Post-Migration Operations
Not just a cutover — a system you can keep maintaining.
Version control, issue management and AI-assisted fixing carry over from the migration into daily operations.
Version control
Source code, compatibility libraries and settings managed on GitHub / GitLab. Every change records who, when and what.
Issue management
Defects, changes and new features registered as issues and linked to the fixes made.
AI-assisted fixing
AI investigates each issue and proposes a fix. Only fixes that pass automated tests and are approved by a person are merged.
Choose where AI runs
If source code must not go to external AI services, use an in-house AI server or our AI server over VPN.
Version control — GitHub / GitLab
The repository used during the migration carries straight over into operations. Cloud and self-hosted editions are both supported.
What is managed
- Converted source (screens, logic, reports)
- Versions of our compatibility libraries
- Build and deployment settings (ClickOnce / web)
- Test cases and results
Branches and releases
- Separate production, development and fix branches
- A tag for every release, so any version can be rebuilt
- Every change reviewed as a pull / merge request
- Build and deployment automated with CI as needed
What you gain
- Change history doubles as audit evidence
- Roll back to the previous version immediately
- Context survives staff changes
- Library updates arrive through the same process
Issue management — defects, changes, new features
Every request is registered as a GitHub / GitLab issue and labelled by type. Commits and pull requests are linked to issues, so the reason for every change can always be traced.
label: bugFixing defects
Symptom, steps to reproduce, expected result and screen are registered. AI isolates the cause — converted code or our library — and fixes it.
label: changeChanging features
Specification changes from new laws or business changes. AI lists the screens and processes affected.
label: featureAdding features
New screens, reports or interfaces. AI drafts an implementation in the style of the existing code for review.
Issue fields (templates provided)
How AI-assisted fixing works
From issue to fix, test and release, the cycle runs on AI and automated testing. People focus on deciding what to fix and whether a fix may go in.
Rules for using AI safely
- AI only proposes fixes; a person always approves the merge
- Every fix is kept as a pull request with its diff
- Fixes that fail automated tests (unit and RPA screen tests) are rejected
- Every fix is linked to an issue, so its history can be reviewed
The approach is the same as during migration. About AI-driven modernization →
Where AI runs — you decide how far your source code goes
For AI to fix code, it has to read the source. If your policies or contracts do not allow sending source code to external AI services, choose B or C.
Operations environment
The split used during migration — a conversion environment for your assets and a separate AI environment — carries over into operations. AI reads only registered issues and the relevant source code, returns fixes as pull requests, and changes go in only after automated tests and human review.
Choosing where AI runs
| A. External AI service | B. In-house AI server | C. Our AI server (VPN) | |
|---|---|---|---|
| Where source goes | External AI service | Stays in your network | Our AI server only (no external AI services) |
| Initial cost | Low (usage fees) | High (hardware and setup) | Low (VPN setup only) |
| AI capability | Latest large models | Open models run in-house | Open models run by us |
| Operations | Minimal | You maintain the hardware | We maintain it |
| Best for | External AI is permitted | Source may never leave / heavy use | No external sharing, but low capex |
B is quoted individually (hardware and cost). C is used over an encrypted VPN under an NDA. In both cases AI runs in a closed environment and your source code is never sent to external AI services. About our in-house AI server service →
Support menu
Operations setup
Repository, branch and permission setup on GitHub / GitLab, issue templates, and CI for build and test automation.
AI environment
Building an in-house AI server, or setting up the VPN connection to our AI server.
Maintenance support
Ongoing support for defects, changes and new features until your own team is up to speed.
Let’s talk about maintenance after migration
Tell us how you maintain your systems today and whether source code may leave your company. We will propose the right setup.
TEL +81-3-5962-3776