Not every old application should be replaced.
An application that is unfashionable but stable, well understood and cheap to run may be the best thing in the estate. The ones that need attention are those where the cost of change is rising, the knowledge sits with one person, the platform is losing support, or the business has outgrown what it does. Those are different problems with different answers.
Rewriting everything at once is how modernisation fails.
A full rewrite competes with the existing system for years, delivers nothing until it delivers everything, and is usually overtaken by a change in requirements before it lands. Meanwhile the original still needs maintaining, so the organisation pays twice and gets neither.
We assess, choose a treatment, then sequence.
Inventory what exists with its cost, risk and business value. Choose a treatment per application rather than one strategy for all: leave it, re-host it, refactor part of it, replace it with a product, or retire it. Sequence by risk and dependency so each step stands on its own and can be stopped without leaving a half-migrated estate.
What you are left with.
A written position on every significant application, a sequence with the reasoning attached, and a first tranche delivered rather than a plan. Where an application is being retired, the data and the obligations attached to it are dealt with explicitly.