A customer wants a scalable implementation plan for modernising a large mainframe estate. What should the consultant recommend doing first?

The recommended first step is to run discovery across the estate to understand dependencies, data flows and transactions, and use those findings to identify sensible modernisation boundaries. A scalable plan needs boundaries that respect real dependencies and transaction flows; discovery is what supplies that information before any code is touched. Rewriting the programs judged most outdated by inspection alone is incorrect because it repeats the mistake of skipping dependency analysis. Modernising every program simultaneously to shorten the overall timeline is incorrect because doing so with no understanding of dependencies multiplies risk rather than shortening delivery. Skipping analysis and modernising strictly in alphabetical order of program name is incorrect because alphabetical order has no relationship to dependency structure or business importance. For the exam, discovery precedes and informs modernisation boundaries, at any scale.