Many organisations still run SharePoint applications built ten or more years ago: a fleet register, a training administration, a complaints log. They work, but they are built on old patterns, are hard to change and will not survive the next platform update. Rebuilding them used to be a project of weeks. With AI coding agents that changes fundamentally – provided you organise it well.

At the Shareflex Partner Dialog 2026 a striking example was shown: a fleet management application, more than ten years old, rebuilt on the modern Shareflex solution patterns with the AI coding tool Cursor and a Shareflex development kit.

The assignment

The instruction to the agent was short: refactor the legacy fleet management solution to the new Shareflex solution patterns, move logic from list events into modern form code where possible, make it bilingual (German and English), provide cockpit and navigation provisioning, and check whether fields or functions are missing that a fleet management system needs.

How it worked

The agent worked in plan mode: first read the legacy solution and the modern patterns, then propose a phased plan, then wait for approval, and only then implement – with a review and a changelog. The development kit guided the agent with rules, skills and reference examples, so the result follows the same structure as a solution built by hand.

The result

• 22 lists with fields, content types, views, forms and a form module each
• 5 server-side rules for the business logic
• German and English texts
• Navigation, a cockpit and one central provisioning script for deployment

Planning, approval and review remained necessary – the numbers describe scope, not a one-click miracle. But the step from “old app nobody dares to touch” to “modern, maintainable solution” has become much smaller.

What it means: from code to expertise

If AI writes most of the code, what does a customer pay for? Not for hours of typing, but for what the AI cannot decide: understanding the business process, choosing what the application should do, governance, data quality, adoption and taking responsibility for the result. That is how we work at Lialis too: we develop Shareflex solutions in VS Code with AI assistants and provisioning scripts (see Shareflex development with VS Code, Claude and provisioning), and spend the time we save on the process and the people.

Good candidates for modernization

• Old SharePoint list applications with custom code or workflows that are no longer supported
• Shareflex solutions that need to move to the New Core Infrastructure
• Applications that should become multilingual or get a modern form and cockpit

Frequently asked questions

Can AI rebuild an old SharePoint application?

An AI coding agent can do most of the implementation work when it is guided by clear patterns and rules and when a consultant plans, approves and reviews the work. It speeds up modernization; it does not replace the business analysis or testing.

Is the result production-ready straight away?

Not automatically. Review, testing in a test tenant and acceptance by the business remain necessary, as in any project.

Which applications are suitable?

Structured list-based applications – registers, case files, request processes – with a clear data model are the best candidates.

An old application that needs a new life?

Tell us what you have. We assess the application, propose a modernization plan and rebuild it on Shareflex in SharePoint.