The customer
A health-tech company.
The problem
How do you put the NIST AI RMF into practice when you cannot see your AI?
Before AIBound, the health-tech company governed AI the way most enterprises do: a policy, a governance committee and a list of approved applications. Then it looked at what was actually running.
AI sat inside SaaS platforms, browser extensions, developer co-pilots, local agents, autonomous workflows and internal models that teams adopted on their own. A policy document alone gave it paperwork, not protection.
The solution
What they did
The company rolled out AIBound with three products, AISeeker, AgentHarness and AIAssure, so each NIST AI RMF function became an operating process.
- 1 AISeeker discovered AI continuously, which produced the inventory that the Map function needs.
- 2 For each system, AIBound recorded the owner, the data it could reach and the identities using it, and judged its risk in context. AIAssure kept the governance decisions and the audit trail.
- 3 When a system exceeded a risk threshold, AgentHarness enforced policy through the existing EDR integration: it blocked access, stopped interaction with sensitive systems and started remediation, with no manual ticket.
The results
What does NIST AI RMF look like as a continuous program?
The company now runs the NIST AI RMF as a living program. AIBound found more than 180 AI systems, including SaaS platforms, browser extensions, developer co-pilots, local agents and internal models, none visible to its existing security tools. Each has an owner, a risk view and a recorded decision.
180+
AI systems found and mapped
4
NIST AI RMF functions covered: Govern, Map, Measure and Manage
Continuous
monitoring, with a full audit trail
Source: AIBound deployment at a health-tech company, as published on aibound.com/case-studies. Checked 15 November 2026.