← Back to AURELIQO

AI Customer Support Agent Pricing in 2026: Per-Resolution vs. Per-Seat Costs Compared

Compare AI customer support pricing by resolution, seat, session, conversation, and credit—and calculate realistic total cost before switching.

AI Customer Support Agent Pricing in 2026: Per-Resolution vs. Per-Seat Costs Compared

Support leaders face a deceptively difficult buying decision: should an AI customer support agent be priced like a software seat, or should it be tied to the number of customer issues it resolves?

Per-seat pricing is familiar and predictable, but it can make teams pay for capacity they do not use. Outcome-based pricing can align spend with automation, yet the invoice rises as the AI resolves more customer requests. Hybrid models add another layer by combining seats, platform fees, tickets, credits, and usage charges.

The right comparison is not the lowest advertised AI rate. It is the model that produces the lowest cost per successfully resolved case while preserving service quality and keeping seasonal invoices manageable.

This guide compares the main pricing structures used by Intercom Fin, Zendesk AI Agents, Gorgias, Freshdesk Freddy AI Agent, Salesforce Agentforce, and HubSpot Breeze Customer Agent. Prices and packaging can change quickly, so treat the figures below as a research baseline and confirm commercial terms with each vendor before purchasing.

The short answer: compare total cost, not the AI unit price

A useful cost model is:

```text Total monthly cost = base platform fees + human-agent seats + AI usage + ticket or conversation fees + implementation and optimization costs + overages or minimum commitments ```

Then calculate:

```text Cost per successfully resolved case = total monthly cost ÷ successfully resolved cases ```

This matters because vendors use different billing units. Industry research identifies five broad approaches: flat usage included in a subscription, per-resolution, per-conversation, hybrid seat-plus-meter, and expiring-credit pricing. It also warns that “resolution” is not standardized across vendors (Drag).

Outcome-based pricing can improve the connection between spend and business results, but it can also create disputes over attribution when several tools or human inputs contribute to the outcome (L.E.K. Consulting).

Comparison of AI customer support pricing models

| Platform | Primary AI billing unit | Published pricing reference | Other cost layers to investigate | Best comparison question | |---|---|---:|---|---| | Intercom Fin | Outcome or resolution | $0.99 per outcome; Intercom seats start at $29 per seat/month on Essential when billed annually | Separate subscription, included allowance, minimums, and non-resolution outcomes | What percentage of billable outcomes are true customer resolutions? | | Zendesk AI Agents | Automated resolution | Suite Team $55 and Professional $115 per agent/month when billed annually | AI usage terms, seat count, plan requirements, and overages | How many automated resolutions will occur under the contracted tier? | | Gorgias AI Agent | Resolved interaction | $0.90 annual-billing rate or $1.00 monthly-billing rate in listed plans | Separate helpdesk ticket charges | Can one interaction create both a ticket fee and an AI fee? | | Freshdesk Freddy AI Agent | 24-hour session | $49 per 100 sessions for a first-time session add-on | Session packs and shared chatbot usage | How many sessions occur per customer issue? | | Salesforce Agentforce | Flex credits, conversation, or user license | $500 per 100,000 Flex Credits; $2 per conversation listed | Salesforce edition, actions, integrations, and license model | How many actions does a complete resolution consume? | | HubSpot Breeze Customer Agent | Credits | 3,000 monthly credits for Pro and 5,000 for Enterprise; extra packages start at $10 per 1,000 credits | HubSpot subscription, credit allowances, and additional capacity | What does one resolved case consume in credits? |

The table is a starting point, not a like-for-like ranking. A rate per resolution cannot be directly compared with a rate per session or credit until you know the average number of billable units required for one successfully resolved case.

Per-resolution pricing: Intercom Fin and Gorgias

Intercom Fin

Intercom lists Fin at $0.99 per outcome, while Intercom seats start at $29 per seat per month on the Essential plan when billed annually (Intercom pricing). Fin also has a separate monthly subscription and an allowance of 50 included monthly resolutions; unused included resolutions do not roll over, and minimum commitments may apply (Intercom’s Fin resolution documentation).

The key commercial detail is that not every billable event should be assumed to be a final support resolution. Intercom defines a resolution as a conversation in which the customer confirms the answer was satisfactory or leaves without requesting more help. Its outcome documentation also identifies procedure handoffs and sales qualification outcomes as separately priced outcomes (Intercom Fin outcomes).

For forecasting, separate these categories:

Resolved support conversations.
Procedure or workflow handoffs.
Qualification or other paid outcomes.
Conversations that escalate to a human.
Reopened or repeat contacts.

Gorgias AI Agent

Gorgias prices its AI Agent per resolved interaction rather than per seat or message. Its published information lists $0.90 per resolved interaction on annual billing and $1.00 on monthly billing, with Starter plans beginning at $1.00 (Gorgias official information).

Gorgias separately charges for helpdesk ticket usage. As a result, an AI-resolved interaction can produce both a ticket charge and an AI Agent charge (Gorgias billing documentation). Gorgias also states that when a shopper contacts a human support agent within 72 hours of an automated resolution, the interaction is charged only as a billable helpdesk ticket rather than receiving a separate automated-interaction charge (Gorgias pricing).

That rule makes handoff and recontact behavior particularly important. A vendor’s headline automation rate may look attractive until you model ticket fees, human follow-up, and repeat contacts together.

Per-seat and automated-resolution pricing: Zendesk

Zendesk measures AI-agent usage through automated resolutions. Its definition is a customer request successfully resolved by the AI agent without escalation to a human agent (Zendesk automated resolution tiers). Zendesk also says each automated resolution is independently verified by a dedicated AI evaluation model in addition to being completed end to end by the AI agent (Zendesk newsroom).

The customer-service pricing page lists Suite Team at $55 per agent per month and Suite Professional at $115 per agent per month, billed annually. Suite Team includes AI Agents, AI Knowledge Base, and AI Action Builder (Zendesk pricing).

Zendesk therefore belongs in the hybrid category for many buyers: the base cost is tied to agent seats, while AI activity is measured separately through automated resolutions. Your forecast should identify:

The number of human-agent seats required.
Which suite is required for your workflows.
The expected number of automated resolutions.
Any AI usage allowance, tier, or overage term in the quote.
Whether agent-assist capabilities have separate licensing requirements.

Do not compare Zendesk’s seat price directly with a $0.99 resolution fee. The two figures pay for different things.

Session, action, and credit pricing

Freshdesk Freddy AI Agent

Freshworks defines a Freddy AI Agent session as all interactions between an end user and the AI agent or chatbot within a 24-hour period. Its official FAQ lists a first-time session add-on at $49 per 100 sessions (Freshworks session FAQ). Session packs are shared between Freddy AI Agents and configured chatbots in the Freshworks account.

A session model can be economical when several messages and steps fit inside one 24-hour window. It can be less predictable when customers open multiple sessions for one unresolved issue. Forecast sessions from historical contact patterns instead of assuming one session equals one resolution.

Salesforce Agentforce

Salesforce supports Flex Credit, per-conversation, and per-user licensing models. Its pricing page lists Flex Credits at $500 per 100,000 credits. One standard Agentforce action consumes 20 credits, equivalent to $0.10 per action, while conversation pricing is listed at $2 per conversation (Salesforce Agentforce pricing).

The difference between an action and a conversation is material. A single customer request may require several actions across CRM, order management, identity, billing, or another system. Ask for an action-level example of your most common workflows before using the conversation price in a business case.

HubSpot Breeze Customer Agent

HubSpot uses credits for Breeze Customer Agent rather than a simple per-resolution fee. Pro subscriptions receive 3,000 monthly credits, Enterprise subscriptions receive 5,000, and additional capacity packages start at $10 per 1,000 credits (HubSpot’s credit announcement).

Credit pricing requires a translation layer: credits per intent, credits per tool call, credits per conversation, and credits per successful resolution. Request that mapping in writing and model unused, expired, or additional credits according to the contract.

A practical calculator framework

Build your comparison spreadsheet with one row per support intent—for example, order status, password reset, refund request, billing question, and technical troubleshooting.

For each intent, record:

Monthly inbound cases.
Expected AI containment rate.
Human escalation rate.
Recontact or reopen rate.
Average billable units per case.
Unit price and included allowance.
Human handling time for escalated cases.
Required seats and platform fees.

Run at least three scenarios:

Low-volume scenario

Use a quieter month to test whether minimum commitments, base subscriptions, and unused allowances dominate total cost.

Expected-volume scenario

Use a realistic monthly average by intent. Do not apply one blended automation rate to every ticket; order-status questions and complex technical cases may behave very differently.

Peak-season scenario

Add seasonal demand, retries, multilingual traffic, reopened cases, and multi-step workflows. This is where automatic upgrades, overage rates, and expiring credits can materially change the result.

As an illustrative calculation—not a vendor forecast—1,000 billable Intercom outcomes at $0.99 would produce $990 in usage before subscription and seat charges. The same 1,000 Gorgias resolved interactions at the listed $0.90 annual-billing rate would produce $900 in AI charges before helpdesk ticket fees. These figures are not equivalent until you confirm that each vendor bills the same type and quality of customer outcome.

Questions to ask before signing

Use these questions in procurement:

What exactly triggers a billable resolution, interaction, session, conversation, action, or credit?
Is a customer who abandons, reopens, or recontacts a case still billable?
Does a human handoff create a second charge?
Do included units roll over, expire, or disappear at the end of the billing period?
Are there minimum commitments, caps, automatic upgrades, or overage rates?
Are workflow handoffs, sales qualification, and tool actions priced separately?
Which integrations and knowledge-base capabilities are included?
What implementation, content cleanup, monitoring, and optimization work is required?
How will quality be measured—containment, recontact rate, escalation rate, CSAT, first-contact resolution, and cost per successful case?
Can the vendor provide an intent-level usage simulation using your historical tickets?

Bottom line

Per-seat pricing is easiest to budget when support volume is stable and the platform already fits your operation. Per-resolution pricing can be compelling when the vendor’s definition closely matches the outcome you value. Session, conversation, action, and credit models can work well, but only when you understand how many billable units a complete customer issue consumes.

The winning platform is not necessarily the one with the cheapest AI rate. Choose the model that makes billing observable, maps cleanly to successful support outcomes, handles peak demand without surprises, and reduces total cost after human labor, platform fees, integrations, and ongoing optimization are included.

Sources

Some links may become affiliate links. If you make a qualifying purchase through one of those links, this site may earn a commission at no additional cost to you.