ISV development automation with AI: cut your X++ backlog

For a Dynamics 365 F&O independent software vendor (ISV), automating development with AI means handing the repetitive X++ items of the roadmap (extensions, data entities, services, batch jobs) to a generator, starting from an approved specification. The AI delivers code, unit tests and documentation in an Azure DevOps pull request; the product team keeps the architecture, the review and the release decision.

The goal is not to replace the X++ team, but to take off its plate the share of the backlog that follows known patterns, so it can focus on what differentiates the product.

Items with a clear scope and patterns documented by Microsoft:

• fields and methods added through extensions: extending a standard table, Chain of Command or event handlers;
• data exchange: data entities and DMF, services and OData actions, business events;
• processing and reports: SysOperation batch jobs, SSRS reports;
• quality debt: missing SysTest tests, BP check warnings.

The team keeps: product architecture, model and packaging choices, performance tuning on large volumes and any feature that is still ill-defined.

Vendor code is installed at many customers: it must stay in extensions, pass best-practice rules and be tested. Unlike overlayering, extensions avoid code conflicts and reduce the cost of upgrades, for the vendor and its customers alike. The CAR (Customization Analysis Report), generated by the xppbp best-practice tool, is one of the requirements of the solution certification process.

AI-generated code goes through the same gates as hand-written code: validation build, BP check with no blocking warning, SysTest tests, human review in a pull request. See the checklist for reviewing generated X++ and the quality gates.

1. The item is written as a precise ticket: context, business rule, acceptance criteria (AI ticket creation in D365 F&O).
2. Qualification and specification: the AI analyzes the ERP context and proposes the design.
3. Generation: X++ code and unit tests.
4. Delivery into your Azure DevOps as a pull request, reviewed by your team.
5. Merge and release on your own cycle.

With Skalp AI, steps 2 and 3 are handled; validation stays with you. On the Pro plan, context comes from your D365 instance and your Azure DevOps; on Basic, context is standard and the package is handed to your team.

Each development uses 1 credit: ≈ €400 on Basic, ≈ €200 on Pro and Max, decreasing with packs. Chat, qualification and design are free; the credit is only used when the code is generated. The stated average lead time is 2 h on Pro and Max, 4 h on Basic. The code, tests and documentation belong to you.

Go further: the Skalp AI for D365 ISVs page, the X++ development in 2026: AI or X++ developer pillar and the Basic, Pro and Max plans.

• Microsoft Learn, "Customization Analysis Report (CAR)", updated 26 Feb 2026, accessed 18 Oct 2026.
• Microsoft Learn, "Customize through extension and overlayering", updated 27 Mar 2026, accessed 18 Oct 2026.
• Microsoft Learn, "Testing and validations", updated 29 Apr 2026, accessed 18 Oct 2026.