Privileged accounts get moved when a privileged-access platform is replaced, when a merger brings in a second estate or when an outsourcer takes over infrastructure. On a plan the job is to copy the account records from the old platform to the new one and switch the old one off.
The quickest way is to move everything as it is, a lift and shift. It carries retired and wrongly configured accounts across with the good ones, and some of them then fail on the new platform.
Until July I led one of three workstreams on a large privileged-access programme at a global bank: hundreds of thousands of accounts migrated off a legacy platform. I did not do this alone. It rested on the team, and in particular on a few very experienced data analysts and specialists. Without them it would not have been possible. What follows is the short list I would work from again.
What to check before an account moves
Before an account moved, we checked two things.
First, it works. The current platform can connect to it, on the system it sits on if it is a local account or in the directory if it is a domain account, and can change its password at the interval set for it. If the password change keeps failing, you cannot tell whether the account is still valid. The system may have been decommissioned, the account may have been retired and never removed from the platform, or its record there may be configured wrongly. An account that is no longer active should not be migrated at all.
Second, its record fits the new platform, which will have its own requirements and its own data model. An account can work on the old platform and still be configured in a way that does not fit the new one. The hardest case to catch is a field that is filled in and works on the old platform but needs a different value on the new one. A check for blanks will not find it. The value has to be looked up in an authoritative source before the move.
What moving bad data costs
The obvious cost is that the account stops working. There are others.
If the new platform cannot change an account’s password, the account is no longer properly managed. Inconsistent records make accounts hard to find. An account that arrives configured outside your standard, or the one your regulator expects, becomes a compliance finding, and its owner may be asked to fix something they did not cause. That then becomes the migration team’s problem.
The worst is production impact. Suppose an account is moved wrongly and nobody notices. It is then needed to fix a live problem on a critical system, it does not work, and the outage lasts longer while access is restored.
Check before and after, and keep a way back
An account had to pass the checks before it was allowed to move. After it moved, it was tested on the new platform. If it was not working, it was backed out to the old one. That only works while the old platform is still running and still able to manage the account, so it stays on until the checks pass.
We automated the checks because of volume: a very large number of accounts, a complex estate and tight timelines. A smaller firm may not need that. With far fewer accounts, the same checks can be done by hand. Manual or automated, every account should pass the same checks, and the process itself should be tested before you rely on it. Test the new platform too, for volume and for the ways your teams use it.
Move the least-used accounts first
The analysts sorted the accounts by how often they were used. We agreed a risk-based order with management, in case something went wrong: rarely and never used accounts first, the most used later. Ownership and sensitivity counted too. We started with some of the accounts owned by our own security teams and by people we knew, and moved to the more sensitive areas afterwards.
Each tranche had criteria an account had to meet before it could move. When a problem was spotted, someone raised a question or priorities changed, the engineers could adjust the criteria and block or unblock accounts.
Decide who does the moving, and tell people
A common approach is for the project to build the tooling, write the guides and ask each team to move its own accounts by a deadline. We judged that this would not work for us, given the complexity of the organisation and of the account estate and the time available, so the programme did the migration itself. A smaller firm can choose differently. Either way, write down what the migration team is accountable for and what the service owners and account owners are accountable for.
I wrote about ownership in an earlier piece, “Orphaned, then promoted”. It matters here because someone has to say whether an account is still needed before it is moved.
We worked out who would be affected by each tranche, among platform owners, service owners and account owners, and wrote to them in advance. People also need a way to reach the migration team, such as drop-in sessions. Any change adds friction for teams with settled ways of working. Management needs to back it and explain why it is being done, so that those teams are brought along.
Five questions before you start
- Do you have a full list of your privileged accounts, including any that sit in other tools or in none?
- Can your platform still connect to each account and change its password? For accounts outside it, how do you know they still work and are still needed?
- Does each record hold what the new platform needs, with the values it expects?
- If an account fails after it moves, how do you put it back?
- Which accounts will you move first, and who is accountable for each step?
Where I am on this
I have since set up on my own to do this kind of work for mid-size regulated firms. My experience of this is at a global bank, so how it scales down is my own judgement. I think most of it applies at a mid-size firm, and that most of it can be done by hand.
If you are about to change platform, which of the five questions can you not answer yet?
First published on LinkedIn on 12 November 2026. Discuss it there, or write to hello@aevixadvisory.com.