Building an AI-Powered Risk Assessment Platform at KiteCyber

During my internship at KiteCyber, I built the frontend for an AI-powered cybersecurity risk assessment platform that would become part of their customer acquisition strategy. The idea was simple but powerful: help businesses understand their current security posture through an interactive assessment, then guide them toward solutions that could actually improve their protection.

The platform I developed starts by asking companies basic questions about their setup — industry, company size, types of devices they use, whether they have remote workers. But it gets more sophisticated from there. The questions adapt based on previous answers, so a healthcare company would get different prompts than a tech startup, and a fully remote team would face different scenarios than an office-based business.

My biggest technical challenge was managing the dynamic flow between frontend and backend through API calls. The AI engine on the backend would analyze responses in real time and determine what questions to ask next, which meant my interface had to handle constantly changing content while maintaining a smooth user experience. I built error handling for network issues, loading states for AI processing delays, and local storage to preserve progress if something went wrong.

Getting the user experience right was crucial because we were asking people to share sensitive information about their security practices. The interface had to feel trustworthy and professional while remaining approachable. We discovered through early user testing that people felt overwhelmed when they saw too many questions at once, so I redesigned the flow to present them in digestible sections with clear progress indicators.

The real magic happened at the end. After users completed the assessment, the AI would generate a personalized report showing their risk level across different areas — endpoint security, data protection, network vulnerabilities, compliance gaps. Instead of generic advice, the system would provide specific recommendations based on their exact situation. Sometimes that meant suggesting KiteCyber's platform, but often it included other tools or practices that would address their particular weak spots.

What made this project especially interesting was seeing how design decisions affected completion rates. When we made the questions feel more conversational and added context about why we were asking, more people finished the assessment. When we showed immediate feedback after each section instead of waiting until the end, users stayed more engaged.

The platform became a valuable lead generation tool for KiteCyber's sales team, but more importantly, it actually helped businesses identify real security gaps they hadn't considered. Working on something that combined AI capabilities with practical cybersecurity guidance taught me that the best frontend work doesn't just look good — it helps people make better decisions about protecting what matters to them.