During development, we faced challenges with the accuracy of AI-driven recommendations for accessibility improvements. Some suggestions from the Groq LLMs were not contextually relevant or feasible. To address this, we implemented a feedback loop for developers to review and rate suggestions, allowing us to refine the model's output.

Additionally, ensuring compatibility with various web technologies required extensive testing and adjustments. By collaborating with our team and conducting user testing, we identified pain points and iteratively improved the tool's functionality, ultimately strengthening Web4All as a robust solution for enhancing web accessibility.