Gorgias
Also known as: Gorgias AI Agent, Gorgias Automate
Ecommerce support platform whose AI agent resolves order, return, shipping and subscription requests end to end across email, chat, social, SMS and native voice, grounded strictly in connected Shopify and knowledge-base data.
Gorgias is a Shopify-native ecommerce support platform whose AI Agent takes order, return and shipping actions directly. Helpdesk plans start at $10 a month, and the AI Agent is a separate add-on billed per resolution, $0.90 on annual billing or $1.00 monthly.
Vendor details
Canonical URL
https://gorgias.com
Category
Customer support agent
Subcategory
Support — e-commerce
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
In practice
A Shopify customer asks where their order is and your rep digs through the admin to find out. Gorgias is Shopify-native and takes order and shipping actions directly in the support flow.
Return requests mean copying details between your support tool and your store. Gorgias handles returns as an action, so the agent processes it without leaving the conversation.
Your support tool and your storefront barely talk. Gorgias is built native to Shopify, so order, return, and shipping actions happen on the store data directly.
Agentic Index coverage score
10.0 / 14 capabilities · 71%
| Integrations & Tool Calling | Full |
|---|---|
|
Not Shopify-locked, and the agent writes rather than reads. Gorgias builds native integrations for Shopify, BigCommerce and Magento, with Shopify order, customer and catalog data living inside Gorgias natively at one-click install and no data syncing, and runs a live App Store spanning ecommerce platforms, subscriptions, returns, shipping, marketing and support channels, each listing labeled first-party or third-party, with named partners including Omnisend, ReturnLogic, ChannelReply, Recharge and Loop, plus Zapier and Make for the long tail. Authenticated action is documented: the AI Agent issues refunds, edits and cancels orders, changes addresses and updates subscriptions using live store data, and the Advanced options editor fires arbitrary HTTP requests to systems Gorgias does not natively integrate. Sourcegorgias.com/appsread 2026-09-05 |
|
| Workflow Orchestration | Full |
|
Agentic logic runs a loop rather than a script: the agent selects between answering from knowledge, performing an Action in a connected app, or starting a sales playbook, then asks whether the resources gathered will fully satisfy the request and launches additional parallel tasks if not, interpreting authored Guidance that branches on conditions such as VIP status, loyalty points and customer location until every step is complete. Sourcedocs.gorgias.com/en-us/how-gorgiass-ai-agent-works-1997817read 2026-09-05 |
|
| Knowledge Grounding & RAG | Full |
|
Knowledge sources are a first-class configuration the customer connects and the platform maintains: the Gorgias Help Center, the public website and product catalog, uploaded documents, connected public URLs and authored Guidance, all persisting across conversations and selected by relevance at answer time. Grounding is strict, with the agent documented as refusing to respond at all when the connected sources do not cover the question. Sourcedocs.gorgias.com/en-us/onboard-ai-agent-with-knowledge-sources-1216498read 2026-09-05 |
|
| Human Oversight & Guardrails | Full |
|
Three vendor-owned gates run before anything reaches a customer: an inbound check for malicious activity and excluded topics, customer-defined exclusion and handover topics that route sensitive subjects to staff, and a second AI model that must clear a confidence threshold or the conversation is handed to a human instead of sent. Every AI message is labeled Automated in chat and can be reviewed with its sources. Sourcedocs.gorgias.com/en-us/customize-how-ai-agent-hands-over-to-your-team-6008591read 2026-09-05 |
|
| Security, Identity & Governance | Full |
|
SOC 2 Type 2 since 2020 and HIPAA compliance published through a live Vanta trust center, with regular third-party penetration testing of application and infrastructure, GDPR and CPRA compliance, encryption and access controls over PII, SSO including Google, and a zero data retention policy required of model providers with interaction data deleted after processing. The attestation report itself is obtainable. Sourceapp.vanta.com/gorgias/trust/jx7664gsj0n01vjw4ubstread 2026-09-05 |
|
| Observability & Auditability | Partial |
|
Every AI-generated message is labeled Automated and can be reviewed alongside the sources it drew on, so which knowledge produced an answer is visible, and performance reports plus an Optimize page surface where the agent underperforms. Source attribution is not a decision trace, and no reasoning log covering which Actions were attempted, in what order, or why a branch was taken is documented, nor is a configuration audit trail. Sourcedocs.gorgias.com/en-us/how-gorgiass-ai-agent-works-1997817read 2026-09-05 |
|
| Memory & State Persistence | Not documented |
|
No memory layer, persistent state store or cross-conversation context is documented anywhere in the AI Agent documentation; the agent reads connected Shopify, helpdesk and knowledge data per request. The one persistence Gorgias does describe is ticket feedback influencing how the AI selects knowledge for similar questions later, which is learning absorbed into an optimization model rather than memory the buyer can locate or delete. Sourcedocs.gorgias.com/en-us/how-gorgiass-ai-agent-works-1997817read 2026-09-05 |
|
| Deployment & Data Residency | Partial |
|
Gorgias documents regional hosting on Google Cloud Platform across multiple servers globally, placing each account's data on the server closest to its location, so an EU customer is hosted in the EU. Placement is automatic by location rather than a customer-selectable region, and no VPC, single-tenant or on-premises option is offered. Sourcedocs.gorgias.com/en-us/how-gorgiass-ai-agent-works-1997817read 2026-09-05 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
Shopping Assistant ships as a named role-specific skill set the customer turns on as a unit and configures on its own terms, with its own selling style and discount strategy settings, sitting alongside the default support behavior and a standard set of ecommerce actions. Only one packaged skill set is documented rather than a browsable library, and Actions are authored by the customer. Sourcedocs.gorgias.com/en-us/customize-your-shopping-assistant-selling-and-discount-strategy-1216116read 2026-09-05 |
|
| Triggers & Channel Coverage | Full |
|
Broad channel coverage landing in one thread. Core channels are email, live chat, contact form and Help Center, with Facebook, Messenger, Instagram, TikTok and WhatsApp connected through the App Store, and SMS and Voice supplied by Gorgias itself rather than requiring a separate telephony vendor. The AI Agent is enabled per channel by toggle across those surfaces, with Gorgias noting SMS is the most tone-sensitive channel and should be tested before activation. SMS and Voice sit behind a separate add-on subscription. Sourcegorgias.com/blog/ai-agent-guardrailsread 2026-09-05 |
|
| Model Flexibility & Routing | Partial |
|
Vendor-internal blending across multiple providers with no customer selection. The AI Agent security and privacy FAQ states that its ability to recognize language and form sentences comes from a blend of state-of-the-art large language models from a few leading providers, including OpenAI and Anthropic, alongside Gorgias's own internal models, and Gorgias separately describes a direct OpenAI partnership giving deeper model access with ecommerce-tuned prompting layered on top. No model selection, routing control or bring-your-own-model option is exposed to the customer anywhere; which model serves a given turn is Gorgias's decision. Sourcedocs.gorgias.com/en-us/security-and-privacy-faq-for-gorgias-ai-agent-1997987read 2026-09-05 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
A public developer platform at developers.gorgias.com: a Developer Portal with sign-up, a Quickstart, a full API Reference and a Changelog, documented REST endpoints with pagination, rate limits and error handling, OAuth2 authentication with published scopes for public apps and HTTP Basic for private ones, and guides for creating integrations and widgets programmatically and for wiring a new channel in from a third-party app. Third parties build and submit public apps for review and publication into the Gorgias App Store, so the platform is extensible from outside rather than only configurable from within. Gorgias also publishes an llms.txt index of every documentation page in Markdown with endpoints in OpenAPI. SourceGorgias developer portal, API reference and llms.txt index; developers.gorgias.comread 2026-09-05 |
|
| Testing, Debugging & Optimization | Partial |
|
The Optimize page evaluates what the agent has already produced, identifying topics where it underperforms and directing the customer to the knowledge worth improving, and ticket-level feedback feeds back into how knowledge is selected; the pre-send confidence check is a runtime gate rather than a test. A test conversation feature at AI Agent > Test simulates customer interactions without affecting real tickets, reporting or customers, but no scored evaluation suite for trialing changes before go live is documented. Sourcedocs.gorgias.com/en-us/how-gorgiass-ai-agent-works-1997817read 2026-09-05 |
|
| Browser & Computer Use | Not documented |
|
No browser control, hosted session or computer use appears in any Gorgias documentation. The agent works programmatically instead: it acts through configured Actions against connected commerce APIs and native integrations, pulling live Shopify, Recharge and Loop data and writing back refunds, order edits and subscription changes, with an Advanced options editor firing HTTP requests to anything not natively integrated. Sourcedocs.gorgias.com/en-us/how-gorgiass-ai-agent-works-1997817read 2026-09-05 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Recent platform changes
Gorgias updated its AI Agent to display product review ratings and review counts directly within the chat interface. The agent automatically retrieves these review averages and counts from Shopify metafields to populate Product Cards and the Product Detail Page (PDP) drawer.
Bears on: Agent capability
View sourceGorgias introduced a 'More like this' feature for its AI Agent, enabling shoppers to explore similar products directly within the chat interface. When the agent recommends an item, users can tap an icon to instantly view a selection of related products without leaving the conversation.
Bears on: Agent capability
View sourcePricing
From $10/mo
hybrid
Included quota
Basic (most common entry, $60/mo / $50 annual): 300 billable tickets, omnichannel inbox (email/chat/social/WhatsApp), core integrations, basic automation rules. AI automation is NOT included — it's an add-on at $0.90–$1.00/resolution. Starter ($10/mo): 50 tickets, up to 3 agents.
What is public
Gorgias (Shopify/ecommerce-native helpdesk, ~15,000 brands) bills by ticket volume, not seats (unlimited agents on paid plans, except Starter's 3). Helpdesk tiers: Starter $10/mo (50 tickets, monthly-only), Basic $60/mo ($50 annual, 300 tickets), Pro $360/mo ($300 annual, 2,000 tickets), Advanced $900/mo ($750 annual, 5,000 tickets), Enterprise custom (>5,000). The AI Agent ('Automate') is always a separate add-on. Annual saves ~16%.
Billing mechanics
Each plan includes a monthly billable-ticket allowance; a ticket is any conversation that gets a reply from a human, a Rule, or the AI Agent (one ticket per thread regardless of message count). The AI Agent is priced per resolution — $1.00/resolution monthly, $0.90 annual — and, per Gorgias's own FAQ, each AI resolution ALSO counts as a billable helpdesk ticket (double-billed), UNLESS the shopper reaches a human within 72h, in which case it's charged only as a ticket. AI interactions beyond the bundled allowance cost $1.50 each. Plans bundle ~90 (small tiers) to 2,500+ (large) automated interactions.
Cost watchouts
AI double-billing (a resolution counts as both a per-resolution fee and a ticket), $1.50 AI overage rate, automatic uncapped ticket overages (BFCM spikes can 2–4x the bill), AI is always an add-on (even at Enterprise), voice/SMS billed separately, USD-only billing (FX risk), annual plans non-refundable
Variable cost rationale
Two coupled meters (tickets + per-resolution AI, both with overages and the AI double-billing) plus seasonal spikes make real spend climb well above list — often 2–3x, occasionally far more
Additional watchouts
The $10 headline is unrepresentative — most teams land on Basic/Pro and real all-in cost runs ~$150–$2,000+/mo once AI resolutions, overages, and voice/SMS are added; AI double-billing has driven cost-shock complaints
Overage / add-ons
Ticket overages bill automatically (no warning, no service suspension): $0.40/ticket on Starter/Basic, $0.36 on Pro/Advanced, $0.32 on Enterprise. AI resolutions beyond the bundled allowance cost $1.50 each. Gorgias may recommend upgrading if a higher tier is cheaper.
Sales call required
Mixed (some tiers require a call)
Lowest paid plan
~$10/mo +$0.90/res
Commercial notes
~15,000 ecommerce brands; deep Shopify/BigCommerce/Magento/WooCommerce + 150+ integrations (Klaviyo, Recharge, Loop, Yotpo); AI Agent (Automate) enabled per channel, with SMS and voice behind a separate add-on; 3-day AI-influenced-sales attribution window
Key ambiguities
Real automation rates run ~26–56% (below marketing), so per-resolution budgeting is hard; bundled-interaction counts and Enterprise pricing aren't fully public; how reopened/inactive threads re-bill as new tickets adds variance
Cancellation / refund
7-day full free trial (no card); monthly or annual (~16% off, '2 months free'); Starter is monthly-only; annual plans are non-refundable (canceling mid-term forfeits prepaid fees)
Support SLA / resale
Revenue analytics unlock at Pro; dedicated success management starts at Advanced; Enterprise adds personalized implementation, optional on-site support, and custom volume; unlimited agent seats on all paid plans (Starter capped at 3)
Missing data
Enterprise pricing is custom (low five figures a year has been cited); exact bundled-interaction counts per tier and real automation rates aren't transparently published. AI resolutions are also billed as helpdesk tickets (double-billing).
Related vendors
- Decagon — AI agent platform purpose-built for customer support, with deep…
- 5.Y — Singapore early-stage platform whose GLUCOSE product deploys…
- Ada — Agentic customer experience platform whose AI agents autonomously…
- Aisera — Enterprise agentic AI platform that orchestrates specialized agents…
- ASAPP — Generative AI for enterprise contact centers that plugs into…
- Assembled — AI customer support orchestration platform that unifies autonomous…
Alternatives to Gorgias
The closest documented capability profiles to Gorgias among customer support agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Tidio9.5 / 14A lighter documented profile than GorgiasGorgias vs Tidio →
- Freshworks10.0 / 14Fuller documented coverage on Deployment & Data Residency
- Five911.5 / 14Adds documented Memory & State Persistence
- Glia10.5 / 14Adds documented Memory & State Persistence
- NICE CXone10.5 / 14Adds documented Memory & State Persistence
- Zendesk10.5 / 14Adds documented Memory & State Persistence
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded