AI as a management tool

Forecasting, scoring, document analysis and generative systems may influence investment, credit, pricing, human resources and compliance. Once AI enters management processes, it becomes part of the company’s organisational architecture.

Article 2086 of the Italian Civil Code makes adequate organisational arrangements central; technology adoption must therefore be governed consistently with the size and risks of the business.

The board and information

Decision quality depends on information quality. Delegations and information flows under Article 2381 should enable directors to understand when AI output is material and what its limitations are.

A board should not turn an algorithm into an unquestionable authority.

Liability and business judgment

The business judgment rule protects the sphere of entrepreneurial discretion, not the absence of a diligent decision-making process. Where AI assists a decision, tool selection, data reliability, available expertise, controls and documentation become relevant.

Responsibility does not disappear behind technical complexity.

Vendors, conflicts and shadow AI

External platforms require contractual governance of data, security, availability, ownership of outputs and liability. Conflicts embedded in data or models also require attention.

Shadow AI — tools used without authorisation or oversight — may create confidentiality and governance risks greater than their operational benefit.

Risk-proportionate governance

Internal policies, tool registers, authorisation levels, output verification and human intervention thresholds should reflect the significance of the decision.

AI may assist managerial power; it should not become an ungoverned space in which it is impossible to identify who decided and why.