Automatic AI Customer-Service Setup in Gorgias
Macros, rules and the knowledge base are set up from the brand's own policies and tickets before any model talks to a customer, so a new helpdesk starts from a working account.

Supplement brand: macro usage from 3% to 50%, 37 agent hours saved a month, ≈ €102K estimated annual impact.
TL;DR
- Macros, rules and the knowledge base are put in place before any model talks to a customer, so a new helpdesk starts from a working account, not a blank screen.
- A toolkit copies views, macros and rules from one Gorgias account into another, while each store keeps its own voice, categories and policies.
- A supplement brand went from 3% to 50% macro usage, saving about 37 agent hours a month, with an estimated €102K annual impact.
- Aidana leads this work.
The problem
Most Shopify brands pay for Gorgias and use a fraction of it. Typical audit findings: no folder structure, single-digit macro usage, unassigned tickets, social tickets auto-closing, no business-hours rules, and an AI agent that is switched off or answers badly because the knowledge base is empty. Agents type the same answer 50 times a day, and first replies take eight hours.
The process
- 1Free audit. Read-only access. We pull ticket volume, first response and resolution times, messages per ticket, macro usage, CSAT and top contact reasons.
- 2Cluster the tickets. Claude groups recent tickets by intent (where's my order, missing item, returns, subscription change, product question) and ranks them by volume.
- 3Draft the macro library. The day-one macros are drafted from the brand's policy documents and recent tickets, so the first replies match how the brand already speaks. We start with about 15 core macros.
- 4Rules and views. Routing by intent, sentiment escalation, VIP detection, business-hours replies, tags. Rules that already exist on the account stay in Gorgias, so a second, conflicting policy is never invented beside them.
- 5Knowledge base. Shipping, returns, tone and the help-center pages. The knowledge base is what a later reply is allowed to promise.
- 6Switch on the AI agent. Start with the safest intents (order status, shipping times), watch the answers, then widen.
- 7Weekly sprint with the client, adjusting rules to what customers actually do.
Under the hood
- Account-to-account toolkit. Exports views, macros and rules from a source Gorgias account and imports them into a target account.
- Per-store configuration. Each store's voice, categories and policies live in that store's configuration, separate from the shared tooling.
- One honest gap. Gorgias has no public API for its own AI Agent settings, so that one screen is still set by hand. Everything around it is prepared so the helpdesk can run the same way on the next store.
- Shopify integration so macros pull live order data and the AI agent can answer "where is my order" by itself.
Results
| Client | Before | After |
|---|---|---|
| Superfood supplement brand | First response 8h 1m · resolution 23h 31m · 3% macro usage | 50% macro usage, 15 core macros, 37.2 agent hours saved per month · ≈ €102,694 estimated annual impact |
| Australian fashion and beauty brand | Messy folders, low macro usage, unassigned tickets | ~20 new rules, refined macros, AI agent live, weekly sprints |
| UK multi-brand pet food group | A separate Gorgias account per brand | One Gorgias for the whole portfolio; first brand live in about 2 days |
Brands we have set up
- Full AI agent and website chat: Zorae, Lammeuld, Nutriment
- Full support management: Modtub, Saber X, Clear Drop, Jewellery Spot
- Macros, rules and routing: ERG Media, Bloom Cooking, WordsWorth and Black, Gioia, Vast Learning, Campervan, Eclipse.bike
Connected projects
- Led to our own model → Automatic customer-service rollout
- Every client is managed in → Artikuno OS
Images.

Ready to build your AI project?
Free 30-minute discovery call. No pitch. Just a conversation about what AI could do for you.
- Free · 30 min
- A concrete first step you can act on
- In German or English, online