We've moved 30+ DTC brands off Zendesk to Gorgias. Some did it because Zendesk had become too expensive. Most did it because their team was spending more time fighting the tool than serving customers. A few because they were on Shopify Plus and the Zendesk integration was, in plain language, an irritant.
This is the field guide. What to port, what to rebuild, where the risk lives, and how to do it without losing tickets or breaking SLAs.
The honest read on cost and time
Two numbers most agencies don't put on the page:
Time. 4–8 weeks end to end. A mid-market account with 50K tickets a year and moderate macro complexity finishes in five weeks. Heavy customization, multiple brands or a complex routing matrix push you to seven or eight.
Money. Setup project at €5K–€15K depending on scope, plus a one-month overlap on Zendesk seats during cutover. Net payback on a typical €120K/year Zendesk bill is 6–9 months.
If a partner quotes you a 2-week migration for under €3K, they're either skipping the discovery phase or planning to drop tickets. Both are bad.
What you're actually moving
A Zendesk account that's been running for three years has accumulated:
- Open and resolved tickets, with conversation history
- Macros (often hundreds, with overlapping coverage)
- Triggers and automations
- SLA policies
- Views and groups
- Custom fields on tickets and users
- Apps, integrations and marketplace add-ons
- Reporting dashboards
You're not going to move all of that, and you shouldn't. The point of a migration is to leave behind the cruft. Here's our split:
Port
- All resolved and open tickets with full conversation history. Searchable in Gorgias on day one. We use the Gorgias native importer plus a custom mapping layer for edge cases.
- Customer records with order history (this comes through the Shopify integration on the Gorgias side, not from Zendesk).
- Custom fields that map cleanly to Gorgias custom fields.
Rebuild
- Macros. Don't 1:1 import. We mine the last 90 days of resolved tickets in Zendesk to find which macros actually got used (you'll be surprised — typically 20% of the macro library handles 80% of volume). We rebuild that 20% in Gorgias's logic, not Zendesk's. Result: 30–60 macros instead of 300.
- Triggers and automations. Same logic. Zendesk's trigger model is more granular than Gorgias's rules engine. Don't try to replicate the granularity — replicate the outcome. You'll usually end up with half as many rules and the same coverage.
- SLA policies. Rebuild from scratch. Gorgias's SLA model is simpler. Map your Zendesk SLAs to Gorgias business hours and priority levels manually.
- Views. Rebuild around your actual team structure. Don't port "Tier 2 Returns Specialists" if that team doesn't exist anymore.
Drop
- Apps you stopped using.
- Integrations to tools you've since replaced.
- Reports nobody's opened in 6 months.
- Triggers that fire on conditions that no longer exist.
If the migration is done well, your Gorgias account is leaner than your Zendesk account, by design.
The cutover
This is where migrations go wrong. Get this part right and the rest is execution.
Two weeks before
- Final macro and rule signoff in Gorgias sandbox.
- Agent training in the new tool. Two sessions, recorded.
- Forwarder rules drafted (where do incoming emails go on cutover day).
- Communication to customers if you have a status page or they expect to email a specific address that's changing.
One week before
- Sandbox tested with live data clones.
- AI Agent enabled on test traffic.
- Agents practicing in the sandbox an hour a day.
- Backout plan documented (if cutover fails at hour X, revert to Zendesk by Y).
Cutover weekend
We do cutovers Friday evening to Sunday afternoon. The team comes in Monday to a working Gorgias.
- Friday 18:00: Zendesk goes read-only for inbound. Forwarder rules redirect new email to Gorgias. Existing Zendesk tickets remain open and editable for the team to clear out the queue.
- Saturday morning: Final import runs to sync any tickets created after the read-only flip.
- Saturday + Sunday: Quality assurance. We test every channel — email, chat, IG, WhatsApp, voice. AI Agent in shadow mode for the first 24 hours.
- Monday morning: Team starts on Gorgias. War-room channel open. AI Agent moves from shadow to live with conservative permissions.
- Monday + Tuesday: Quality monitoring every hour. We ship hotfix rules as needed.
30 days post-cutover
Zendesk stays read-only for 30 days. You can search old tickets, refer back to historical context. After 30 days, export a final archive and shut it down.
Where things go wrong
Three failure modes we've seen:
The "let's import every macro" team. They want every macro that ever existed in Zendesk to exist in Gorgias. The migration takes 50% longer, the Gorgias account is bloated from day one, and the team uses 20 macros while 280 collect dust. Don't.
The "we'll train the team after go-live" team. Cutover happens, agents log into Gorgias for the first time on Monday, productivity craters for a week, customer experience suffers, and the new tool gets blamed. Train two weeks before. No exceptions.
The "we'll just port the SLAs" team. Zendesk's SLA model has more knobs than Gorgias's. Trying to replicate every Zendesk SLA condition in Gorgias produces fragile rules that break in week three. Rebuild SLAs around the outcomes you actually need to hit.
The AI Agent question
Most brands moving from Zendesk to Gorgias also want to ship AI Agent at the same time. We support this — but with one rule:
AI Agent goes live in shadow mode for the first week post-cutover.
It watches every ticket, drafts replies that go to a queue for human approval, and learns. After a week, you flip on auto-send for the highest-confidence categories. Two weeks later, the next tier. By week four, you're at full AI Agent coverage.
This is the difference between an AI Agent that's useful in month one and an AI Agent that gets switched off in week two because it answered three things wrong on cutover day.
Communication to your team
The thing nobody warns you about: the team that's been running Zendesk for three years has muscle memory you're about to break. Some of them will hate the move for two weeks before they love it. Plan for that.
What helps:
- A clear "why" memo. Speed, native Shopify, AI Agent — whatever the reason. Write it down.
- Two weeks of sandbox practice before cutover.
- A "things you'll miss from Zendesk and what we're replacing them with" doc.
- A war-room channel where agents can flag friction in real time during week one.
Done right, the team is back to baseline productivity in five days and above it in three weeks.
A short list of things you should not do
- Do not migrate during peak season. Ever. Cutover in February or July, not November.
- Do not skip the discovery phase to save two weeks. The discovery phase is what makes the rest fast.
- Do not go live without the AI Agent in shadow mode first.
- Do not run both tools in parallel for more than 30 days. The team will sit in the comfortable old one.
- Do not let your sales team migrate themselves "because we're technical". The hidden cost is the productivity loss across the team during ramp-up. It always exceeds the agency fee.
When migration isn't the answer
Sometimes Zendesk is the right tool and the problem is elsewhere. We've sent two prospects back to Zendesk in the last 18 months — both above €100M GMV with multi-brand requirements that Gorgias couldn't meet cleanly.
If your problem is "Zendesk is too expensive and our team hates it", migration is probably the answer. If your problem is "we don't have a CX strategy and our inbox is a mess", the tool isn't the issue. Fix the strategy first, then evaluate the tool.
What we'd actually do for you
We'll run a two-hour discovery call, look at your Zendesk account in screen-share, and tell you on the call whether a migration is worth it. If it is, we scope it. If it isn't, we tell you what to fix in your Zendesk first.
If you sell on Shopify and you're serious about CX, you talk to custo.tech.
Book a discovery call — 30 minutes, no deck.