Problem guide
Legacy modernization discovery should reveal the work the old system quietly performs.
The hardest modernization risks are often not visible in the codebase. They live in workarounds, exceptions, reports, permissions, data habits, and institutional memory.
Why modernization scope is deceptive
Legacy systems usually do more than their official feature list. They encode local policy, informal shortcuts, compliance habits, reporting rituals, and workarounds that users may not mention until the replacement fails to support them.
Discovery that relies only on inventories and interviews can miss the behavior people depend on. RDD adds a practical test: show a candidate replacement slice early and watch where reality pushes back.
What to test before commitment
- Critical workflow slicesWhich paths must work because the business depends on them daily?
- Exception behaviorWhich edge cases are common enough to be operationally normal?
- Data trustWhich fields, histories, imports, exports, and reports carry decision value?
- Permissions and authorityWho can do what, who overrides, and who gets audited?
- Integration dependenciesWhich upstream and downstream systems quietly shape the replacement?
RDD modernization pattern
Start with one risky slice, not the whole platform. Build candidate reality that represents a future version of the workflow. Let real users and stakeholders operate it. Capture disagreement as evidence.
The output should include accepted behavior, rejected replacement assumptions, open migration risks, required Reality Checks, and a delivery baseline candidate. That is more useful than another modernization roadmap that has not met the work.
When to use RDD in modernization
Complex workflow, high compliance consequence, fragmented stakeholder knowledge, undocumented rules, or AI-enabled replacement pressure.
Simple replatforming, stable commodity workflow, absent users, or a purely technical infrastructure upgrade.
FAQs
Does RDD replace system analysis?
No. It complements analysis by testing how a future workflow behaves with the people who understand the legacy reality.
Can this reduce migration risk?
Yes. Candidate reality can reveal data, reporting, permission, and process assumptions before the migration plan becomes expensive to change.
Should the candidate connect to live systems?
Only when that is necessary and governed. Many early loops can use representative data to expose workflow and scope disagreement safely.