Peak Defence Leverages Generative AI to Transform Cybersecurity Audits

As an organization that provides cybersecurity management and compliance consulting services, Peak Defence needed a way to meet increasing demand for its services while ensuring the highest level of security and accuracy. By freeing up time spent on manual processes such as compliance and information security management, as well as operational tasks such as responding to requests for proposal (RFPs), they would be able to focus on doing what they do best: providing their clients with cutting-edge solutions.

Secondly, Peak Defence aimed to achieve operational excellence by minimizing the costs associated with the manual execution of compliance processes. Automating manual tasks would allow the company to unlock new capacity and efficiency gains, enabling it to scale and meet client needs faster. Ultimately, targeting these two key needs would allow Peak Defence to remain highly competitive in their industry while also accelerating growth. The company saw an opportunity to incorporate generative AI into its operations to achieve these goals. It aimed to leverage AI to provide timely insights, enhance security measures, and streamline compliance, all while maintaining strict privacy standards for sensitive customer data.

To tackle these challenges, Peak Defence enlisted Neurons Lab to build an AI-powered platform designed to automate compliance and security operations. It utilized Amazon Bedrock as the foundation to manage multiple large language models (LLMs), including Anthropic’s Claude for core AI processes. Extensive model experimentation and training were done using Amazon SageMaker, allowing the team to quickly develop prototypes.

One of the platform’s most innovative features was its ability to process large volumes of sensitive security data quickly. The platform employed a serverless architecture with AWS Lambda and AWS Step Functions to automate workflows and efficiently manage large datasets. This infrastructure enabled Peak Defence to dynamically scale operations without the need for additional infrastructure investments. Security and privacy were at the forefront of the solution, with strict controls ensuring that only authorized personnel could access sensitive data, and secure data management processes were implemented to handle increasing demand while maintaining flexibility in system scaling. By utilizing AI models, Peak Defence aimed to reduce the time required for data analysis and significantly improve decision-making processes related to security threats and compliance risks.

In terms of technical design, the system was centered around automating data processing workflows. Large datasets were efficiently fed into AI models, where each step in the process was dynamically managed and sequenced. This system ensured that relevant data was routed to the appropriate LLMs for analysis, with results passed back to the security team for action. Moreover, Peak Defence continually tested and adapted their AI models using the machine learning framework, giving them the ability to rapidly integrate new and emerging models. This provided flexibility, enabling the team to swap or chain models together for optimal performance depending on specific use cases.