Modernise critical software without stopping operations

Usually modernise. A rebuild discards years of accumulated business logic, much of it undocumented and all of it load-bearing. We only recommend rebuilding when keeping the architecture costs more than replacing it.

In most cases yes. Components are replaced incrementally while the system stays in service, often with both paths running and compared. Where a short outage is unavoidable it's planned and rehearsed, not discovered on the night.

We confirm a backup exists and that restoring it works before anything changes. Migrations are rehearsed against a copy and reconciled against the source, so you can verify the totals agree before the old system is retired.

Yes, and it's much of this work. We read the code, trace the data and write tests capturing what it currently does, including behaviour nobody intended. Those tests become the missing specification.

Existing test coverage more than anything, then documentation, data quality and how much downtime you can tolerate. We scope from the assessment rather than estimating blind.

Yes, that's the normal starting position. We begin with technical onboarding: reading the code, mapping infrastructure, verifying backups and documenting what we find.