure rates are genuinely low and scope is narrow FAQ deflection (rare for complex support).
Choose per resolution if you want cost to track outcomes and can tolerate a variable invoice as automation improves.
Choose flat platform if volume is high enough to use the floor and finance needs a fixed number.
Choose IrisAgent flexible usage or resolution ifyou want a helpdesk overlay with public Standard from $500/month and a structure you can match to volume. See plan cards.

/pricing/(live Free / Standard / Enterprise plan cards)
IrisAgent vs Intercom (Fin)
IrisAgent vs Ada
IrisAgent vs Zendesk AI
Blog: AI chatbot pricing cost guide (2026)(measure article; this page is the living packaging reference)
Blog: conversational AI platform pricing

AI support pricing questions?

We have answers.

How much does AI customer support cost in 2026?

It depends on the pricing model more than the sticker. Per-seat plans often run about $29 to $150 per agent plus AI add-ons. Per-conversation fees commonly sit around $0.10 to $1.50 per attempt. Per-resolution rates that are publicly reported often fall between about $0.50 and $6.00. Flat platform fees typically start in the mid-market thousands per month. IrisAgent publishes Free at $0/month, Standard from $500/month, and Enterprise custom, with flexible usage-based or resolution-based structures. Compare models against your ticket volume rather than a single vendor list price.

What are the four AI support pricing models?

How does IrisAgent pricing compare to Fin, Ada, and Zendesk AI?

IrisAgent publishes Free at $0/month, Standard from $500/month, and Enterprise custom, with flexible usage-based or resolution-based plans. Fin by Intercom is publicly reported around $0.99 per resolution on top of seats (as of dates cited on our comparison pages). Ada is often modeled around about $3.50 per resolution in IrisAgent blogs drawing on publicly reported figures, and is otherwise quote-only. Zendesk AI is typically suite-tied and quote-based on seats, Advanced AI add-ons, and automated resolutions. Confirm competitor rates on vendor sites before you buy.

Is per-resolution pricing cheaper than per-seat?

Where can I see IrisAgent plan cards?

Book a demo or see plan cards

Match usage-based or resolution-based structure to your volume. Public bands stay on /pricing/.