Gorgias and Shopify are a match made in heaven for brands looking for a solid integration between their ecommerce platform and customer service operations. While linking these two platforms is relatively straightforward, you should regularly verify the integration works as expected as your business grows and changes.
Many of the issues that can arise between the platforms aren’t really the fault of either, as well: A macro might insert a blank customer name, two automations might reply to the same ticket, a cancellation could update Shopify without reaching your 3PL, a ticket that should reach a human could quietly land in the wrong queue, or an AI Agent might answer confidently using an outdated return policy, for example.
That’s why a Gorgias Shopify integration audit can help you catch the small details that go far beyond confirming that the two platforms are connected. Instead, a proper audit should verify your Shopify and channel mappings, user permissions, macros and variables, tags and ticket fields, views and routing rules, AI Agent Knowledge and Skills, Actions, tone of voice, handover conditions, and reporting. Most importantly, test these settings together using realistic customer scenarios.
In this blog, we’ll break down everything you should audit in your Shopify and Gorgias implementation to ensure it all works as it should, and you’re getting the most out of this powerful integration.
What Should a Gorgias Audit Include?
Before we get into how to audit Gorgias and Shopify, let’s first discuss what exactly you’ll need to include in the process.
A comprehensive Gorgias implementation audit should cover the entire path between an incoming customer message and its final resolution. This means you should be checking your Shopify store and support-channel mappings; integration permissions and customer/order data; macros and variables; rules and rule ordering; tags, ticket fields and views; routing and assignment; AI Agent Knowledge, Skills, Actions and tone; AI handovers and exclusions; Shopify order actions; and reporting and production performance.
All these areas are interconnected. Gorgias rules, for example, operate separately from the AI Agent. Rules process tickets before the AI Agent does, so a poorly scoped auto-reply rule can send one message, and the AI Agent can then send another. In fact, Gorgias specifically identifies this as a common source of duplicate customer replies.
To catch these issues, set up your audit to confirm your setup, deliberately test situations where things could go wrong, and then check whether the integration works as expected in that scenario.
Audit Your Shopify Store Connection, Channel Mapping, and Permissions
Your first layer is the Shopify connection itself. Connecting Shopify to Gorgias can surface customer profiles, purchase history, shipping information, fulfilment status, Shopify metafields and other order data inside tickets. It can also enable Shopify variables in macros, Shopify actions from tickets and macros, revenue reporting, and AI Agent functionality. Gorgias currently requires a connected Shopify store to use AI Agent.
First, if you run multiple Shopify stores, brands, or regional storefronts, confirm each one is connected to the right channels. Test tickets from each major customer-facing channel and confirm that they are attributed to the intended store. For example, you don’t want a Canadian customer to email your Canadian support address and be treated as a U.S. customer.
Next, validate the data that your agents need. Open controlled customer records and check whether the Shopify information your team relies on is available and accurate. That could include order history, fulfilment status, shipping information, Shopify notes, imported metafields and other customer context.
If Shopify customer matching is enabled in Gorgias, the platform can use phone numbers found in a customer’s default Shopify shipping address to help match and merge customer profiles. This setting is off by default, so include it in your integration audit rather than assuming it’s active. Customer notes can also be synchronized between Gorgias and Shopify, though Gorgias notes an exception for customers with profiles across multiple Shopify stores.
At this point, you’ll also want to audit permissions and access. Current Gorgias roles differentiate substantially between ticket access and account administration. For example, Basic users can perform third-party actions directly from tickets, including Shopify refunds or cancellations. Lead users can create and manage macros, rules, views and tags, while Admins have broader account, user and integration-management permissions.
Only the Gorgias account owner and Admins can manage connected stores and their channels, while ticket auto-assignment settings are also restricted to Admin-level access. Keep in mind that updating Gorgias’ Shopify app permissions additionally requires Admin access in Shopify.
During your audit, confirm that each role can do everything necessary for their job, but doesn’t have any additional access to tools and actions they don’t need or shouldn’t have access to.
Audit Gorgias Macros, Variables, Tags, Ticket Fields, Views, and Routing
Next up is your human-agent workflow. Over time, mature Gorgias accounts can accumulate years of macros, rules, tags, and views. Many will still be essential and relevant, while others are duplicates, remnants of old campaigns, previous policies or workflows nobody uses or needs. Your audit will help you clear out the clutter.
First, let’s audit your macros, since they can do more than paste text into a response. When Shopify is connected, macros can also contain actions that duplicate, cancel, refund or edit orders. Review your active macros for outdated wording, discontinued promotions, incorrect return windows, old shipping expectations, duplicate macros and transactional actions that are no longer appropriate.
Pay extra attention to variables, as Gorgias replaces variables without relevant data with a blank space, which can look pretty bad in customer communications.
Next, we’ll clean up tag taxonomy. Check for spelling variants, duplicates, legacy campaign tags, tags with unclear ownership and tags that nobody uses for routing or analysis.
Once you've cleaned up your tags, check your ticket fields. Gorgias currently allows unlimited ticket fields, but only 25 can be active at a time. Fields can be optional, required or conditionally displayed.
Gorgias also provides ecommerce-oriented managed fields such as Contact Reason, Resolution, and Product. You can customize or archive these if they don’t support your reporting needs. Gorgias also creates separate managed AI fields, including AI Intent and AI Agent Outcome, that AI Agent uses for classification and reporting. You can’t edit or archive these AI-managed fields. Before changing any editable field, confirm that rules, macros, or reports don’t depend on it!
Lastly, review test views and ticket routing, since your views are only useful when you can trust the tickets inside them. Create test tickets representing your most important conditions and verify that they appear where expected. That could include different stores, channels, VIP customers, priority levels, contact reasons, AI handovers or fulfilment exceptions. Then, verify assignment: A correctly classified escalation that sits unassigned in an overlooked queue is still terrible customer experience (CX).
Audit Gorgias Rules Before They Collide With AI Agent
Gorgias rules use triggers, conditions and actions to automate ticket behaviour. Since Gorgias prioritizes rules based on trigger type (and executes them from top to bottom in the order shown on the Rules page when multiple rules share the same trigger), the order of these rules should be verified.
Also double-check whether the Gorgias AI agent and your rules send duplicate replies: Gorgias rules run before the AI Agent, so if a rule includes a “reply to customer” action or applies a macro that sends a response, the rule can reply first. AI Agent may then process the ticket and send its own response, resulting in two messages.
Audit every active rule containing customer-facing replies or macros and, for each one, decide whether conventional automation or AI Agent should own that conversation. Where AI Agent should never participate, Gorgias provides a Prevent AI Agent rule that applies the ai_ignore tag. This could be useful for situations such as VIP customers, selected sender addresses, tickets containing particular words or tickets already handled by another process.
Pay close attention to older implementations as well: Gorgias’ legacy Autoresponder Rules are also worth checking for during an audit. Gorgias discontinued them for new accounts after September 2024, and they stopped working for existing accounts on January 30, 2026. If your implementation still contains documentation, workflows, views, or processes built around these deprecated rules, review and remove or replace those dependencies.
Audit AI Agent Knowledge, Skills, Actions, and Tone
As of September 2026, Gorgias organizes AI Agent behaviour around several major components: Knowledge, Skills, Actions, tone of voice and handovers. Each connected Shopify store can have its own AI Agent configuration. As part of your Gorgias integration audit, you’ll want to confirm how all those pieces work together.
Start with the knowledge your AI agent is using. This can include Gorgias Help Center content, guidance created inside Gorgias, website content, public URLs and uploaded documents. Each Shopify store with AI Agent enabled can have its own Knowledge library. Confirm that the important information is there and also correct/up-to-date.
Look specifically for ended promotions, old shipping thresholds, discontinued products, contradictory policies, regional differences and internal documents that shouldn’t have been in there in the first place.
Next, look at your agent’s Skills and ensure that the AI is taking the correct actions when encountering key situations. Skills and Knowledge are similar but still separate. For example, Knowledge tells the agent everything it needs to know about a topic, while Skills tell it the exact steps to take when someone asks about that topic.
Actions are the next thing to audit. These allow your AI Agent to make changes in the other applications you have connected to your account. This includes cancelling Shopify orders, editing shipping addresses, or managing workflows. Because Actions can affect real business data, it's important to ensure they run correctly.
For pre-built Actions from connected apps, Gorgias requires you to add them to Skill or Guidance instructions before they can run. Some custom Actions can also run automatically, although Gorgias recommends attaching Actions to Skills when you need tighter control. Audit the conditions under which every important Action can execute and ask whether the AI has enough information to act, whether confirmation is required where appropriate, whether the customer is eligible, and what happens when one step fails.
Finally, let’s look at Tone. Gorgias allows AI Agent to use a defined personality and tone of voice, with channel-specific behaviour available in its current tone-of-voice configuration. Some of the latest controls, including more granular wording and channel instructions, are still being released as beta capabilities. Review your settings to ensure they still align with your brand voice.
How Do You Test Gorgias Before Enabling AI Automation?
If you’re looking to get started with Gorgias’ AI features, you’ll want to take advantage of the AI Playground offered by the platform to test conversations without creating real tickets or affecting your reporting.
There’s an exception, though: Actions tested in the Playground can interact with real data in connected apps. Gorgias recommends using a dedicated test customer profile so you don’t accidentally modify a real customer’s order or subscription. Where an Action requires order data, use controlled test orders tied to that profile.
Come up with common scenarios, log your expected answer, action, escalation, and failure condition, and then run down the list to ensure that you’re getting the results you want before activating the AI Agent. To get you started, we’ve listed some common test scenarios below.
|
# |
Scenario / Input |
Expected Answer |
Expected Action |
Expected Tag / Field |
Expected Escalation |
Failure Condition |
|
1 |
Ticket arrives through Store A’s support channel |
Uses Store A policies and context |
Associates ticket with Store A |
Correct store metadata |
None |
Store B data or policy appears |
|
2 |
Customer/order belongs to another connected storefront |
Avoids presenting unrelated order details |
No wrong-store transaction |
Merchant identity/store field |
Human review if unresolved |
Cross-store information exposed or changed |
|
3 |
Shopify connection lacks a required permission |
Explains failure appropriately if customer-facing |
Action should fail safely |
QA defect recorded |
Admin/technical review |
Silent failure or false success |
|
4 |
User with insufficient role attempts an integration-management task |
N/A |
Access remains restricted |
Audit result |
Admin |
Excess privilege or required access missing |
|
5 |
Macro used with complete customer/order data |
Correct personalized response |
Intended macro actions only |
Expected business tags |
None |
Incorrect variable or action |
|
6 |
Same macro used with missing first name/order field |
Response remains readable |
No unsupported transaction |
QA issue if malformed |
Human edit if needed |
Blank produces broken customer-facing copy |
|
7 |
Macro contains Shopify refund/cancel/edit action |
Clearly communicates outcome |
Correct Shopify action runs once |
Merchant action field/tag |
Human on failure |
Wrong action, duplicate action or misleading message |
|
8 |
Ticket matches a key routing rule |
N/A |
Reaches correct queue/team |
Expected tag/field |
Correct specialist team |
Ticket lands unassigned or invisible |
|
9 |
Straightforward policy question covered by Knowledge |
Accurate current answer |
No unnecessary transaction |
ai_answered; ai_close if resolved |
None |
Incorrect, stale or unnecessary handover |
|
10 |
Ticket matches auto-reply rule and AI Agent |
Customer receives one response |
Only intended system replies |
Expected AI/rule state |
As designed |
Duplicate messages |
Conclusion
Gorgias and Shopify are both platforms that are constantly evolving. When you audit your implementation, you’re not just doing so to catch mistakes or errors; you’re also ensuring that everything you set up last month or last year is using the latest capabilities of each platform, in addition to updating everything to meet the current version of your business and policies.
It’s normal that policies and workflows change, new support channels appear, macros accumulate, or your knowledge gets a little stale. As long as you’re regularly testing and updating your integration, you can catch and fix issues before your AI Agent gives a customer the wrong information, or worse, takes action in your store that it shouldn’t have.
At Blue Badger, we help Shopify merchants implement, audit and maintain Gorgias across the full customer experience stack, including Shopify integrations, rules and ticket routing, macros, AI Agent setup, Knowledge and Skills, Actions, reporting and ongoing workflow optimization. Get in touch with us today to learn more.