Building the Cyber Risk Intelligence Layer: From AI Models to Actionable Security

Security teams are overwhelmed with data but still struggle to answer a fundamental question: what actually matters?

In this fireside chat, CyberSaint's Padraic O’Reilly and IBM's Srinivas Tummalapenta explore how cybersecurity is evolving from fragmented data collection to a unified cyber risk intelligence layer. They break down how layered AI architectures, combining NLP, GNNs, and LLMs, enable organizations to normalize massive volumes of security data and transform it into real-time, actionable insight.

The conversation dives into what it takes to connect telemetry, controls, and threat intelligence into a system that continuously prioritizes risk, supports decision-making, and aligns cybersecurity with business outcomes.

For more information about CyberSaint , please visit: https://securityweekly.com/cybersaintrsac.

Show Notes: https://securityweekly.com/rsac26-5

Timestamps:
00:00 - Introduction to Cyber Risk Intelligence Layer
01:45 - Shift from Point-in-Time to Continuous Risk
03:02 - Rapid Innovation with AI & 7-Week Sprint
03:53 - Where Security Architecture Breaks Down
05:57 - AI Consumption Models & Shared Responsibility
06:41 - Combining AI Techniques: Agents, NLP, GNNs, LLMs
09:11 - From Static Risk Assessments to Real-Time Insights
09:55 - Building a Layered AI Security Architecture
10:45 - Breaking Down Silos with Interoperability
11:33 - Normalizing Security Data Across Systems
13:51 - Using Data Efficiently for Risk & Compliance
15:33 - Mapping Telemetry, Controls & Risk Exposure
16:40 - Managing Global Regulations & AI Policies
18:12 - Risk Quantification & Business Context
20:03 - Inside the Cyber Risk Intelligence Layer
22:10 - Beyond SIEM & GRC: A New Security Model
24:06 - Outcome-Driven Security vs Tool Sprawl
25:08 - How Security Operations Will Change
26:59 - Autonomous vs Human-in-the-Loop Security
29:06 - Continuous Monitoring & Autonomous Controls

Видео Building the Cyber Risk Intelligence Layer: From AI Models to Actionable Security канала CyberRisk TV