Building a fully agentic operations platform using AWS with Strike48

Security and IT teams have long faced a costly visibility issue: traditional SIEM solutions can make storing and analyzing all logs prohibitively expensive. As a result, organizations often operate with significant visibility gaps, with some threats left unseen while analysts manage alert fatigue.

Businesses have been gradually adopting AI in security operations. But without visibility into the entire log surface area, security leaders can’t trust AI agents to act on the complete data picture. So some organizations keep humans in the loop to guard against unexpected behavior, slowing security processes down.

Strike48 saw an opportunity to address both challenges, applying its expertise in SIEM solutions and AWS services. To create and manage autonomous agents, workflows, and knowledge bases, the company built Prospector Studio on AWS. The agentic AI–powered platform serves as the backend orchestration layer that coordinates all Strike48 AI agents and connector applications.

A longtime user of AWS services, Devo Technology ran its legacy solution on Amazon Elastic Compute Cloud (Amazon EC2), which offers secure and resizable compute capacity. The company stored its log data in Amazon Simple Storage Service (Amazon S3), which provides industry-leading scalability, durability, availability, and security for virtually any amount of data.

For the core intelligence layer that fuels Prospector Studio, the Strike48 team chose Amazon Bedrock, which gives access to hundreds of foundation models from leading AI companies. “Using Amazon Bedrock, we can build and deploy powerful agents for our customers while protecting their data and—most importantly—earning their trust,” says Tim Leehealey, cofounder and vice president of strategy at Strike48.

Amazon Bedrock provides built-in guardrails and supports seamless switching between models. Built on this model-agnostic foundation, Prospector Studio can orchestrate agents across micro-agent workflows, dynamically matching the right model to each agent and task. The solution offers connectivity to over 312 third-party tools, including Grafana, through Model Context Protocol (MCP) integration. Powered by Amazon Bedrock reporting capabilities, Prospector Studio’s dashboards display token usage data, delivering full transparency into cost and performance across the entire agentic workflow.

The collaboration between the AWS and Strike48 teams relies on close, continuous communication, including regular meetings and joint problem-solving across storage, cost, and architecture. As the two teams test, refine, and iterate on Prospector Studio around customer feedback, they move forward together—confident, aligned, and committed to what comes next. “Working alongside AWS, we’ve always had the tool sets to overcome obstacles and continue on,” says Leehealey.

With its AI-driven, human-led log analysis solution, Strike48 is helping customers affordably achieve virtually 100 percent log coverage. Compared with traditional security systems, the company’s solution sped up time to remediation by over 90 percent. Customers now save about USD $3 for every USD $1 they spend on the solution. Along with Prospector Studio, Strike48’s log intelligence layer and agentic packages support natural language interactions, facilitating human oversight and helping reduce training burden for new users.

The company will expand beyond security by incorporating agentic AI into the whole IT organization, targeting network operations centers, observability, DevOps, compliance, and IT support functions. Continuing its day one pursuit of global scale, Strike48 is evaluating Amazon Bedrock AgentCore for building, connecting, and optimizing agents.

“With the support of the AWS team, we’ve been able to think strategically and operate globally from the beginning,” says Aaron Dun, vice president of marketing at Strike48.