The technical half of a helpdesk migration is the half that gets planned. The half that decides whether it worked happens two weeks later, when some of the team is answering in the new inbox and some has quietly gone back to the old one.
Why people go back
Not because the new tool is worse. Because on day 9 an agent has a customer waiting and cannot find the saved reply they use eleven times a day, and the old tool is still open in a tab. That is a 40-second problem, and it gets solved by going backwards every single time, unless the tab is gone. Read-only access is not gone. An agent can still read the thread there, and reading a thread somewhere is most of the way to answering it there.
Set the shutdown date before the import, not after
Pick it while people still care about the migration, and send it in the same message as the launch date. ‘New inbox live on the 3rd, old one closes on the 13th’ is a plan. ‘We will switch the old one off once everyone is comfortable’ is a permanent second inbox. Nadia Haddad ran a 40-agent move that way and had all volume in the new tool by day 12. The attempt before it, with no end date, was still at 60 percent four months later.
Rebuild the top 15 saved replies before launch
Whatever else is unfinished on day one, the fifteen most-used saved replies need to exist. They are most of what an agent touches in an hour, and their absence is the single most common reason somebody reopens the old tab. Everything else — reporting views and the tag structure — can land in week three without anyone noticing.
Watch one number for a month
Share of conversations answered in the new tool, plotted by day. It should climb steeply and then flatten close to 100 percent. A curve that flattens at 70 is telling you one team has one blocker, and asking them directly on day 6 costs nothing at all.
Leave a Reply