Mike HeapMike is an experienced Product Manager who focuses on all the “non-development” areas of My AskAI, from finance and customer success to product design, copywriting, testing and more.
Table of Contents
- What is Gorgias Automate AI in 2026 (and is it the same thing as “AI Agent”)?
- How easy is it to set up Gorgias Automate AI?
- What channels does Gorgias Automate AI work in?
- What are the limitations of Gorgias Automate AI?
- What knowledge sources can I train Gorgias Automate AI on?
- What features does Gorgias Automate AI have?
- How do I improve Gorgias Automate AI responses?
- What resolution rate can I expect from Gorgias Automate AI?
- What AI model does Gorgias Automate AI use?
- What languages does Gorgias Automate AI work in?
- How secure is Gorgias Automate AI?
- Who is using Gorgias Automate AI?
- How much does Gorgias Automate AI cost?
- Does Gorgias Automate AI have a free trial?
- Is Gorgias Automate AI worth it?
- FAQs
Last Updated
Created time
Mar 3, 2026 02:10 PM
Title length (<60)
Author
Ecomm?
Tags
Image
Publish date
Mar 3, 2026
Slug
gorgias-automate-ai-agent-complete-guide-2026
Featured
Featured
Type
Article
Ready to Publish
Ready to Publish
I know why you’re here.
You either switched on Gorgias’ AI Agent (the thing most people still loosely call “Automate”), saw your first bill, and briefly considered a new career as a lighthouse keeper… or you tried to set it up and realized “plug-and-play” still involves a lot of knobs.
Either way, you’re trying to answer a very practical set of questions:
- What exactly is “Gorgias Automate AI” in 2026?
- Where does it work (email, chat, SMS, help center)?
- What can it learn from?
- How do you actually improve it without sacrificing your weekends?
- And, bluntly: how much is this going to cost?
This post is the straight-through answer, using what Gorgias publicly documents about AI Agent, plus the “Automate-adjacent” automation features that still show up in reporting and workflows.
What is Gorgias Automate AI in 2026 (and is it the same thing as “AI Agent”)?
TL;DR: In 2026, “Automate AI” is best understood as the Gorgias AI Agent (Support Agent + Shopping Assistant) plus adjacent automation features like Flows and Order Management that are measured alongside it. Gorgias has repurposed the Automate tab into a dedicated AI Agent area, while other automation tools moved under Settings.
Gorgias has been shipping automation for a while.
They launched automation features back in 2020, then rebranded those capabilities as Automate in 2023.
The bigger leap happened later, when Gorgias introduced its AI Agent in mid-2024 as a “conversational AI platform” for ecommerce support and sales, and has continued shipping updates since.
By 2026, the naming gets slightly confusing because “Automate” historically referred to multiple things.
In Gorgias’ own product updates, the UI has been reorganized so that the Automate tab is repurposed into a dedicated AI Agent tab, and other related automation functionality (including Flows and Order Management) has been moved elsewhere in settings, while still being reported together in automation analytics.
You can see this shift reflected in other updates, including how the product areas are described and reorganized in their release notes about improved AI Agent settings and the new navigation structure.
So if you’re evaluating “Gorgias Automate AI,” you’re really evaluating two layers:
First is the AI Agent itself, which Gorgias positions as an ecommerce expert that can resolve support questions and perform certain automated actions without a human in the loop.
Second is the surrounding automation layer that often sits beside (or underneath) AI Agent in real deployments and in reporting, including Flows and Order Management, plus a legacy/retired feature called Article Recommendations.
One critical scope detail that catches people: per Gorgias, the AI Agent is explicitly Shopify-only. Gorgias’ broader helpdesk supports more than Shopify, and some automation features may support other platforms, but the AI Agent itself is positioned as Shopify-only in current docs.
How easy is it to set up Gorgias Automate AI?
TL;DR: Setup is mostly UI-driven: connect Shopify, add knowledge, test conversations, then deploy to Email/Chat/SMS. Gorgias doesn’t publish a time estimate, but the documented steps make it realistic to launch quickly and then iterate for weeks if you want strong performance.
Gorgias has a structured setup checklist for AI Agent with clear prerequisites: you need an AI Agent subscription, a Shopify store connected to Gorgias, and the right permissions (Lead/Admin/Account Owner) to configure it.
The official setup path is laid out in set up and go live with AI Agent. In practice, it follows a predictable sequence:
You update Shopify permissions, you add knowledge, you optionally configure tone of voice and handover behavior, you test, and then you choose where it should go live.
This is the part Gorgias generally gets right: because AI Agent is native to the helpdesk, the initial activation flow can be kept simple.
Where “setup” starts to feel less simple is when you move past “it’s enabled” and into “it’s good.”
The docs put a lot of emphasis on ongoing optimization loops—Guidance quality, content enable/disable, topic performance, coaching feedback, and opportunities analysis—because that’s how you raise automation rate without the agent hallucinating or escalating everything.
If you want a faithful mental model of the lifecycle, Gorgias describes AI Agent as Analyze → Train → Test → Deploy, and then you monitor + coach.
That’s not marketing fluff; it maps closely to how you’ll actually run it.
My AskAI is a direct swap for Gorgias AI Agent, with an approved app in the Gorgias marketplace, entirely no-code, you can set it up in under 10 minutes and save 5-8x the cost.
What channels does Gorgias Automate AI work in?
TL;DR: AI Agent can be deployed on Email, Chat, and SMS (SMS requires the SMS add-on). Flows and Order Management run on Chat, Help Centre, and Contact Form rather than email.
Channel coverage depends on which piece of “Automate AI” you mean.
For the AI Agent itself, Gorgias documents deployment on Email, Chat, and SMS, with “more channels to come.” SMS support requires the SMS add-on, and then you can attach AI Agent to one or more SMS-enabled phone numbers once SMS is active.
For the broader automation features:
Flows are described as channel agnostic, and can be deployed in Chat, Help Centre, and Contact Form. That “channel agnostic” wording matters: Flows aren’t just a chat-only toy; they can be used across shopper-facing surfaces that aren’t email.
Order Management is similarly exposed to shoppers via Chat, Help Centre, or Contact Form, and is Shopify-only.
One other channel nuance: Gorgias as a helpdesk supports a wider set of channels overall (Gorgias markets omnichannel support including things like social channels and add-ons), but AI Agent automation is explicitly documented for Email/Chat/SMS.
If you’re hoping AI Agent will autonomously reply everywhere you receive tickets in Gorgias, you’ll want to confirm what’s currently supported in your exact plan and channel mix.
My AskAI works in all Gorgias channels (including social).
What are the limitations of Gorgias Automate AI?
TL;DR: The hard constraints are Shopify-only for AI Agent, knowledge ingestion limits (no embedded media extraction from URLs, no image processing inside uploaded docs), and channel-specific gaps like Vision only working via Email. Billing/reporting also has a 72-hour attribution delay for automated interactions.
Gorgias is unusually explicit about several limitations that you should plan around up front, because they affect both performance and operational fit.
The biggest platform constraint is that AI Agent is Shopify-only and explicitly not supported on BigCommerce, Magento, or WooCommerce, even though some other automation features (like Flows) may support additional ecommerce platforms.
If you’re not Shopify, AI Agent itself isn’t an option in the way Gorgias documents it.
On data access, Gorgias notes that when answering shopper questions, AI Agent can only access Shopify order data from the last 10 orders. That’s a very specific constraint, but it matters for repeat buyers and edge-case order history questions.
On knowledge extraction, there are multiple practical constraints:
Public URL syncing can’t extract embedded media (images/video/PDFs) from the page, and the URL must be public and not block crawlers. If your policy pages rely on embedded PDFs or heavy visual content, you may need to rewrite the content into text-forward pages or help center articles.
Uploaded documents support formats like pdf, docx, pptx, and xlsx, but images inside documents can’t be processed, and Gorgias advises against uploading sensitive information.
On channel feature gaps, “Vision” (image understanding) is documented as available only via Email, not Chat. So if a large share of your “what is this?” or “look at this screenshot” issues come in via live chat, Vision won’t help there.
On automation capability constraints, Order Management specifically cannot trigger automatic refunds or cancellations. The refund/cancellation buttons submit a request that requires agent intervention. That’s not necessarily bad—it’s a safety rail—but it does mean “automated” isn’t always “fully self-serve end-to-end.”
Flows have their own limits too, such as the maximum number of flows enabled per integration and the number of answers per options step.
Finally, there’s a limitation that isn’t about functionality—it’s about measurement.
In the AI Agent Overview report, “automated interactions” are only counted after 72 hours have passed without human intervention, and the report shows no data for the past 72 hours. If you’re the type of operator who wants to ship a change and see results the same afternoon, this will slow down your iteration loop.
My AskAI charges based on ticket volume, so you can more easily predict your usage and keep your costs stable as your resolution increases. You can also connect your My AskAI to more knowledge sources for direct replies, including Google Drive, Notion, OneDrive and more.
What knowledge sources can I train Gorgias Automate AI on?
TL;DR: AI Agent can use Shopify store data plus your connected Knowledge sources: Help Centre articles, store website snippets, single-page public URLs, uploaded documents, and Guidance. Constraints include public-only URLs and limited extraction for embedded media and images in documents.
If you want AI Agent to behave like a competent tier-1 agent, the real work is knowledge hygiene.
Gorgias describes AI Agent as being trained on a blend of your Shopify store data, your policies and brand guidelines, and the “Knowledge” you connect. In the Gorgias setup flow, Knowledge is a first-class configuration area, and Gorgias is explicit that you need at least one knowledge source before you can enable the agent.
The supported knowledge types Gorgias documents include:
Help Centre articles, with visibility controls affecting whether AI Agent can use them. Public and unlisted visibility matter here; if your help center content is hidden in a way the agent can’t access, don’t expect miracles.
Store website content, which is synced during setup and transformed into snippets (Gorgias describes this as generating Q&A-like snippets from synced content that the agent can use).
Single-page public URLs (your returns page, warranty page, blog posts, FAQs), with the constraint that the content must be publicly accessible and not block crawlers, and embedded media can’t be extracted.
Uploaded documents (pdf, docx, pptx, xlsx), with the constraint that images inside documents can’t be processed and Gorgias advises against uploading sensitive information.
Guidance, which is where you give explicit instructions. Gorgias documents that Guidance is considered before other knowledge sources or actions, and there are specific limits like a maximum number of enabled Guidance items per store and character limits per Guidance.
There’s also product information sourcing via Shopify catalogue and product pages, plus the ability to manage product information within Gorgias.
The practical takeaway: you’ll get the best results if your policies and FAQs are written as clean, plain-language text, and if you avoid relying on “the answer is inside this embedded PDF” as your primary source of truth.
My AskAI can create tasks, actions and connect to any API to improve responses. You can also connect your My AskAI to more knowledge sources for direct replies, including Google Drive, Notion, OneDrive and more.
What features does Gorgias Automate AI have?
TL;DR: The core features are AI Agent with two skill modes (Support Agent and Shopping Assistant), Guidance-driven behavior control, Actions for connected apps, multilingual resolution, Vision for email images, and analytics by intent/topic. Adjacent Automate features include Flows and Order Management.
Gorgias AI Agent is not just “a bot that replies.”
It’s positioned as two different skillsets you can choose from: Support Agent and Shopping Assistant, which is an important distinction if you care about pre-sales conversion versus post-sale operational support.
On the ecommerce side, Gorgias leans hard into shopping features: you can define selling styles, offers, and discount strategies for the Shopping Assistant, and there are also proactive engagement settings that allow outreach while customers are on-site.
On the support side, the features that actually matter operationally are:
Guidance, which is effectively your instruction layer. Gorgias documents best practices for writing Guidance and encourages structure that reduces ambiguity. It also supports using Shopify variables inside Guidance, which is useful if you want the agent to reference dynamic store data when responding.
Actions, which connect AI Agent to other apps and tools. Gorgias supports pre-made integrations and also supports building custom actions with advanced options. This is how you get from “answers questions” to “does things.”
Handover and exclusions, where you customize when AI Agent should escalate, what it should avoid, and how it should behave around handover.
Multilingual resolution, with Gorgias stating AI Agent supports over 80+ languages, along with language settings control.
Vision/image understanding, which Gorgias documents for email (not chat).
Ticket tagging, where Gorgias supports AI ticket tagging behavior configuration.
On the automation-adjacent side:
Flows provide multi-step Q&A scenarios that can guide, recommend, and triage across multiple shopper-facing surfaces.
Order Management provides self-serve order status and issue-reporting flows for shoppers, with the constraint that certain actions still require agent intervention.
My AskAI has these features plus Insights and analytics to review responses, can reply in internal as well as direct modes and has a self-learning feature to automatically improve AI knowledge.
How do I improve Gorgias Automate AI responses?
TL;DR: Improvement is driven by iterating Guidance, coaching AI Agent on real tickets, analyzing intents/topics and opportunities, and enabling/disabling knowledge sources to remove conflicts. AI Reasoning adds visibility into why responses happened for eligible tickets.
Most teams underestimate this part.
Turning AI Agent on is a moment. Improving it is a process.
Gorgias provides several first-party loops that are meant to keep you in control instead of “set it and pray.”
The first lever is Guidance. Gorgias states Guidance is considered before any other knowledge source or action, and they publish structured practices for writing it clearly to reduce ambiguity. If you take nothing else from “improvement,” take this: sloppy Guidance creates sloppy behavior, even if your help center is perfect.
The second lever is ticket-level feedback and coaching. Gorgias documents a coaching workflow where feedback helps AI Agent select the right knowledge sources and improve responses for similar questions in the future. This is basically “correct it while it’s wrong,” but inside the product.
The third lever is AI Reasoning visibility. Gorgias documents that AI Reasoning provides visibility into decision-making and lets you edit knowledge directly from the ticket view to correct behavior.
The fourth lever is knowledge governance: enabling/disabling content sources, using help center visibility toggles, and bulk managing what’s allowed. This is how you handle conflicts between sources (Guidance says one thing, help center says another, snippets say a third).
The fifth lever is performance analysis by intent/topic. Gorgias groups conversations by intent and shows intent-level performance, including improvement potential, success rate, and CSAT, and provides workflows for exploring performance by topic and optimizing based on insights from tickets. In plain language: it’s their way of telling you where the agent is failing and where the upside is.
One practical caution: measurement of automated interactions has a 72-hour attribution delay in reporting. So improvement cycles won’t feel “real-time” in the dashboards.
My AskAI allows you to inspect responses to better understand answers, has QA and testing features and also self-learns to fix knowledge gaps.
What resolution rate can I expect from Gorgias Automate AI?
TL;DR: Gorgias frames performance as “automation rate” rather than a universal resolution rate, with claims of resolving up to 60% of inquiries and case studies showing results like 26% resolved for one brand. Order Management documentation claims it can automate up to 30% of live chat ticket volume.
If you’re looking for a single number that applies to every business, it doesn’t exist (and anyone who gives it to you is selling something).
Gorgias measures performance primarily through automation rate and automated interactions, and tracks it across AI Agent and other automation features in the AI Agent Overview report. That report explicitly counts automated interactions only after 72 hours have passed without agent involvement.
In terms of public claims and examples, the range is wide:
Gorgias’ main site makes claims around AI Agent resolving a large share of support inquiries (positioned at 60% in marketing language).
Specific customer stories publicly show lower or mid-range results depending on the brand. For example, Gorgias case studies cite automation outcomes such as 56%, 30%, and 26% in different stories, and one documented AI Agent customer story reports 26% of tickets resolved for Psycho Bunny.
Order Management documentation separately claims it can automate up to 30% of live chat ticket volume.
The honest expectation to set internally is a range, not a promise. Your automation rate will be strongly shaped by the completeness and consistency of your policies, how repetitive your ticket mix is, and how aggressively you configure exclusions/handover.
My AskAI has resolved over 1.1m tickets to date with an average resolution rate of 72%.
What AI model does Gorgias Automate AI use?
TL;DR: Gorgias states AI Agent uses multiple LLM providers including OpenAI and Anthropic plus internally fine-tuned models, and runs a second AI model for QA with confidence thresholds. Gorgias does not specify exact model names/versions publicly.
Gorgias is relatively clear at the provider level and relatively opaque at the exact model-version level (which is pretty normal, even if slightly annoying).
In Gorgias’ own AI Agent documentation, they state the system uses “a combination of multiple secure LLM providers, including OpenAI, Anthropic, and internally fine-tuned LLMs.” They also describe a safety/quality mechanism where each response is checked by another AI model, and responses are only sent above a confidence threshold; otherwise the agent hands over.
What you won’t find in the public docs is a list like “we use GPT-X for email and Claude-Y for chat.” Gorgias describes providers and architecture patterns rather than pinning to specific model names or versions.
My AskAI use a suite of different models, namely from OpenAI and Google. We are also constantly running A/B tests and benchmarking to determine whether better quality, faster models can be used.
What languages does Gorgias Automate AI work in?
TL;DR: Gorgias states AI Agent supports multi-language resolution in 80+ languages, with control via language settings. For chat and automation surfaces, Gorgias documents a specific list of supported UI languages and notes Guidance is best written in English.
Gorgias makes two different language claims depending on what you’re talking about.
For AI Agent, Gorgias documents multi-language resolution in over 80+ languages, and notes you can control which language AI Agent uses via language settings.
Gorgias also states the agent can use knowledge content even if that content hasn’t been localized into the reply language, which is helpful if your help center is English-only but you receive multilingual inquiries.
Separately, for multi-language support in chat and related automation experiences, Gorgias documents a defined list of currently supported languages and notes that chat can auto-detect shoppers’ browser language.
Their list includes English (US), French (FR), French (Canada), Spanish, Danish, Swedish, Italian, Dutch, German, Norwegian, Czech, Brazilian Portuguese, English (UK), Japanese, and Finnish.
And a subtle but important operational note: even though AI Agent can read and understand multiple languages, Gorgias recommends writing Guidance in English for optimal results.
My AskAI is multilingual in 95 different languages.
How secure is Gorgias Automate AI?
TL;DR: Gorgias is SOC 2 Type II certified and HIPAA compliant (reports available on request), and documents AI safety measures like grounded responses and confidence-threshold QA. Gorgias also states AI Agent data isn’t used to train external LLMs and claims “zero data retention” for OpenAI API interactions—something you should verify in the security/DPA pack.
If you’re deploying AI in customer support, security isn’t a checkbox—it’s the thing your legal team will bring up at the worst possible moment.
Gorgias positions security across both the core platform and the AI layer.
On platform controls and compliance, Gorgias publicly highlights SOC 2 Type II certification, HIPAA compliance (with reports available on request), SSO options (including Google and Microsoft 365), 2FA, HTTPS/TLS encryption, encrypted backups, uptime monitoring, and Google Cloud Platform infrastructure.
On data location, Gorgias states data is maintained on Google Cloud Platform and hosted on servers closest to your location, and their pricing materials list regions including the US, EU, and Australia.
On data ownership and deletion, Gorgias states customers own their data, can delete shoppers’ data and tickets, and that customer data is removed from Gorgias data stores within 90 days of contract termination.
On AI-specific safety, Gorgias describes “grounded responses,” meaning the AI only answers based on verified content from your knowledge sources and Shopify data, and won’t respond if it can’t find an answer. They also describe the second-model QA system and confidence thresholds.
On external model data usage, Gorgias states it does not use your data to train external LLMs and describes “zero data retention” for OpenAI API interactions. There’s also an important nuance surfaced in Gorgias’ own security/privacy FAQ: the retention terms depend on provider configurations, so if this matters to you, you should validate the exact retention mode and subprocessors for each LLM provider in your DPA/security materials.
My AskAI is SOC-2 Type II certified and GDPR compliant.
Who is using Gorgias Automate AI?
TL;DR: Gorgias states it’s trusted by 15,000+ ecommerce brands, and its release notes claim 2,000+ brands adopted AI Agent in under a year. Public customer stories include brands reporting outcomes like 26% AI Agent ticket resolution.
Gorgias is widely adopted in ecommerce support. Publicly, Gorgias claims it’s trusted by 15,000+ ecommerce brands globally.
For AI Agent specifically, Gorgias states that in less than a year, over 2,000 brands have started using AI Agent to automate support and reduce resolution time.
There are also public customer stories that discuss automation outcomes, including one reporting that AI Agent resolved 26% of tickets for Psycho Bunny, alongside other cited automation figures like 56% and 26% in additional case studies.
As always, case studies are directional, not predictive. But they’re useful for setting a realistic expectation that “automation rate” can vary a lot across brands.
My AskAI is used by over 200 businesses around the world, with ticket volumes from a few hundred per month up to 100k per month. You can read our recent case studies here.
How much does Gorgias Automate AI cost?
TL;DR: Gorgias Helpdesk pricing is ticket-volume based (starting at $10/month for 50 tickets), while AI Agent is priced per interaction at $1.00 (monthly) or $0.90 (annual), and each AI Agent interaction also counts as a helpdesk ticket. Total cost depends on both ticket volume and automated interaction volume, plus add-ons like SMS.
Gorgias pricing has two layers you need to keep separate in your head: the helpdesk subscription and the AI Agent usage.
On the Helpdesk side, Gorgias publishes ticket-volume tiers, with starting points such as Starter from $10/month (50 tickets/month), Basic from $50/month (300 tickets/month), Pro from $300/month (2,000 tickets/month), Advanced from $750/month (5,000 tickets/month), and Enterprise being custom.
Gorgias also defines what counts as a billable ticket: any conversation started on any Gorgias channel that includes a message from a human agent, a rule, or AI Agent can count as a billable ticket, and replies after an inactivity period can count as a new billable ticket.
On the AI Agent side, the key line in Gorgias pricing is that AI Agent interactions are priced per interaction: $1.00 per interaction on monthly plans and $0.90 on annual plans. And importantly, Gorgias states that each AI Agent interaction also counts as a helpdesk ticket and is billed accordingly. This is the part that tends to surprise people: you’re paying AI usage, and those interactions also roll into ticket billing.
Gorgias also presents AI Agent tiers with included interactions and overage interaction costs on the pricing page, including examples where overages can be priced higher than the base per-interaction cost.
If you’re deploying AI Agent to SMS, SMS is an add-on with its own pricing schedule by SMS ticket volume, and AI Agent on SMS requires that SMS add-on.
The most reliable way to model your spend is to map:
- your current ticket volume into a Helpdesk tier, then
- estimate how many of those tickets will become automated interactions (and remember the 72-hour rule for “automated” attribution), then
- multiply automated interactions by $0.90–$1.00 (and watch for tier/overage pricing).
My AskAI plans start from $199/mo, with additional tickets costing from $0.10 per ticket. This works out around 5-8x cheaper than Gorgias AI. You can compare your ROI for Gorgias AI vs My AskAI here.
Does Gorgias Automate AI have a free trial?
TL;DR: Yes—Gorgias states it offers a 7-day trial with access to all features and no payment info required, with cancellation managed in Settings > Billing.
Gorgias’ pricing page states the trial lasts 7 days, includes access to all features during trial, and doesn’t require payment information to start.
If you decide not to continue, trial/subscription cancellation is managed in Settings > Billing.
My AskAI has a 30 day, unlimited free trial. Test every feature to your heart’s content and don’t pay a dime. No credit card required.
Is Gorgias Automate AI worth it?
TL;DR: It’s worth serious evaluation if you’re Shopify-based, live in Gorgias, and have lots of repetitive inquiries that can be fully resolved without agents. The main risks are Shopify-only constraints, knowledge/ingestion limitations, the 72-hour reporting delay, and the pricing structure where AI interactions are charged and also count as tickets.
“Worth it” comes down to one thing: can you convert enough of your ticket volume into fully automated interactions to justify the per-interaction cost, given that those interactions also count as helpdesk tickets?
The evidence-based reasons it can be worth it are pretty clear.
Gorgias is built for ecommerce, and AI Agent is designed to be grounded in your Shopify data plus your policy knowledge, with explicit handover and exclusions. It also has a real optimization loop inside the helpdesk: Guidance-first behavior, ticket feedback/coaching, reasoning visibility, and intent-level performance exploration.
But there are also real constraints you can’t “configure away.” AI Agent is Shopify-only. Knowledge ingestion has hard limits around embedded media and image processing inside docs. Vision is email-only. Automated interaction reporting has a 72-hour delay. And pricing is layered: helpdesk tiers plus AI interaction charges, plus add-ons if you want SMS/Voice coverage.
If you’re a Shopify-first brand running meaningful volume through Gorgias and your top issues are repeatable and policy-driven, it’s at least worth running the 7-day trial and measuring what happens. If you’re not Shopify, or your edge cases dominate, the fit will be much weaker based on Gorgias’ own documented scope.
If you want a full comparison of My AskAI against Gorgias AI, read one here.
FAQs
Is “Gorgias Automate AI” the same thing as Gorgias AI Agent?
In 2026, “Automate AI” is best understood as AI Agent plus adjacent automation features that are measured alongside it (notably Flows and Order Management). Gorgias has also reorganized the UI so AI Agent has its own dedicated area, with other automation tools moved into settings.
What’s the difference between Support Agent and Shopping Assistant in Gorgias AI Agent?
Gorgias lets you choose a skill for the AI Agent, making it either a Support Agent or a Shopping Assistant. Shopping Assistant includes ecommerce-specific capabilities like selling styles, offers, and discount strategies, while Support Agent focuses on handling support inquiries with policy and store knowledge.
Which channels can Gorgias AI Agent reply on?
Gorgias documents AI Agent deployment on Email, Chat, and SMS (SMS requires the SMS add-on). Other automation features like Flows and Order Management run on Chat, Help Centre, and Contact Form.
Can Gorgias AI Agent work if I’m not on Shopify?
Gorgias documentation explicitly positions AI Agent as Shopify-only and not supported on BigCommerce, Magento, or WooCommerce.
What knowledge can I connect to Gorgias AI Agent?
Gorgias documents multiple knowledge inputs: Help Centre articles, store website snippets generated from synced content, single-page public URLs, uploaded documents (pdf/docx/pptx/xlsx), Shopify store data, and Guidance.
Can Gorgias AI Agent extract information from PDFs linked inside a webpage?
Gorgias documents that URL syncing cannot extract embedded media such as images, video, or PDFs embedded in a page. If the information only exists inside embedded files, the AI Agent may not be able to use it as expected.
Can Gorgias AI Agent understand screenshots or images customers email in?
Yes—Gorgias documents a Vision capability for understanding images and screenshots, but it’s available via Email only, not Chat.
How does Gorgias measure AI performance—resolution rate or automation rate?
Gorgias frames performance primarily through automation rate and automated interactions, tracked in the AI Agent Overview report across AI Agent and other automation features. Automated interactions are counted after 72 hours have passed without human intervention.
What kind of automation rate should I expect?
Gorgias marketing includes claims around resolving a large share of support inquiries (e.g., 60%), while public case studies show a range including results like 26% tickets resolved for one brand. Order Management documentation claims it can automate up to 30% of live chat ticket volume.
What LLMs/models does Gorgias AI Agent use?
Gorgias states AI Agent uses multiple LLM providers including OpenAI and Anthropic, plus internally fine-tuned models. Gorgias also describes a second AI model used for QA with confidence thresholds, but does not specify exact model names/versions publicly.
Does Gorgias use my data to train external AI models?
Gorgias states it does not use data accessed by AI Agent to train external LLMs. Gorgias also describes “zero data retention” for OpenAI API interactions, but if retention controls are critical for you, you’ll want to confirm exact terms for each provider in the security/DPA materials.
How does Gorgias AI Agent pricing work?
Gorgias publishes AI Agent pricing per interaction: $1.00 per interaction on monthly plans and $0.90 on annual plans. Gorgias also states each AI Agent interaction counts as a helpdesk ticket and is billed accordingly, so your total cost depends on both helpdesk tier and AI interaction volume.
Does Gorgias offer a free trial for AI Agent?
Yes—Gorgias states the trial is 7 days, includes access to all features, and does not require payment information to start, with cancellation managed inside Settings > Billing.
