Automating Software Documentation: A Multimodal AI Agent
Project description
Software evolves rapidly, yet maintaining up-to-date documentation for end-users remains a persistent manual bottleneck. This graduation project, conducted at Cape.io, investigates the application of Multimodal AI to address this challenge by acting as an intelligent production assistant.
The research focuses on the design and development of a "Documentation Agent" capable of streamlining the workflow from raw usage to publishable content. The project explores an automated workflow where, instead of manually recording and describing every step, an autonomous agent "watches" recorded videos of platform features to interpret user intent and context.
Context
AI in optimizing existing workflows
Results
The core objective is to determine how effectively an AI can process this visual input to populate a structured data repository. The prototype aims to isolate key user interactions and automatically collect the necessary assets—such as labeled screenshots and clips—storing them in a database to ensure they are readily retrievable for future documentation updates and platform integration.
By analyzing the transition from raw video to organized content, this project demonstrates a proof-of-concept for generating accurate, visually detailed documentation with significantly reduced manual effort.
About the project group
A graduate student at Fontys ICT is in the final phase of the bachelor programme and works independently on a graduation assignment rooted in professional practice. The assignment addresses a concrete ICT-related challenge and requires the integration of technical knowledge, research skills, and professional competencies.
During the final semester, the student analyses a real-world problem, develops and implements a substantiated solution, and reflects critically on both the process and the outcome. As part of the graduation moment, the student presents and demonstrates their work at Innovations Insight, explaining the relevance, approach, and results to a diverse audience of professionals, students, and teachers. This presentation forms an essential part of the assessment and demonstrates the student’s readiness to enter the ICT profession.