AI Strategy and AI Governance at the Board Level: Governing the Age of Cognitive Enterprises

The board's role is not to oversee AI implementation, but to determine how AI changes competitive advantage, risk profiles, organizational capabilities and corporate value creation.

At first sight, the strategy approach could be similar to how previous technology disruptions (like the advent of the internet) were factored into the strategy process. The exco would prepare recommendations after reviewing how AI affects the core value drivers (use cases driving revenue, productivity, or customer experience). Externalities such competitive and regulatory changes would be taken into account, threats of industry disruption would be evaluated, and defensive moats would be analysed. From the analysis would then emerge a set of programmes/initiatives to change the course of the enterprise, leveraging the potential of AI.

However, this time it’s different. AI is a cognition and intelligence disruption (it automates execution, reasoning, and judgment). AI changes how work is performed and decisions are made. The strategy process therefore needs to take three factors more prominently into account:

• Knowledge: What knowledge assets create advantage, powering AI?
• People: What should humans do, and what can AI do?
• Governance: How far can we responsibly delegate cognition, decisions and action to AI?

Knowledge Sovereignty

Strategic Question: Which information, data assets, and learning cycles must the company own or control to power AI? Do we control the knowledge and intelligence that can be derived from our data?

Data assets range from proprietary knowledge, research data, customer knowledge to operational knowledge. Which control mechanisms need to be in place to be able to use and protect these assets? These cover all aspects of ownership and rights. AI heavily relies on training to increase in accuracy. Who will guide and control these learning cycles?  

Board-level implications range from competitive differentiation, dependence on external AI providers, to sovereignty risks. The geographic location will also offer opportunities or challenges, given data residency constraints. The impact on the business will also depend on its activity: SaaS business might be forced to make more dramatic changes compared to local industrial activities.

The key conclusion is that knowledge sovereignty is a new topic that becomes a strategic asset, similar to intellectual property.

Human Capital Transformation

According to BCG (AI Transformation Is a Workforce Transformation | BCG), value comes less from the technology itself (around 30%) and far more from organizational adaptation and upskilling (around 70%). 

Strategic question: how should the workforce evolve in an AI-centric enterprise?

Traditionally, a large part of the organization was defining processes and solving specific issues. These are activities that can be done by AI Agents. Humans now need to step into higher-order roles: shaping strategy, setting intent, managing risk, and intervening when judgment, ethics, or accountability is needed. 

Work needs to be redesigned, considering Human-AI collaboration, Agent-assisted work and Agent-supervised workflows. New capabilities required include AI literacy, AI oversight, prompting and orchestration and risk judgement. In tandem, new roles such as agent managers, AI product owners and AI governance specialists will be created. 

Board-level implications are that workforce strategy, talent acquisition, upskilling investments and change management need to be addressed. AI transformation is primarily a workforce transformation.

Dynamic AI Governance

Strategic Question: How far can we responsibly delegate cognition, decision-making and action to AI?

Traditional Governance is static, rule-based, periodically reviewed and implemented by deterministic systems (e.g. ERP, HR). In contrast, AI Governance requires continuous monitoring, model evaluation, human oversight and risk-based controls in a probabilistic environment. Models can drift over time, Agents will take decisions autonomously, explainability will vary widely. As a result, Governance will need to become equally AI driven, with new Governance Agents auditing autonomously how policies and rules are adhered to by humans and AI agents. 

The Strategic Interdependence Between Strategy and Governance

The ability to govern AI determines the strategic options available. Poor governance limits achievable strategy and strong governance expands strategic possibilities.

Board Oversight Framework

In light of the opportunities offered by AI, we advise boards to lean in the following topics: 

1. Strategy

• Review AI investment portfolio: to what extend does it cover fundamental enablers like data platforms, compute infrastructure and operational changes for example?
• Review competitive implications: which core competency will be strengthened or weakened, 
• Approve ambitions for AI: increase revenue, achieve operational efficiencies

2. Risk

• Review AI risk appetite: which regulations apply, which operational aspects must be protected at all items against AI threats, 
• Review governance effectiveness: how are frameworks adapted for AI, how can rules be enforced, which area requires continuous governance
• Monitor regulatory exposure: to what extend is this done automatically, what is the approach to minimise risk

3. Capability Building

• Review scope of knowledge sovereignty:  for example: how are rights determined for data used with AI, is it clear which model will be enriched with usage and who will own the models, are there export control limitations
• Review readiness of human capital: what learning programmes are planned, how will the organisation evolve, which roles are becoming critical 

This is an article by the GUBERNA Sounding Board on Artificial Intelligence.