RESEARCH
Enterprise Legacy Modernization with Generative AI
Investigating how Generative AI can help recover, preserve, validate, and transform the business knowledge embedded within legacy enterprise systems.
These are working propositions—not conclusions. They are intended to be investigated, challenged, refined, and, where the evidence requires it, rejected.
Research Propositions
The propositions below explore how Generative AI might assist organizations in understanding, preserving, validating, and modernizing the business knowledge embedded within legacy enterprise systems. Each proposition is a subject for investigation—not an assumption of fact—and may be supported, refined, challenged, or rejected as evidence develops.
P-001
Legacy Systems Contain Institutional Knowledge
UNDER INVESTIGATIGATION
Legacy enterprise systems contain institutional knowledge accumulated through decades of business rules, policies, decisions, exceptions, processes, and operational experience.
P-002
Business Rules Are the Core Value of Legacy Systems
EVIDENCE EMERGING
The enduring value of a legacy system lies not primarily in its technology, but in the business rules, policies, decisions, exceptions, and operational knowledge that the system implements and preserves.
P-003
Documentation Is Incomplete by Nature
UNDER INVESTIGATIGATION
Documentation of legacy enterprise systems is inherently incomplete because not all business rules, exceptions, decisions, dependencies, and operational knowledge are formally recorded as systems evolve over time.
P-004
GenAI Can Accelerate Knowledge Discovery
PROPOSED
Generative AI can accelerate the discovery of business knowledge embedded within legacy enterprise systems by analyzing and connecting information across code, data, documentation, rules, dependencies, and other system artifacts.
P-005
Human Validation Creates Meaning and Trust
EVIDENCE EMERGING
Only human experts can validate meaning, resolve ambiguity, and establish authority.
P-006
What If Code Is All We Have?
FOR INVESTIGATION
When source code is the primary surviving record of a legacy system, Generative AI may help recover candidate business rules and operational knowledge, but what can be reliably reconstructed—and what may already have been lost—requires investigation.
P-007
Institutional Memory Must Become a Living Asset
PROPOSED
Recovering institutional knowledge should not become another static archive. It should evolve as the enterprise system changes.
P-008
Every change should become part of institutional memory.
PROPOSED
Changes to an enterprise system should be captured with their purpose, authorization, implementation, impact and validation so that today's change becomes tomorrow's institutional knowledge.
P-009
Governance Must Remain Human Accountability
UNDER INVESTIGATION
GenAI may assist investigation, documentation, and decision support, but authority for consequential changes must remain attributable to authorized human decision makers.