Gorgias Automate vs an AI Agent Layer: The Real Cost
Gorgias bills its AI the way many ecommerce teams say they want to be billed: you only pay when the AI fully resolves a conversation on its own. It sounds clean, and for a store drowning in "where is my order?" tickets it can genuinely lower the cost of the easy stuff. But resolution-based billing hides a second question that matters more as you automate deeper work like returns, exchanges, and subscription changes: what happens to your bill when the AI does five useful things but doesn't cleanly "resolve," and what happens when a resolved ticket gets charged twice? This is an honest, first-party cost model — where native Gorgias Automate earns its keep, where a dedicated AI agent layer that runs on top of Gorgias has the edge, and how the two pricing philosophies (per resolution versus per action) actually behave on a real ecommerce queue.
How Gorgias Automate is priced
Gorgias splits your bill into two meters. The first is the helpdesk ticket meter you already know: per Gorgias' billing documentation, a ticket becomes billable when at least one message in it is sent from your Gorgias helpdesk — whether that message came from a human agent, from AI Agent, or from an automated Rule.
The second meter is the AI one, and it is outcome-based. Gorgias charges an additional automation fee only when AI Agent resolves a ticket without handing the conversation over to a human. Fully autonomous resolutions — through AI Agent, Flows, Order Management, or Article Recommendations — trigger it; a ticket the AI hands off to a person does not. Gorgias' own AI Agent pricing page defines the billable event plainly: "the AI resolves a customer conversation entirely on its own. The customer asks something, the AI handles it, the conversation closes. That's one interaction."
That is a real advantage worth crediting. You don't pay for the AI's failed attempts, and emails with no replies, spam, and human-resolved tickets don't count. For a store where 40% of contacts are order-status questions the AI can close end to end, you pay for outcomes and nothing else.
The part the "only pay for resolutions" pitch skips
Two mechanics change the arithmetic once you look closely.
Dual billing on the same ticket. When AI Agent responds to and completely resolves a ticket on its own, Gorgias' docs confirm you pay both a ticket fee and an automation fee on that one ticket. A fully automated WISMO reply isn't "free automation" — it's a helpdesk ticket plus a resolution charge. (If the AI hands off to a human, you pay only the ticket fee, which is the more forgiving case.)
Overages above your included resolutions. AI Agent plans bundle a set number of automated interactions, and once you exceed them, overages are billed at $1.50 per interaction — higher than the discounted base rate you signed up at. A seasonal spike (BFCM, a viral product) is exactly when your resolution count runs hottest and your per-unit price steps up.
Here is the current shape of the two meters, with the AI Agent add-on shown alongside the helpdesk tiers.
None of this makes Gorgias' model bad — it makes it outcome-shaped. It rewards clean, one-and-done automations and quietly penalises the messier, multi-step work that defines returns and subscriptions, where the AI often does a lot of valuable work without a tidy autonomous "resolution." For the full plan-by-plan breakdown, our Gorgias pricing explained guide walks the tiers, and Gorgias Automate & AI complete guide covers what each automation surface actually does.
Per-resolution vs per-action: what the meters reward
The cleanest way to see the difference is to stop thinking about "AI cost" as one number and look at what each model charges for. Gorgias charges per resolution (an outcome). An AI agent layer like Macha charges per action (a unit of work the agent performs — a reply drafted, a tool call made, a lookup executed). The Macha pricing page details the credit model; the screenshot below is the real billing screen from our demo org.
Neither model is universally cheaper — they behave differently by ticket type:
| Ecommerce scenario | Gorgias Automate (per resolution) | AI agent layer (per action) |
|---|---|---|
| WISMO — one lookup, one reply, closed | One clean automated interaction billed (+ a ticket fee) | A few actions (order lookup + reply) drawn from the credit pool |
| Return/exchange — check policy, verify order, issue label, update tags | If fully autonomous, one resolution; if it hands off, no automation fee despite real AI work | Every action the agent took is metered whether or not it "resolved" |
| Subscription change — pause, swap SKU, adjust date via API | Same all-or-nothing outcome billing | Metered by the API calls and steps performed |
| Peak season overflow | Overages at $1.50/interaction above your bundle | Actions draw from your monthly credit pool at one rate |
| AI does work but hands off | Not billed for the AI portion (good) | Billed for the actions it took (you pay for the work) |
The honest read: if most of your volume is simple, self-contained deflections, per-resolution billing is elegant and predictable. As soon as you automate deeper, multi-tool workflows — the kind where the agent reasons across your store's order, returns, and subscription systems — per-action billing tells you what the automation actually cost, and it doesn't double-charge a resolution as both a ticket and an interaction. Credits are consumed per AI action, never "per resolution."
Deflection vs resolution vs automation
The pricing debate usually collapses three different things into one word. Separating them is the point.
Deflection is preventing a ticket from ever reaching an agent — a help-center article or a self-serve flow that answers before a human is needed. Resolution is the AI closing a conversation autonomously, which is the event Gorgias meters. Automation is broader than both: it's any AI action that moves a ticket forward — triaging by intent, pulling an order via a custom tool, drafting a grounded reply for an agent to approve, updating a tag, or handing off cleanly with full context.
Gorgias' billing optimises for the middle one, resolution, because that's the outcome it can charge for. But a lot of the value in an ecommerce queue lives in automation that isn't a clean resolution: the return the AI investigates and teed up for a human, the subscription edit it half-completed before a policy check, the triage that routed a VIP to the right person. That work is real, and a per-action model prices it honestly rather than treating it as unbilled-but-unrewarded. This is the whole reason the broader category of AI agents for customer service exists — to do reasoning-heavy work, not just close easy tickets.
Where Macha fits — on top of Gorgias, not instead of it
Macha is an AI agent layer that runs on top of the Gorgias you already use. The connector is live: you connect Macha to Gorgias, and it reads and writes the same tickets your team already works — it does not replace Gorgias, your Rules, your Flows, or your billing relationship with them. Setup is covered in connect Gorgias to Macha for automated ticket replies, and what native Automate does on its own is in Gorgias AI Agent explained.
What the layer adds is the multi-step, tool-using automation that per-resolution billing tends to under-serve: an agent that looks up an order and issues a return label through a custom tool, pauses or swaps a subscription via API, triages by intent before a timer or a rule fires, and drafts grounded replies your agents approve. You keep Gorgias as the system of record and the helpdesk; Macha handles the reasoning-and-action work on top. And Macha's cost is metered per AI action against a monthly credit pool — not per resolution — so a return that took six useful actions is priced as six actions, not billed as a resolution one month and invisible the next. There are no hard-coded numbers to memorise here; the pricing page carries the current credit model, and the screenshot above shows the real billing screen.
FAQ
How does Gorgias bill for AI Agent? On an outcome basis. Per Gorgias' billing docs, an automation fee applies only when AI Agent resolves a ticket without handing it to a human — roughly $0.90 per resolved interaction on annual billing and $1.00 on monthly, with overages at $1.50 per interaction above your bundled amount. Confirm current figures on gorgias.com/pricing.
Does one resolved ticket get charged twice? It can. Gorgias' documentation states that when AI Agent responds to and completely resolves a ticket on its own, you pay both a helpdesk ticket fee and an automation fee on that single ticket. If the AI hands the ticket to a human instead, only the ticket fee applies.
What's the difference between per-resolution and per-action billing? Per-resolution (Gorgias Automate) charges for a clean autonomous outcome. Per-action (an AI agent layer like Macha) charges for each unit of work the agent performs — a lookup, a tool call, a drafted reply — from a monthly credit pool. Simple deflections favour per-resolution; deep multi-step ecommerce workflows are priced more honestly per action.
Is Macha a replacement for Gorgias? No. Macha is an AI agent layer that runs on top of Gorgias through a live connector. It reads and writes your existing Gorgias tickets and complements Gorgias' native Automate — it does not replace your helpdesk.
Which model is cheaper for my store? It depends on your ticket mix. If most contacts are one-and-done deflections, per-resolution billing is elegant and predictable. If you automate returns, exchanges, and subscription edits — multi-tool workflows — per-action billing shows what the automation actually cost and avoids double-charging a resolution. Model your top ticket types before committing.
Want to see per-action automation on top of your own Gorgias queue? Start a free trial of Macha and connect it to Gorgias in minutes.
Add AI agents to your Gorgias
Macha resolves tickets end to end on Gorgias — no migration, no code.
Zendesk
Freshdesk
Gorgias
Front
Shopify
Stripe
Slack
Notion
Google Workspace
Confluence

