The two-part path
| Data | Native route | Notes |
|---|---|---|
| Organizations | Data importer (CSV) | Load first |
| Users (requesters, agents) | Data importer (CSV); agents created with roles | Use skip_verify_email to suppress emails |
| Tickets and comments | Ticket Import API (/api/v2/imports/tickets, create_many for up to 100) | No UI import |
| Attachments | Upload through the API, then reference the token in a comment | Files can’t be referenced from CSV |
| Custom fields, tags | Set in the ticket payload | Fields must exist first |
Steps
- Create custom ticket fields and the tags you’ll use. Add a field for the legacy ticket ID.
- Import organizations, then users, keeping each one’s source ID in a user field.
- Build a lookup from source user IDs to Zendesk user IDs, and from organizations to organization IDs.
- Turn each CSV ticket into JSON: requester, assignee, subject, status, tags, custom fields,
created_at,updated_at,solved_at, and acommentsarray. Each comment gets anauthor_id, body,publicflag and its originalcreated_at. - Upload attachments and add their tokens to the right comments.
- Post in batches of up to 100 with
create_many, and usearchive_immediatelyfor closed tickets. - Reconcile: compare ticket counts by status, and spot-check threads and dates.
What the native tool does / where it stops
The Ticket Import API is built for history. It keeps timestamps, accepts several comments with their authors and skips triggers for imported tickets that aren’t closed. What it won’t do is read a CSV. One CSV row per ticket rarely holds the whole conversation either. Most exports put comments in a separate file, or leave them out: Freshdesk’s ticket export, for example, doesn’t include the entire conversation. Check that your source export includes comments before you plan around it.
Where FlitStack AI fits
FlitStack AI loads tickets through the import API with full threads, attachments and original timestamps. It keeps the legacy ticket ID in a custom field and produces a 301 redirect map. See the Zendesk migration page.