In tandem with the RSA Conference, Bricklayer AI, the first Automated AI Security Team for Security Operations Centers (SOCs), unveiled three groundbreaking new features designed to both infuse “humanlike” intelligence into their AI Agents and expedite security outcomes for our customers.
Long term memory is a human-like concept where the agents learn from doing more work. Onboarding helps customers deploy agents in a timely manner, just like they would onboard a new analyst. Dynamic procedures are designed to allow customers to move from a desired security outcome to an actual implementation with as little effort as possible.
Bricklayer AI defines intelligence the same way you would evaluate if a human was good at their job: their knowledge of the industry, your specific business, and a collection of experiences that shape the context they use to make decisions.
These new features deepen Bricklayer’s ability to understand, learn from, and make decisions for their specific organization, bringing AI Agents to a new level of intelligence to work even more closely with their human counterparts.
Provide Business Context with Guided Onboarding
Just as you would onboard new members to your team, Bricklayer has released a new, streamlined onboarding process for Bricklayer Agents.
The tailored onboarding process begins by describing your organization’s security goals. From there, Bricklayer will list suitable procedures, suggest which AI agents are needed and walk through the configuration of the tools required to reach the desired security outcomes. The entire experience is dynamically guided by your organization’s priorities—so Bricklayer agents configure themselves to not just operate, but to operate in alignment with your strategy.
By capturing this context early, Bricklayer enables its AI agents to immediately work in sync with your environment, reducing manual setup and accelerating time to value.
Build Institutional Knowledge that Never Leaves with Long Term Memory
Bricklayer has also introduced a powerful way for their AI agents to remember, reuse, and learn from their work across workflows.
As AI agents execute a procedure, whether it’s triaging an alert, enriching a vulnerability, or responding to an incident, they can now store key outputs and context into a shared long term memory store. This includes final datasets and intermediate findings, contextual decisions and rationale, and explanations of tools used, outcomes, and confidence levels.
Once captured, this memory becomes accessible across your entire Bricklayer environment. Any future procedure can include a task to recall relevant memories, enabling agents to avoid redundancies, reference historical insights, and adapt based on prior outcomes. Bricklayer now operates as a continuously learning system, where AI Agents don’t start from scratch—they start from experience.
Plan Collaboratively with AI Agents through Dynamic Procedures
At Bricklayer, we believe that the AI agent with the deepest domain knowledge should lead the planning process, not a centralized orchestrator or static playbook. Each agent brings its own specialized perspective, and they collaboratively decide and delegate tasks dynamically based on who’s best suited to contribute.
Today, we are announcing Dynamic Procedures, a patent-pending, collaborative planning system where AI agents co-develop and refine operational procedures customized to your business and environment.
With Dynamic Procedures, SOCs can move from describing a set of security goals in plain language to having Bricklayer Agents build a custom designed security workflow, based on knowledge of your business and external threats, in mere minutes.
This creates a more intelligent, collaborative system where agents don’t just complete work—they own and optimize it. The result is workflows that are adaptive, explainable, and constantly improving.
Want to learn more? Contact [email protected] to schedule a meeting at RSA, or schedule a demo for after the conference.
Request a Demo →
