While working at PlexTrac, I designed and built an internal prototype demonstrating how modern AI tooling could interact directly with customer security data using PlexTrac’s existing public APIs.

The goal was simple:
show how quickly a secure, useful AI integration could be stood up using the tools customers already have access to.

This prototype uses Anthropic’s Model Context Protocol (MCP) to allow Claude to query PlexTrac data in real time and assist with reporting, analysis, and executive communication — without introducing new proprietary dependencies or requiring special customer access.

What This Prototype Demonstrates

• Natural language querying of PlexTrac clients, reports, and findings

• AI-assisted report summaries, trend analysis, and remediation roadmaps

• Executive-ready outputs generated directly from live data

• Secure authentication using existing PlexTrac credentials and APIs

• Clear separation between prototype logic and production systems

The prototype was intentionally designed as a proof-of-concept, not a production feature — focused on clarity, speed, and capability exploration rather than scale or hardening.

My Role

• Designed the MCP integration architecture

• Built the initial server and tool definitions

• Defined practical, real-world use cases for customers and internal teams

• Created onboarding and internal documentation for non-AI specialists

• Transitioned the prototype to PlexTrac engineering for further development

This work has since been handed off to the PlexTrac engineering team and is being evaluated for future productization.

Why This Matters

This project reflects how I approach innovation as a Sales Engineer and aspiring Field CTO:

• Start with real customer workflows

• Use existing platform capabilities first

• Prototype quickly to reduce ambiguity

• Communicate value through working examples, not slides

It also reinforces a core belief of mine: AI becomes truly valuable when it meets practitioners where they already work.

Availability

This prototype is not publicly available and is shared here only as an example of internal innovation and technical exploration completed in my role at PlexTrac. No customer data or production customer environments were used in developing or demonstrating this prototype. This work was created in the course of my employment with PlexTrac and is owned by PlexTrac. Any views expressed here reflect my personal perspective on product innovation and do not represent official product commitments or roadmaps.