From the quarterly report to continuous intelligence: the new corporate governance
AI in the boardroom makes it possible to detect risks, connect information and strengthen continuous oversight of companies, with traceability and human control.
Georgina NuñezDoctorado en Economía | Partner Estratégico Gobierno Corporativo @Yunt
Column by Georgina Núñez, originally published in Spanish in Altavoz (Peru) on 11 November 2026.
For decades, boards have governed with periodic snapshots of the company: quarterly financial statements, management presentations, risk maps and scheduled audits. That model is still indispensable, but it is no longer enough for organizations exposed to regulatory changes, cyberattacks, reputational crises, logistics disruptions and automated decisions that evolve between one board meeting and the next.
Artificial intelligence opens a possibility that goes beyond summarizing documents: moving from periodic oversight to continuous governance intelligence. An AI collaborator such as YUNT (YUNT.ai), connected only to authorized sources, can watch relevant signals, relate them to the strategy and the risk appetite, and raise material matters before they reach the next board information package. This collaborator does not seek to replace the director's judgment, but to widen their field of vision.
The need is concrete. In 2026, PwC reported in its Annual Corporate Directors Survey that 71% identify AI as the main capability their boards need to strengthen; 82% see opportunities to improve the materials the board uses to make decisions, and 81% believe they could challenge management more effectively on strategic matters.
The problem is not a lack of information
Boards receive enormous volumes of data, but almost always after someone has selected it, ranked it and turned it into a narrative. That mediation is legitimate, since informing is management's responsibility; however, it also creates an asymmetry. The board knows the company through a version prepared by those who execute the strategy that the board itself must evaluate.
A well-designed digital collaborator such as YUNT (YUNT.ai) can contribute a second analytical perspective. Faced with a significant investment, for example, it could contrast the growth hypotheses with market movements, new rules, competitor behaviour and geopolitical risks; point out which assumptions explain most of the expected return; and show what changes under alternative scenarios. The aim is not to turn AI into a competitor of management, but to improve the quality of constructive disagreement and support decision-making.
PwC noted that 35% of directors already reported some integration of AI into oversight activities (PwC 2026). Its emerging uses include digesting voluminous materials, discovering patterns, comparing with peers and supporting scenarios. The decisive gain is not delivering more pages, but helping to ask better questions and reducing the information gap between the board and management.
The next step is to connect that capability with risk management. A company with Latin American operations can face, at the same time, a new tax, a sector investigation, community protests, accusations against a supplier, cyberattacks and logistics strain. Each isolated signal may seem minor; together they can materially change the business's exposure to higher levels of risk. Continuous intelligence serves precisely to recognize connections that separate reports tend to hide or omit.
A collaborator, not an oracle
It is important to distinguish an AI collaborator from a generic chatbot. The former has a defined role, limited permissions and a manual for how to act. It accesses authorized internal documents and verifiable external sources; monitors changes; compares new evidence with previous decisions, controls and commitments; generates scenarios and questions and prepares reports; records its sources; and follows up on agreements, owners and deadlines.
Tools such as YUNT.ai illustrate this approach: collaborators integrated into the company's systems and channels that carry out tasks according to defined processes, connect scattered information and preserve auditable evidence. For a board, the initial case should not mean broad autonomy, but a bounded function: preparing, for example, a monthly report on material changes in regulatory, reputational, cyber and third-party risks, with links to the evidence and the matters that require human validation.
The recommendation, therefore, is to adopt AI collaborators, not to buy a black box. The board must demand five conditions:
identifiable sources
access and confidentiality controls
a clear separation between facts, inferences and uncertainties
escalation thresholds
human approval for every relevant decision.
It must also know what data is missing, what assumptions the system uses and when it was last reviewed.
How to start without losing control
The starting point is not the technology, but a concrete information gap. Once the problem is identified, the company must order and prioritize its sources, and define data owners, permissions, retention, quality and traceability. Without that base, AI will only automate the disorder.
Run a pilot with clear scope and metrics. A collaborator could compare, each month, the new signals with the risk register and the strategic hypotheses in force. Its usefulness should be measured by the reduction in the time needed to detect material matters, the quality of the questions generated, the proportion of findings validated and the timeliness of the follow-up, not by the number of alerts. More signals can mean more noise.
The framework published in 2026 by INSEAD and KPMG offers pertinent guidance: strategic oversight to create long-term value; active vigilance over technology and security; workforce transformation with human responsibility; trustworthy AI; and review of the board's own work. Its central idea is correct: trust, transparency and accountability do not slow innovation down; on the contrary, they scale it (INSEAD, KPMG 2026).
The real change will not be that directors read less, but that they spend less time reconstructing what happened and more time examining alternatives, tensions and future impacts. A decision may have been reasonable when it was made and stop being so when its assumptions change. An AI collaborator can watch that distance between the original decision and the current reality.
Governance will remain human, because responsibility, experience, prudence and judgment cannot be delegated. But it does not have to remain intermittent. Boards that combine reliable data, traceable AI collaborators and human review will be able to move from receiving information about the company to building, continuously, a better-informed understanding of it. AI must not, and does not seek to, reduce the board's responsibility to understand the business; it must help the board understand it in time.
Georgina Núñez is an economist and a former economic affairs officer at the United Nations Economic Commission for Latin America and the Caribbean (ECLAC). She is a member of the Governing Board of the Centre for Regional Studies of the University of Tarapacá, Chile (CEUTA), and is certified in corporate governance by the Corporate Governance Institute (United Kingdom).
Sources
PwC. (2026). 2026 corporate governance trends: five priorities for directors.
PwC. (2026). 2026 Annual Corporate Directors Survey.
Gan, K. C. J., Wilczek-Stronczek, B. and Papasava, A. (2026). Governing the algorithm: a conceptual review of strategic AI oversight in global corporations. AI & Society.
Tipurić, D., Hruška, D. and Kovač, I. (2026). Board-level AI literacy as a missing governance capability: a pilot study of boards and executives. Frontiers in Artificial Intelligence.
INSEAD Corporate Governance Centre and KPMG International. (2026). AI Governance Principles for Boards.
YUNT. (2026). Digital collaborators with AI for companies. https://www.yunt.ai/en/riesgos-dinamicos
