Why it matters
The risk in a migration lies in what the customer notices.
An open conversation that gets lost, an email that stops arriving or an agent who cannot see the history turn into a bad experience just as the team is learning the new tool. That is why the plan is built around switch-over day.
What gets migrated and what does not
Not everything is worth the trip, and deciding beforehand shortens the migration.
| What | Usually migrated | Worth reviewing first |
|---|---|---|
| Customers and companies | Yes, in full | Duplicates and unused fields |
| Open conversations | Yes, or closed in the old tool | Who owns them |
| Closed history | The period people look up | What law or contract says to keep |
| Knowledge base | Yes | Out-of-date articles |
| Saved replies and rules | Rebuilt | Which are still in use |
The phases of a migration
A migration runs in six phases, and the third is the one most often skipped.
- Inventory. What is in the current tool and what is used.
- Set-up. Channels, teams, hours and permissions in the new one.
- Trial. A test import and a small group working in it.
- Training. The team practises on cases before the switch.
- Switch-over. Channels are redirected on a low-volume day.
- Overlap. The old tool stays available to look things up for a while.
Switch-over day
On switch-over day email, chat and the other channels are redirected, and each is checked with a test message.
It is worth having written down how to roll back and who decides.
What tends to go wrong
Three failures keep recurring in migrations.
- Rules and automatic replies left active in both tools.
- Integrations nobody remembered that stop receiving data.
- Reports that cannot be compared because metrics are calculated differently.
