From Source Code to Customer Training Videos: An AI-Powered Approach

Turning Application Knowledge into Practical Training Resources

Software applications often include complex workflows that require clear instructions to help users get the most out of their features. As applications evolve, keeping customer training materials accurate and up to date can become a time-consuming task.

Visus explored an AI-powered approach to streamline this process for a client operating a SaaS platform that supports religious congregations with membership management, pastoral events, tasks, follow-ups, and administrative workflows.

The goal was to demonstrate how AI could transform existing application code into customer-facing training videos, reducing the manual effort required to create instructional content.

The Challenge: Creating Accurate Training Materials Efficiently

The client's application included multiple workflows that required clear, practical instructions for end users. Creating training videos manually presented a challenge, particularly as new features were introduced and existing functionality changed.

Producing an effective tutorial required more than explaining how a feature worked. The process also involved understanding the application's underlying business logic, identifying the correct sequence of user actions, capturing the actual interface, and presenting the information in a format that non-technical users could easily follow.

The application also consisted of separate frontend and backend projects, with business logic distributed across both. Understanding the complete workflow required examining these components together rather than relying on a single source of documentation.

Visus set out to explore whether AI tools could connect these technical details with visual demonstrations to create accurate, accessible training materials more efficiently.

The Strategy: Combining AI-Powered Documentation and Video Creation

Visus developed a proof of concept using three AI-powered tools: OpenWiki, Claude, and Claude Motion. Each tool supported a different stage of the process, from understanding the application to producing a sample training video.

Generating Application Documentation with OpenWiki

The first step involved using OpenWiki to generate documentation directly from the application's source code.

This documentation helped establish a clearer understanding of the application's features, workflows, and underlying business logic. By examining the codebase, the team could identify the steps involved in specific processes and use that information to guide the tutorial.

This approach connected the application's technical implementation with the practical instructions needed by end users.

Capturing the User Workflow with Claude

With the documentation available, Claude navigated the application in the development (DEV) environment and captured screenshots of the "Create Event" workflow.

These screenshots provided visual examples of the actual application interface and the sequence of actions required to complete the task.

Combining code-generated documentation with screenshots helped bridge the gap between how the application worked behind the scenes and how customers interacted with it.

Producing a Training Video with Claude Motion

The final stage involved using Claude Motion to combine the screenshots and instructional content into a sample training video.

The resulting video demonstrated the "Create Event" process in a format designed to help users understand and follow the workflow.

By connecting documentation generation, application walkthroughs, and video production, the proof of concept demonstrated a potential workflow for turning existing software knowledge into practical customer training resources.

The Results: Validating an AI-Assisted Training Workflow

The proof of concept successfully produced a sample training video demonstrating the application's "Create Event" process.

The result validated the feasibility of combining AI-generated documentation, automated application navigation, screenshot capture, and AI-assisted video production within a single workflow.

The approach also demonstrated how existing source code could serve as a starting point for developing customer-facing instructional materials, rather than requiring every tutorial to be created entirely through manual documentation and recording.

Although the initial proof of concept did not measure time savings or production-scale efficiency, it established a foundation for further exploration. With additional validation and refinement, the approach could help organizations accelerate tutorial development, reduce repetitive manual work, and maintain training materials as their applications evolve.

Looking Ahead: Making Customer Training Easier to Maintain

As software applications continue to grow in complexity, organizations need practical ways to help customers understand new features and complete important tasks.

AI-assisted documentation and video creation offer one potential path forward. By using source code to inform instructions and application screenshots to demonstrate workflows, teams can explore ways to make training content easier to develop and update.

The next step is to evaluate the approach across additional workflows, verify the accuracy of generated instructions and demonstrations, and measure the effort required compared with traditional video production.

For Visus, this proof of concept illustrates how AI can help connect technical knowledge with real-world user needs. With further development, similar workflows could support more efficient customer education and help organizations keep their training resources aligned with their evolving software applications.