One person holds the map
Often one or two people hold the full picture of what the code does — the end-to-end relationship between code, jobs, data and operational exceptions.
Concentrated knowledgeAS/400 Decommissioning & Core Banking Replacement.
The platform still works. The people who can safely change it are retiring.
The central risk
The risk isn't the platform. It's the roster.
IBM i (AS/400) is still one of the most reliable platforms in banking — most shops report better ROI than on competing platforms, and most plan hardware or software upgrades this year. The platform isn't the problem. For the first time in nine years, industry-wide surveys show skills availability has overtaken cybersecurity as the #1 IBM i concern. The people who wrote and still understand your core banking logic are retiring — and almost nobody is being trained to replace them.
Why this matters now
The system may be stable while the organisation's options are narrowing.
Retirement dates, scarce internal knowledge and undocumented production behaviour can set the programme's timeline long before the technology fails.
The talent pool isn't refilling
RPG and CL are rarely taught, so the specialists who can still safely read and change the system aren't being replaced as they retire.
Structural skills gapProduction is the specification
Decades of undocumented patches sit between the spec and the production behaviour, making current behaviour more authoritative than the documents intended to describe it.
Undocumented driftRetirement changes the clock
A retirement date, not a project plan, ends up setting the timeline for what must be recovered while the people who know it can still challenge the reconstruction.
Succession riskThe false default
Traditional replacement starts by choosing the destination.
It treats the problem as a hiring gap when the pool isn't refilling, chooses the target core banking platform before the current logic is understood, assumes vendor documentation matches decades of production patches, and waits for a full rewrite before capturing what the last expert knows.
The working path
Capture. Replace or extend. Prove.
The path should first make the current system reviewable, then let the organisation move at the pace the business can absorb.
- Capture
Get the knowledge out of the people and into evidence
Interview the remaining RPG/CL/DB2 specialists — often just one or two people — and build a living map of what the system actually does, ahead of a retirement date.
- Replace or Extend
Move at the pace the business can absorb
A structured, evidence-based replacement onto a modern core banking platform — or a managed extension of the current system's safe life while the target is prepared. Client's choice.
- Prove
Validate every migrated rule against production history
Deterministic testing against what the AS/400 system actually did — not just what it was specified to do.
Technology landscape
Understand the source before a target platform narrows the question.
Most IBM i shops report strong ROI and keep investing in the platform. The constraint isn't the technology — it's finding people who can still safely read and change it.
Source estate
- IBM i (AS/400)
- OS/400
- RPG (III / IV / ILE)
- CL
- DB2 for i
- COBOL
Target core banking platforms
- Avaloq
- Temenos
- Finnova
- Olympic (ERI)
- TCS BaNCS
- Cloud-native cores
Decisions, not theatre
How we help.
Four ways to work with the current estate, depending on what the organisation is ready to decide.
Knowledge Capture & Documentation
We interview the remaining RPG/CL/DB2 specialists and build a living map of what the system actually does, ahead of a retirement date.
Code Modernisation
Move RPG/CL logic to modern languages, or refactor it within ILE, once business meaning is separated from implementation history.
Core Banking Platform Replacement
A structured, evidence-based replacement onto a modern core banking platform, validated against production history.
Managed Extension
Keep-safe support that extends the current system's safe life while the target platform is prepared.
Questions worth asking
Before the programme writes its own answer.
Do we need to pick the target platform before starting?
No. Understanding the current logic comes first — choosing a destination too early lets the target model quietly decide which history and behaviour no longer count.
What if only one person understands the system?
That's the typical starting point, not an edge case. A single specialist holding the full picture is common, and it's exactly why capturing that knowledge early matters.
Can Qubiz keep the AS/400 running while we plan?
Yes — managed extension. We can extend the current system's safe life while the target platform is prepared, on the client's timeline.
Where does AI fit?
AI accelerates code reading and comparison, but release decisions stay with the people who understand what the system means.
A useful first conversation
Find out which part of the AS/400 estate depends on knowledge the organisation cannot easily replace.
Bring the system, history and decision the organisation cannot afford to misunderstand. The first step is to establish what evidence exists and what must become reviewable.