Pricing models

AI Customer Support Pricing Explained

Published September 10, 2026 · Reviewed by the Fidiora team

Short answer

AI support is billed on several units: per resolution, per conversation, per message, per seat add-on, or a quoted enterprise agreement. The unit matters more than the rate, because a cheap price on an easily inflated unit is not a discount.

Key takeaways

  • Per-message billing is the easiest unit to inflate and the least aligned to value.
  • Quoted enterprise agreements hide implementation cost that rarely appears upfront.
  • A spend cap is what makes any usage model safe to sign.
  • Ask for the resolution definition in writing before comparing rates.

The billing units

Per resolution charges when an issue is solved. Per conversation charges on contact regardless of outcome. Per message charges by exchange, which a verbose model or a chatty customer inflates without adding value. Per-seat add-ons charge for capability rather than usage. Quoted agreements vary and typically hide implementation cost.

Why the unit matters more than the rate

The same conversation might be thirty billable messages or one billable resolution. A headline rate that looks thirty times cheaper is identical in practice, or worse. Always model your own volume against the specific unit before comparing any two prices.

The definition question

Where a vendor prices on resolutions, they wrote the definition. Ask whether abandonment counts, whether handoffs are billed, whether a reopened conversation retracts the charge, and whether you may audit a sample of billed resolutions. Get all four in the contract.

The implementation line

Enterprise AI platforms carry real implementation and ownership cost that rarely appears in the quote. Ask who does the work, how long it takes, and what internal role is required to run it afterwards. That answer often changes the comparison entirely.

The five questions

What is the billable unit and exactly how is it defined. Is abandonment billed. Are handoffs billed. Is there a hard spend cap. Can I audit billed usage. A vendor unwilling to answer all five in writing is telling you something useful.

Side by side

AI support billing units and what each one rewards.
Criterion Unit What it rewards
Per resolution Charged when an issue is genuinely solved Successful outcomes, if the definition excludes abandonment and handoffs
Per conversation Charged when a conversation happens Volume of contact, regardless of whether it helped
Per message Charged per exchange Verbosity, which is the least aligned unit available
Per-seat add-on Charged per agent licence Headcount, with no relationship to AI usage or value
Quoted enterprise Negotiated, often outcome linked Whatever the negotiated definition rewards, so read it closely
How Fidiora prices

Fidiora charges $0.59 per genuine resolution. No seat fees, no add-on modules, no channel surcharges, and nothing billed for abandoned chats, timeouts, or handoffs to a human. Set a monthly spend cap and the bill cannot exceed it.

FAQ

Questions

How much does AI customer support cost?
It depends entirely on the billing unit and your volume. Fidiora publishes a rate of $0.59 per genuine resolution with no seat fee. Other vendors bill per conversation, per message, or through quoted agreements, so convert everything to cost per resolved issue before comparing.
What is the safest AI support billing unit?
Per genuine resolution with a published definition and a hard spend cap, because it is the unit closest to customer value and the hardest to inflate. Per message is the least aligned.
Should I worry about unpredictable AI costs?
Only without a cap. Ask what mechanism stops the bill rising, not whether it can. A hard ceiling plus proactive alerts turns a variable bill into a bounded one.
Get Started

See Fidiora resolve a ticket in 60 seconds.

No credit card, no sales call required. Connect your docs and watch it work.

No credit card · Live in under an hour