Three things shipped to production.

01Track record
90.7%	M365 requests automated A fleet of scheduled workers that drains the M365 ticket queue: licences, mailboxes, shared access, distribution lists, new hire accounts. Every rule was derived from the real request history, not invented. The share handled without human intervention went from 77% to 90.7% during 2026.
Merges and splits	Directory restructuring Script-driven cross-forest migration with SIDHistory preserved, so file server access keeps working the next morning. Accounts move in capped waves, with the real state verified before and after each one, ServiceNow profiles repaired automatically and duplicates sorted by rules. The commercial migration tool was replaced by script and its licence was not renewed. The same tooling runs the other way for splitting an environment after a sale, and similar work has been delivered on other engagements.
31 scripts	Moved onto one foundation A fleet already running in production, with passwords sitting inside the scripts and no way to tell which ones ran last night, moved onto a single foundation in one day: centralized configuration, encrypted secrets, structured logging with retention, failure alerts, and a watchdog that confirms every morning that the night's jobs actually ran. No cleartext credential left in a script, and silence stopped meaning success.

02What comes next

If your admins spend their days on tickets a script should be handling, we should talk.

M365 requests automated