The helpdesk migration checklist: everything to verify before you switch

Most helpdesk migrations don't fail during the import. They fail three weeks earlier, when nobody checked whether the old plan allows bulk export, or three weeks later, when the SLA report quietly shifted by two hours and nobody could say why. The import itself is mechanical. The verification around it is where projects are won or lost.

This is the master checklist for that verification, organised by stage rather than by date. Whether your cutover is a single weekend or a month of parallel running, the boxes are the same — only the calendar changes. Work through it top to bottom and the move becomes boringly predictable, which is exactly what you want from a project that touches every customer conversation you have.

Stage 1: verify the destination before you commit

The cheapest place to catch a problem is before the contract. Confirm, with evidence rather than sales assurances:

• Every channel you run today has a home. Email addresses, website chat, messengers, social DMs. A channel without a landing spot becomes an emergency two days before cutover.
• The integrations that earn money exist. Billing lookups, order status, CRM sync. Nice-to-have integrations can wait; revenue-touching ones cannot.
• Import support covers your objects. Tickets, contacts, knowledge base articles, macros, tags, custom fields — ask for the list in writing, and ask what happens to attachments.
• The pricing model matches your team's shape. Per-seat pricing punishes growth; flat plans punish very small teams. Model your headcount 12 months out, not today's.

If you're weighing specific vendors, a structured side-by-side like our Freshdesk comparison is faster than twenty tabs of feature pages.

Stage 2: verify your exit path

Your current helpdesk decides how easy it is to leave, and you want those answers while you're still a paying customer in good standing:

• Can your current plan export conversations, contacts, articles and macros in bulk — or is export gated behind a higher tier?
• What formats do exports arrive in, and do they include attachments and inline images?
• When does your contract renew, and how many days of notice does cancellation require? Auto-renewal windows have trapped more teams than any technical issue.
• What does the vendor's policy say about data deletion after departure? You may want a grace period to re-export something you missed.

Stage 3: inventory what you actually own

You cannot verify a migration against an inventory you never took. One person, one document, half a day:

• Every email address that forwards into support — including the aliases someone created years ago and forgot.
• Every widget install and its pages, every connected messenger.
• The macro census: how many exist, which ten are actually used.
• Automation rules — written down in plain language, because rules rarely export cleanly and undocumented ones die silently.
• Custom fields on tickets and contacts, with their types.
• Knowledge base structure: categories, article count, drafts versus published.
• API consumers and webhooks — anything that reads from or writes to the old helpdesk programmatically.

Stage 4: verify the import path field by field

This is the stage teams skip, and it is the stage that decides whether your reports mean anything next quarter:

• Build a mapping table: every source field to its destination field, including type. A dropdown that arrives as free text will haunt your reporting for a year.
• Decide how ticket statuses translate — most tools have five to seven states and they never align one-to-one.
• Pick the deduplication key for contacts (almost always email) and decide what wins when two records collide.
• Confirm attachments and inline images are copied to the new tool's storage, not hotlinked to the old one.
• Run a sample import of 50–100 tickets first. Check timestamps against the source — a silent timezone shift moves every SLA metric by hours. Then check threading, authorship and attachment integrity before committing the full run.

Stage 5: verify the technical seams

The seams between systems are where downtime hides:

• Email routing: lower DNS TTL days in advance, and test forwarding end to end with a real message.
• Widget swap: know exactly which commit changes the snippet, and who can deploy it.
• Sign-in: confirm every agent can authenticate before cutover day, not during it.
• Webhooks and API consumers from Stage 3 all point at the new tool — an integration still writing to the old helpdesk after cutover is the classic source of ghost tickets.
• CSAT surveys send from the new tool and land in its reporting.

Stage 6: verify the people side

• Roles and permissions rebuilt, with at least two admins.
• A one-page cheat sheet mapping old habits to new ones: where views went, how to escalate, the new shortcut keys.
• A named champion per team who learns the tool a week early.
• First-week staffing that assumes a modest productivity dip, because there will be one.

Stage 7: verify the cutover itself

• Replay 25 golden tickets — real conversations from the last quarter — through the new setup and read every answer critically.
• Rehearse the rollback: the two or three steps that revert routing, written down, with a named owner.
• Freeze configuration changes on the old tool during the transfer window so exports stay consistent.
• Prepare the customer note, and keep it to two sentences. Customers care about answer quality, not your tooling.

Stage 8: verify after go-live

The migration isn't done at cutover; it's done when the numbers are trustworthy:

• Watch the queue for ghost silence — a channel producing zero tickets is usually misrouted, not quiet.
• Sanity-check the first SLA report against the old tool's baseline.
• Confirm search finds migrated history and articles.
• Keep the old tool read-only until the billing cycle ends, then confirm the cancellation in writing.

The one-page version

Fit before contract. Exit terms in writing. Full inventory. Field-by-field mapping with a sample import. Seams tested end to end. People prepared, not surprised. Cutover rehearsed with a rollback owner. Numbers verified after go-live. Eight stages, no heroics.

Where MoveDesk fits

MoveDesk exists because this checklist used to take a consultant. The white-glove migration is free on every plan: our team runs the export, mapping and import from Zendesk, Intercom or Freshdesk — attachments, articles, macros and history included — while both tools run in parallel for as long as you want to compare. The 14-day trial is long enough to work through every stage above on your real data before you commit to anything.

Print the checklist, book a quiet week, and start at Stage 1. The teams that verify twice migrate once.