Within just 6 weeks, the system demonstrated above-benchmark early-agent performance: 25% of autonomous actions were executed correctly.
Meet our client
Client:
Industry:
Market:
Technology:
Client’s Challenge
A leading SASE platform needed a unified AI agent solution that could accelerate onboarding, improve knowledge retrieval, and autonomously perform actions across multiple disconnected systems, and ensure support responses were consistently generated in a standardized format without hallucinations.
Our Solution
We built a Self-Improving AI system with a controllable level of autonomy – ranging from full human-in-the-loop oversight to fully automated operation (Claude and Agentic Skills + LangGraph for orchestration + AWS Bedrock Claude models). The solution fuses all existing knowledge sources and support systems into a consistent set of atomic skills, each designed to return outputs in a standardized format. The entire platform was securely deployed in a cloud environment within a tunneled VPC.
Client’s Benefits
Within just 6 weeks, the system demonstrated above-benchmark early-agent performance: 25% of autonomous actions were executed correctly. 60% of response, crafting atomic skill outputs were rated as good. The output format has also been especially well received by the support team.
AI Agents: Lessons Learned in the Field
What engineering teams
learned while building
and scaling AI agents
