Make AI part of the business. Keep it governed.

01

Select the platform, open up its use.

I contributed to selecting an AI agent-building platform, managing the RFP and contracting. I also coordinated contributions to the AI charter and helped write it.

The platform quickly gained traction: business teams created more than 2,000 agents within months. That adoption exposed a need for governance at scale, reinforced by the EU AI Act. AIR was conceived to meet that need.

02

Build a shared framework.

I started with the first complete diagrams, dimensions and rule sets, then brought Digital, IT, Compliance, Security, Legal and Regulatory teams around those artifacts. The group tested and changed them until a common model was validated.

The core of the work is a shared framework: business, Legal, Compliance and Information Security rules become one centralized, versioned and evolving rule set. The same foundation can be reused as skills or applied by a dedicated harness such as AIR. Reviews and decisions improve the framework, and those changes become available to the systems that use it.

The framework assesses any AI system. In AIR, the LLM structures the facts; deterministic rules produce the assessment and preserve its rationale.

03

Focus the oversight.

I estimate that a manual agent review would take at least four hours of combined work across Compliance, Legal, Information Security and digital or business teams. For 2,000 agents, that is roughly 8,000 hours. This is an estimate of workload, not measured time savings.

AIR assesses the portfolio of more than 2,000 systems in about ten minutes, including the applied rules, rationale and obligations. Teams then check quality through sample reviews and investigate cases that need attention. Under the active framework, 96–97% of the portfolio follows the green path.

Experts can focus on the few cases that require deeper control and review samples periodically. Questionnaires are shorter, duplicate requests disappear and the time saved compounds across every function previously involved.

04

Give teams the tools to build.

I train and coach employees so they can build their own projects with AI. Depending on their needs, this covers advanced prompting, agents, reusable skills, MCP connections and development with Claude Code or Codex. I run group workshops and individual sessions, including for management, executive committee members, Legal and Compliance.

AIR is the dedicated harness for applying that framework across the portfolio. It imports systems from existing sources and keeps their configuration, assessment, triggered rules, obligations and framework version in one registry. Its exports provide the reference for the EU AI Act register, AI Gateway and governance tools.

I also designed agents, skills and processes to reuse company rules in business workflows, including investment, legal and contract analysis.

The standard assistant on the first agent platform was replicated across environments and is now its most-used agent. Backed by a shared skill library, it helps employees produce the prompt and operating guide for an agent, checks whether the use is permitted and adds specialized guardrails. An MCP extension is now being designed to connect it directly to the rules engine, registry and similar agents, skills, connectors and prior decisions.