INITIATIVE · AI · FINOPS · JIT PAM · LEADERSHIP
An enterprise FinOps prototype that combined multiple AI agents with just-in-time privileged access to help users obtain cost and operational insight securely and automatically.
I initiated and led this prototype while leading the sales engineering team at a Cloud PAM startup. The work was designed for an enterprise customer and helped my team demonstrate how secure access, AI orchestration, and FinOps could become a practical customer solution.
The initiative focused on increasing employee productivity, improving the help provided to field and partner teams, and turning a technically differentiated prototype into customer value that supported a successful enterprise opportunity.
FinOps users need timely answers about cloud cost and usage, but investigation often requires multiple systems, specialist knowledge, and privileged credentials that should not remain active.
A supervisor agent receives the request, understands the intent, and delegates to specialized compliance and FinOps agents. JIT PAM checks out short-lived credentials only for the approved task, then the workflow returns an explainable answer to the user.
The supervisor coordinates the agents; compliance constrains the action; the FinOps specialist gathers the relevant information; and the response layer presents the result with enough context for a human to review it.
AI handled request interpretation, delegation, and explanation. AgenticOps supplied the operating model: scoped access, approval boundaries, evidence capture, failure handling, and a clear human escalation path.
This initiative connected technical excellence to measurable operating outcomes: less manual investigation, faster enablement for employees and field partners, safer access to enterprise data, and a stronger customer story for the sales engineering team.
Customer names, credentials, account identifiers, and implementation-specific secrets are intentionally excluded from this public overview.