While exploring different AI possibilities around Business Central, I started looking at another area where consultants spend a considerable amount of time — documentation.
During implementations and support activities, I noticed that the information required for documentation is usually already available. The solution is built, the processes are defined, and the technical details exist within the system.However, transforming that information into clear and meaningful documentation is still a manual effort.
A consultant needs to understand the functionality first, identify the business purpose, and then create different documents depending on the audience.
A functional consultant may need process documentation.
A business user may need a step-by-step user manual.
A testing team may need detailed UAT scenarios.
Each document requires a different perspective, even though the source information is often the same. This made me think about a question:
Can AI help bridge the gap between Business Central technical information and business-ready documentation?
That question became the foundation of my next AI experiment — building an assistant that could understand Business Central functionality and help transform technical knowledge into practical documents.
The Challenge Behind Documentation
One thing I realized while working on this idea was that documentation is often underestimated. Creating good documentation is not simply about writing descriptions of pages and fields. It requires understanding the business process behind the functionality and explaining it in a way that makes sense to the person reading it.
A developer looking at a Business Central page may see objects, fields, actions, and technical details.
A functional consultant sees how that page supports a business process.
A business user sees a task they need to complete in their daily work.
A tester sees scenarios that need to be validated before the solution goes live.
The same functionality can have completely different meanings depending on who is using the document.
This was the interesting challenge. The information already existed, but converting technical details into business understanding required human effort and experience. That is where I started exploring how AI could help bridge this gap — not by replacing the consultant’s understanding, but by helping transform existing knowledge into useful documentation faster.
Providing AI With the Right Context
One important lesson I learned from my previous AI experiments was that the quality of the output depends heavily on the quality of the input.
AI can generate impressive results, but it needs the right context to understand the actual problem. Without meaningful information, it may only create a general explanation instead of something useful for a real Business Central scenario. Instead of providing screenshots or manually explaining every detail, I started working with structured information generated from Business Central page scripting in YAML format.This gave the AI agent a clearer understanding of the actual functionality behind the page.
The information included details such as:
- Page structure and information
- Fields and their properties
- Available actions
- Controls and user interactions
- Navigation flow
The goal was not simply to ask AI to write documentation.
The goal was to help the AI understand what the functionality represents, how users interact with it, and how that knowledge should be converted into meaningful business documentation. This approach changed the role of AI from a text generator into a more context-aware assistant.
From Technical Details to Consultant Thinking
The most interesting challenge was teaching the agent to think beyond technical metadata.
A field name, page action, or control definition alone does not explain the real purpose behind the functionality. The agent needed to understand the business meaning behind these technical details and how they connect with the user’s daily activities. Instead of only identifying what exists on a page, the agent needed to understand why a field exists, when a user would interact with it, which business process it supports, and what should happen after a user performs a specific action.
This is where instruction design became a critical part of the solution. The agent needed guidance on how a Business Central consultant approaches documentation. A consultant does not simply describe screens and fields; they explain the purpose, the process, and the expected outcome.
The focus was not just generating technical descriptions. The goal was to create documentation that helps users, consultants, and testers understand how the functionality works in a real business scenario. This changed the role of the AI agent from a simple content generator into a business-aware assistant that can help convert technical information into practical knowledge.
Creating Multiple Documents from One Source
One of the most valuable parts of this approach was realizing that the same source information could be transformed into multiple useful outputs.
A single Business Central functionality can require different types of documentation depending on the audience. A functional consultant may need a document that explains the business process, while an end user may need a simple user manual that describes daily activities. At the same time, the testing team may need UAT scenarios to validate whether the functionality works as expected, and training teams may require material to help users adopt the solution.
Traditionally, creating each of these documents separately requires consultants to review the same information multiple times and rewrite it in different formats.
With this AI approach, the same structured knowledge can be transformed into different document types based on the requirement. The goal is not only to automate documentation creation, but also to reduce repetitive effort and allow consultants to spend more time on analysis, solution improvements, and customer discussions.
My Biggest Learning
The biggest learning from this experiment was that AI becomes much more valuable when it is connected to real workflows and practical business scenarios. The goal is not simply to make AI generate more content. The real objective is to help AI understand the purpose behind the information and transform that knowledge into something useful for people who need it.
A good AI assistant should not replace the consultant who brings business understanding and implementation experience. Instead, it should support the consultant by reducing repetitive effort, organizing information, and helping convert existing knowledge into meaningful outcomes faster.
I believe documentation is one of the areas where AI can create immediate value in Business Central projects. The challenge is not that documentation cannot be created manually. The challenge is that the same information often needs to be transformed into multiple formats for different users, teams, and project stages.
With the right context, well-designed instructions, and a structured workflow, AI can become a practical assistant for Business Central consultants and development teams.
This experiment is another step in my journey of exploring how AI can support the way we implement, maintain, and continuously improve Business Central solutions.
The future is not about replacing human expertise. It is about combining human experience with AI capabilities to work more effectively.
Leave a Reply