Helpdesk Migration Checklist: What Moves, What Breaks, and What It Costs
What actually migrates between helpdesks like Zendesk, Kustomer, Freshdesk, HubSpot and Intercom, what has to be rebuilt, the hidden costs, and a cutover plan.
On this page 14 sections
- Start with why you are moving
- What moves and what does not
- Common migration paths
- Kustomer to Zendesk, or Zendesk to Kustomer
- Freshdesk to HubSpot
- Help Scout to Freshdesk
- tawk.to to Tidio
- Intercom via the conversations API
- The tools that do the moving
- A realistic timeline
- Why costs spike after a migration
- The cutover checklist
- The option of not migrating at all
- Frequently asked questions
In most helpdesk migrations, tickets, conversations, contacts, organisations, tags, custom field values, knowledge base articles and attachments can be moved. Macros, automations, triggers, SLAs, views, reports and integrations usually cannot, and must be rebuilt by hand. That rebuild, not the data transfer, is where most of the time and cost goes.
Key takeaways
- Data moves. Configuration is rebuilt. Plan the budget around the rebuild.
- Run a sample migration on 20 real records before committing to any tool.
- Costs often spike after migration because of parallel licences, higher tiers and AI add-ons.
- Freeze configuration changes in the old system two weeks before cutover.
- If the goal is AI resolution or lower cost, you may not need to migrate at all.
In most helpdesk migrations, your data moves and your configuration does not. Tickets, conversation history, contacts, organisations, tags, custom field values, knowledge base articles and attachments can usually be transferred with a native importer, the API or a migration service. Macros, automations, triggers, SLAs, views, reports and integrations usually have to be rebuilt by hand. Budget for the rebuild, because that is where the weeks go.
This checklist covers what moves, the tools that move it, a realistic timeline, the hidden costs that make bills spike afterwards, and a cutover plan. It also covers the option most teams skip: not migrating at all.
Start with why you are moving
Before exporting anything, write down the reason in one sentence. The common ones:
- A missing capability you genuinely need, such as omnichannel or a CRM integration.
- Cost, usually seat fees rising faster than headcount.
- Consolidation, for example moving support into the CRM your sales team already uses.
- Wanting AI resolution that your current tool charges extra for or does poorly.
If the reason is cost or AI, a migration may not fix it. A new helpdesk has seats too, and the queue follows you. We cover this at the end.
What moves and what does not
This table reflects typical behaviour across the major platform pairs. Every pair differs, so confirm with your target vendor or migration tool before planning.
| Data type | Usually migrates? | Notes |
|---|---|---|
| Tickets and conversation history | Yes | Replies, internal notes and timestamps usually carry over via API or a service. CSV exports often flatten threads. |
| Contacts (end users) | Yes | Watch for duplicates when one person has several emails. |
| Organisations or companies | Yes | Mapping to the target’s company object may need tidying. |
| Agents and groups | Yes, as records | Agents need new logins and permissions set by hand. |
| Custom fields | Values yes, definitions partly | Create the fields in the target first, then map. Dropdown options must match. |
| Tags | Yes | A good moment to prune years of unused tags. |
| Attachments and inline images | Usually | Check size limits in the target, and check that CSV routes do not drop them. |
| Knowledge base articles | Usually | Categories and folders move, but formatting, internal links and images need review. |
| Macros and canned replies | Sometimes | Some tools move text. Actions inside macros rarely map cleanly. |
| Satisfaction ratings | Sometimes | Often migrated as a field or note rather than native data. |
| Automations, triggers, workflows | No | Rebuild by hand. Logic differs between platforms. |
| SLAs and business hours | No | Recreate policies and calendars in the target. |
| Views, queues and routing rules | No | Rebuild, ideally from a reviewed list rather than a copy. |
| Reports and dashboards | No | Historical metrics rarely survive. Export what you need for year-on-year comparisons. |
| Integrations and apps | No | Reconnect every one, and find the ones other teams quietly depend on. |
| Chatbot flows | No | Rebuild. Different platforms model bots differently. |
Help Desk Migration, one of the best-known services, lists tickets, contacts, companies, agents, groups and knowledge base articles and categories as core record types, and says whether custom fields and other options are available depends on the platform pair.
Common migration paths
These are the pairs we see people search for most. None of them is unusual, and all of them follow the table above.
Kustomer to Zendesk, or Zendesk to Kustomer
Kustomer models support around a customer timeline, while Zendesk models it around tickets. Moving between them is less about export and more about mapping: which Kustomer conversations become which Zendesk tickets, and how customer attributes map to user and organisation fields. For Zendesk as a target, the Ticket Import API lets you set historical created_at, updated_at and solved_at timestamps and does not fire triggers on import, which is exactly what you want for history. See our guides for moving off Kustomer and moving off Zendesk.
Freshdesk to HubSpot
Usually a consolidation move. HubSpot supports CSV import of tickets, contacts and companies, but CSV is a blunt tool here: threaded replies, private notes and attachments are the usual casualties. For full history, teams use the API or a service; Help Desk Migration and ClonePartner both publish Freshdesk to HubSpot Service Hub guides. Our Freshdesk migration guide and HubSpot Service Hub guide cover the decision before the move.
Help Scout to Freshdesk
Help Scout mailboxes map to Freshdesk groups, and Docs articles map to the Freshdesk knowledge base. Saved replies and workflows need rebuilding. See moving off Help Scout.
tawk.to to Tidio
tawk.to exports chat history as JSON files inside a ZIP, and exports contacts separately. Tidio lets you import contacts from CSV. Moving historical chat transcripts into Tidio as conversations generally needs a migration service rather than a native importer, so decide whether you really need old chats in the new inbox or just an archive. See moving off tawk.to and moving off Tidio.
Intercom via the conversations API
Intercom’s REST API exposes conversations with their individual parts, which is how most scripts and services extract full threads. Rate limits apply per app and per workspace; check Intercom’s rate limit documentation for the current figures before estimating how long a large export will take. See moving off Intercom.
The tools that do the moving
| Route | Good for | Watch for |
|---|---|---|
| Native importer in the target | Simple pairs the vendor supports directly | Coverage varies. Some only import from one or two competitors. |
| CSV export and import | Contacts, companies, small ticket volumes | Flattened threads, lost notes and attachments. |
| Vendor APIs with your own script | Engineering teams wanting full control | Rate limits, pagination, error handling and re-runs are your problem. |
| Migration service (Help Desk Migration, ClonePartner and others) | Most teams with real history | Priced per record or per project. Run a sample migration first. |
On cost, Help Desk Migration’s published example is $100 for 1,000 records and $514 for 10,000 records on its Essential plan for a basic migration, with a free demo migration of 20 tickets and 20 articles. Custom options and higher support tiers cost more, so treat that as a floor and get a quote on your real volumes.
A realistic timeline
| Phase | Typical duration | What happens |
|---|---|---|
| Audit and decide | 1 to 2 weeks | Inventory config, integrations and data volumes. Write the reason down. |
| Build the target | 1 to 4 weeks | Create fields, groups, SLAs, views, macros and automations. |
| Sample migration | 2 to 5 days | Migrate 20 to 100 real records. Check threads, attachments and fields. |
| Integrations | 1 to 3 weeks | Reconnect billing, CRM, Slack, Jira and anything else. |
| Full migration and delta | Hours to days | Move history, then sync anything created during the move. |
| Training and hypercare | 1 to 2 weeks | Agents learn the new tool while volume continues. |
End to end, two to twelve weeks is normal, which matches what we set out in our migration guides. The spread comes almost entirely from configuration and integrations, not data volume.
Why costs spike after a migration
“Why did our costs spike after the migration?” is a question plenty of support leaders end up asking. The usual causes:
- Paying for two systems. Parallel running for a month or more doubles licence spend.
- A higher tier than expected. The feature you had on your old plan sits one tier up in the new one.
- AI priced separately. Seat plus usage stacking, which we explain in seat plus usage double billing.
- Migration fees. Per-record pricing on years of history adds up, especially with attachments or custom options.
- Productivity dip. Agents are slower for the first weeks. Backlog grows, and overtime or contractors fill the gap.
- Integrations rebuilt at a cost. Paid connectors or developer time to replace what used to work.
- Overage and usage billing. Some platforms now bill automatically for usage above commitments; see Zendesk’s automatic overage billing.
Put each of these in the business case before you sign, not after the first invoice.
The cutover checklist
Four weeks before
- Export everything from the old system and store it somewhere you control.
- Inventory triggers, automations, macros, SLAs and integrations. Mark each as keep, rebuild or drop.
- Audit the knowledge base: remove contradictions on refunds, pricing and cancellation.
Two weeks before
- Freeze configuration changes in the old system.
- Complete the sample migration and sign off on how threads, fields and attachments look.
- Build and test every automation in the target with real sample tickets.
Cutover week
- Run the full migration outside peak hours.
- Switch email forwarding, chat widgets and web forms to the new system.
- Run a delta migration for anything created during the move.
- Keep the old system read-only for at least one billing cycle.
After cutover
- Compare ticket counts, contact counts and article counts against the source.
- Watch first response time and backlog daily for two weeks.
- Cancel the old contract only after the export is verified and stored.
The option of not migrating at all
If your reason for moving is cost or AI resolution, consider whether a migration is the right tool at all.
An AI resolution layer can sit in front of the helpdesk you already have, answer repetitive questions from your existing help content, and create tickets in your current system only for what needs a person. Nothing is exported, no configuration is rebuilt, and your agents keep the tool they know. We compared the options in ways to add AI support without migrating.
Fidiora works this way: it runs alongside Zendesk, Freshdesk, Help Scout, Zoho Desk and other helpdesks, charges $0.69 per genuine resolution, dropping to $0.45 above 1,000 a month, with no seat fees, and hands off to humans for free. Setup is designed to take under an hour.
When a migration is still the right call: if your current helpdesk is missing a capability you genuinely need, if you are consolidating into a CRM for reporting reasons, or if the vendor is sunsetting your plan. A layer does not fix a missing feature in the tool your agents work in. In that case, use the checklist above and budget honestly for the rebuild.
A practical sequence many teams use: add AI resolution in front of the current helpdesk first, measure for a month how much volume genuinely needs a person, then decide on migration with that number in hand. It often changes the seat count you would be migrating to, and sometimes it removes the reason to move entirely.