RSAC 2025: A Data Security Inflection Point – Key Takeaways from Bedrock’s CEO

RSA 2025 Key Insights: Addressing Data Risk in the Age of AI

Our presence at RSA 2025 crystallized a fundamental understanding: the exponential growth of enterprise data across diverse environments is critically straining the capabilities of established data discovery and classification tools. This was not just a hypothesis but validated over 75+ conversations with customers, prospects and analysts, our recent survey data, and through the feedback on the talks and panels our team gave. All this data consistently pointed to enterprises struggling with visibility and context for data spanning multiple clouds, applications, and AI models, including new enterprise copilots. Legacy tools lack the necessary scale, manual efforts to correct inaccuracies are unsustainable, and emerging AI further complicates governance.

Key Challenges Voiced at RSA 2025

Our discussions consistently highlighted several urgent enterprise needs:

1. Scalable Data Discovery: Traditional tools cannot keep pace with data proliferation. Enterprises require continuous, comprehensive visibility across cloud-scale environments.

2. Cross-Functional Data Security: Data protection is now a recognized shared responsibility across security, IT, compliance, and data governance teams. Siloed efforts are failing.

3. Actionable Insights over Alert Fatigue: Significant frustration exists with legacy DSPM tools that generate excessive, low-context alerts, increasing manual investigation burdens rather than providing clear solutions.

4. AI Governance & Copilot Risk Management: The rapid adoption of AI, especially enterprise copilots, introduces substantial risks of sensitive data exposure. Secure deployment necessitates a deep understanding of existing sensitive data and user/AI access patterns.

How Bedrock Security's Approach Resonated

Bedrock's platform and its foundational Metadata Lake architecture were recognized as directly addressing these pressing concerns:

• Actionable Context, Not Just Alerts: The Bedrock Metadata Lake resonated deeply by offering a solution to alert fatigue. It consolidates comprehensive data context (sensitivity, entitlements, lineage, usage) into a unified repository. This enables "agentic investigation"—the system itself correlates findings and helps answer critical follow-up questions, drastically reducing manual effort and providing clear, actionable insights.

• Enabling Secure AI & Copilot Deployment: Bedrock enables AI governance for enterprise copilots by delivering deep data context through its Metadata Lake. Its context-aware classification distinguishes sensitive data from other content, while full entitlement analysis maps who can access what data. This empowers organizations to implement precise controls—such as copilot training exclusions, entitlement pruning (e.g. least privilege), and data deletion—for safer AI adoption.

• Innovations Validating the Approach: Key Bedrock capabilities highlighted this differentiated approach:

  • Model Context Protocol (MCP) Server: Enables AI agents to securely access the Metadata Lake's rich context for governed AI operations. 

  • API-First Architecture: Facilitates seamless integration, embedding Bedrock's intelligence into the broader security ecosystem. 

  • Push-Down Tags: Enforces consistent governance by propagating classifications to source systems.

  • Bedrock Copilot for Security: Allows natural language queries for rapid, actionable data security insights from the Metadata Lake. 

Industry recognition, including the Publisher’s Choice in Data Security and DSPM at the Global InfoSec Awards, underscores that Bedrock's focus on data-centric, scalable, and actionable security meets critical enterprise needs.

Conclusion: RSA 2025 confirmed that modern data security demands scalable discovery, cross-functional collaboration, context-rich integrated architectures, and AI-driven insights that lead to decisive action.