Japan’s Information-technology Promotion Agency (IPA) reported in its 2026 DX survey that AI adoption is spreading, particularly among larger companies, but outcomes remain concentrated in efficiency and speed rather than new value creation or business transformation. That gap is a useful warning for the wider region: adding AI to a fragmented system landscape does not automatically create an AI-ready enterprise.
The practical answer Before an enterprise can scale AI, it needs visibility into IT assets, cleaner system boundaries, reliable data flows and architecture that can change without destabilising core operations. AI readiness is increasingly a modernisation outcome, not a software-licensing decision.
Why legacy systems become more visible in the AI era
Japan’s Ministry of Economy, Trade and Industry (METI) has warned that legacy systems can block the introduction of newer digital technologies and reduce industrial competitiveness. Its legacy-modernisation report links better modernisation outcomes with stronger management visibility, clearer IT-asset understanding and greater internal capability.
Generative AI and agents intensify the issue because they need access to current data and executable systems. A chatbot can sit outside the core architecture. An agent that needs to create an order, change a customer record or trigger a workflow must interact with the underlying systems safely.
Four signs the enterprise is not AI-ready
- Critical business logic is undocumented and concentrated in a small number of legacy specialists.
- Data is copied manually between systems rather than exposed through controlled interfaces.
- Teams cannot easily map which applications depend on which databases or integrations.
- Every new automation requires custom point-to-point work and regression testing across multiple systems.
Modernisation is not the same as replacing everything
A common mistake is to frame modernisation as a binary choice between keeping the legacy system and replacing it. In practice, enterprises can sequence the work.
- Make the estate visible: inventory systems, dependencies, owners, interfaces and data flows.
- Separate systems of record from systems of engagement so customer-facing change can move faster.
- Expose stable APIs around high-value business capabilities before attempting large-scale replacement.
- Modernise the highest-risk or highest-change areas first rather than rewriting the entire stack.
- Create observability and automated testing so change becomes safer and more frequent.
IPA’s 2026 draft indicators for data and IT-system utilisation reflect the same direction: connect DX strategy to the actual state of enterprise data and IT systems, rather than treating transformation as a separate innovation programme.
What this means for enterprise-tech startups
Legacy complexity creates opportunity for startups, but only when the product fits real operating constraints. Enterprises need tools that reduce migration risk, improve observability, expose APIs, automate testing, map dependencies, govern data and support phased change.
The strongest startup proposition is rarely “replace your entire core system.” It is often “remove one source of friction while making the architecture more adaptable.”
The founder GTM challenge
Selling modernisation technology requires a different story from selling a productivity app. Buyers care about operational risk, compatibility, security and migration path. A founder should be able to explain what can remain unchanged, what must change, how rollback works and how value appears before the full programme is complete.
Global Apex Tech’s Enterprise Tech desk and software-defined warehousing analysis provide useful internal context: connected operations create value only when the underlying systems and data can work together.
A practical modernisation sequence for AI readiness
- Inventory and dependency mapping.
- Identity, access and API boundary definition.
- Data-quality and master-data remediation.
- Observability and automated testing.
- Incremental service extraction or platform migration.
- AI or agent deployment only where system actions can be governed and audited.
The enterprise decision
The AI era makes technical debt more expensive because every new intelligent workflow depends on the reliability of the systems beneath it. Enterprises that modernise deliberately can add AI as another operating capability. Enterprises that skip the foundation risk turning AI into another integration layer on top of existing complexity.
Published by the Global Apex Tech Editorial Desk. Partner involvement, when applicable, is disclosed above the headline. For editorial questions or source material, contact editor@globalapextech.org.