Check Point / Lakera – AI Runtime Security Discovery Workshop

The AI Runtime Security Discovery Workshop is a professional services engagement designed to help organizations understand and address the runtime security risks associated with modern AI applications. As enterprises adopt Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and agent-based systems, new attack surfaces emerge across prompt handling, data pipelines, APIs, and integrated tools. This engagement provides a structured approach to identifying and prioritizing these risks without requiring changes to production systems.

The workshop evaluates AI applications deployed on AWS environments, including integrations with services such as Amazon Bedrock, AWS Lambda, Amazon API Gateway, and data sources used within RAG pipelines. ControlPlane works with customer teams to map architecture, data flows, and trust boundaries, and to identify runtime threats including prompt injection, data leakage, hallucination risks, and tool misuse. Existing controls are assessed, and gaps in protection are identified.

Lakera capabilities are then mapped to the identified risks, defining how runtime protection can be integrated into the application architecture. By the end of the engagement, customers receive a prioritized risk assessment, control gap analysis, and a defined pilot plan with success criteria. This enables organizations to move from uncertainty to a structured, actionable approach for securing AI applications at runtime.