FyreWall AI

Inspiration

Security teams face thousands of alerts daily, making it difficult to identify and respond to critical threats quickly. We wanted to build a smarter cybersecurity solution that uses Agentic AI to investigate threats, reduce manual effort, and enable faster, safer responses.

What it does

FyreWall AI is an AI powered security platform that monitors security alerts, identifies potential attacks, and plans appropriate responses. It simulates the impact of remediation actions before execution, supports human approval for high risk decisions, and provides rollback capabilities to help prevent unintended disruptions.

How we built it

We designed FyreWall AI around Agentic AI workflows, local Large Language Models , backend services, and an interactive frontend. The proposed system correlates security events, maps threats to the MITRE ATT&CK framework, generates remediation playbooks, and visualizes network topology through interactive 3D threat graphs.

Challenges we ran into

Our key challenges include correlating large volumes of security events, enabling reliable AI-driven threat investigation, simulating the potential impact of defensive actions, and balancing automation with safety. Preventing false positives and avoiding disruptions to legitimate services are also important considerations.

Accomplishments that we're proud of

We developed a solution concept that combines autonomous threat investigation, pre-action simulation, human oversight, and rollback capabilities in one platform. We are particularly proud of our approach to making automated cybersecurity responses more controlled, transparent, and safer.

What we learned

We explored how Agentic AI can support cybersecurity workflows, how security events can be correlated to investigate threats, and how frameworks such as MITRE ATT&CK can structure threat analysis. We also gained insights into the importance of simulation, human oversight, and safeguards in autonomous systems.

What's next for FyreWall AI

We aim to improve threat detection and investigation, enhance the accuracy of AI-generated remediation plans, and develop a more realistic simulation sandbox. Future improvements include stronger rollback mechanisms, richer 3D threat visualization, and expanded security posture and compliance reporting.