Understanding AI
For corporate leaders, AI is not an IT matter. It is a human one.
OCNUS approaches AI from the perspective of the people who use it and the leaders who are responsible for how it is used. This is not about technology. It is about capability, governance and strategy.
The OCNUS Taxonomy
Not all AI is the same. Governing it as though it were is a governance failure.
OCNUS has developed an AI governance framework and taxonomy that gives boards and executives a precise and practical language for understanding what their organisations are actually using and governing. The taxonomy classifies AI systems into three categories, defined by the degree to which the system acts autonomously.
Instrumental AI
Processes input and produces output. The system never acts. A human independently evaluates it and decides what, if anything, to do with it.
Decisional AI
Output constitutes or directly drives a decision, with or without nominal human review. Stakes are organisational: accountability for the outcome sits with the organisation, whatever the personal consequences to the people affected by it.
Agentic AI
Decides and acts autonomously in the world. Executes tasks, triggers systems or controls processes without requiring human approval at each step.
Generative
Produces novel, non-predetermined outputs. Raises distinct risks around accuracy, originality and intellectual property.
Deterministic
Outputs are rule-bound and reproducible. More auditable, but may embed bias structurally through its logic.
The taxonomy gives boards and executive teams the precise language they need to understand what their organisations are using, put in place sound AI governance for each category and make confident decisions about where AI can deliver the most value.
Executive AI Coaching
The most powerful AI tools in the world are already available to you. OCNUS will help you use them.
Senior leaders do not need to become AI engineers. They need to become sophisticated users of the tools that are already available to them — and to understand how to configure those tools for the specific, recurring demands of their role. That is what OCNUS coaching delivers.
Most executives are using AI tools as they come out of the box, if they are using them at all. That is the first level of capability, and it delivers real value. But the major platforms — ChatGPT, Claude, Gemini and Microsoft Copilot — all now allow users to create customised AI assistants built around specific tasks and contexts: custom GPTs, Projects, Gems and Copilot agents. These require no coding, no technical knowledge and no IT involvement. They require only the understanding of how to build them and the judgement to know what to build them for. That is the second level of capability, and very few senior leaders have yet developed it.
The use cases for a senior leader are specific and immediate. Developing and stress-testing strategy. Synthesising large volumes of information quickly and reliably. Preparing for board meetings and interrogating board papers. Modelling scenarios and examining options. Navigating complex HR challenges. Drafting and refining high-stakes communications. Supporting the professional development of teams. Each of these can be transformed by a well-configured AI assistant built around the leader's own context, priorities and ways of working.
Every engagement begins with a clear understanding of the leader's role, their priorities and the tasks that consume the most time and demand the most judgement. The coaching then builds capability directly in that context — not through a programme designed for someone else's challenges, but through sustained, practical work on the leader's own.
You work directly with Dr Rob Leach. The pace is yours. The application is immediate. The capability you build is yours to keep and develop.
AI Governance
Three risk threads. Four organisational levels. One coherent framework.
AI risk falls into three categories: Internal Risk, External Risk and Bad Actor Risk. Each is real and distinct. OCNUS specialises in Internal Risk — the risk that arises from how an organisation's own people use AI. It runs across three distinct threads and four organisational levels. The OCNUS governance framework maps all of them, giving boards and executive teams the precise understanding they need to govern AI use responsibly, discharge their directors' duties with confidence and close the AI governance gap before it becomes a source of director liability.
Active Risk arises from ungoverned AI use that is already happening throughout most organisations. Staff are choosing tools without guidance, handling sensitive data without oversight and relying on AI outputs without the critical evaluation needed to catch errors. The governance response is not restriction. It is the policy, training and oversight that allows AI to be used safely and productively — and that protects the organisation when questions of liability arise.
Passive Risk is the risk of unrealised capability. Organisations that do not invest in AI capability across their workforce lose ground to those that do. The gap compounds over time and becomes harder to close the longer it is left unaddressed. Boards that govern AI well create the conditions in which that capability can be developed safely, consistently and to the organisation's lasting advantage.
Skill and Judgement Erosion is the third thread, running through both of the above. Consistent delegation of cognitive work to AI gradually erodes the human capabilities that practice develops and maintains — without any visible signal that anything is wrong. Governance that understands this risk creates the conditions for AI use that preserves and develops human capability rather than quietly replacing it.
The three threads run across every level of the organisation. At the operational level, the risk is primarily about errors and data exposure. At the managerial level, it is about decision quality and accountability. At the executive level, it is about strategic judgement and the culture the executive creates, deliberately or by default. At the board level, the question is whether the board can perform its governance function at all where AI is concerned.
Risk does not stay at the level where it originates. It travels upward, shaping decisions at every level above. Sound board AI governance requires a framework that covers all four levels. That is what the OCNUS framework provides.