Augmented decision-making (and leadership): how AI is changing the C-suite

After an initial phase of unstoppable 'hype', artificial intelligence is increasingly quietly but concretely entering the decision-making processes of companies. No longer as an individual productivity tool, but as a strategic lever capable of influencing - and in some cases challenging - the choices of theC-suite. Less proclamation and more real use, less isolated experimentation, more integration in the processes that really count. Starting with decision-making processes. This is the picture that emerges from a recent report by the Capgemini Research Institute, "How AI is quietly reshaping executive decisions", based on a survey of 500 C-level executives (including 100 CEOs) from companies with revenues in excess of USD 10 billion. The assumption of this study, which we will analyse in detail in this article, may be as follows: when it comes to artificial intelligence, we are still living in a transitional phase, in which the technology is already widely used and generates measurable value, but at the same time raises several questions about governance, accountability and the human-machine relationship.

More than half of the interviewed CXOs (Chief X Officers, the full C-suite) already use AI to support or guide high-impact strategic decisions and this percentage is destined to grow rapidly: if 38% of the executives confirm a structured and integrated use of the technology in decision-making processes (almost a third are still experimenting), within the next one to three years the rate of 'active adoption' is set to rise to 78%.

Going into how algorithms and generative models are deployed in the enterprise, the report tells us that at present, operational activities such as drafting emails and minutes or researching and analysing information are prevalent; in the next three years, it will no longer be just a question of operational efficiency or individual productivity because the role of AI will change radically, entering into the heart of leadership and supporting CEOs and CXOs in defining strategies, assessing scenarios and managing increasingly complex and uncertain contexts. In fact, managers plan to exploit it as a support tool to simulate scenarios, stress decision hypotheses, identify hidden patterns in data and anticipate market trends. A step that marks, at least on paper, the entry of artificial intelligence into the flow of executive decision-making, were it not for the fact that the conditions for its mature adoption are far from established.

The data collected in the Capgemini Research Institute study therefore tell of a rapid evolution, but far from any idea of total delegation to technology. Only one per cent of CXOs, in fact, consider it plausible that chatbots and agents will be able to take certain strategic decisions autonomously in the near future; the model that is emerging is rather that (now well known) of the human in the loop, and therefore of an AI operating as a support co-pilot, capable of broadening the perimeter of analysis and assessing the various viable options, while leaving the final choice to the leader in the flesh. An approach, the latter, that reflects not only prudence from a technological point of view (AI is not infallible), but also awareness of the legal, ethical and reputational responsibilities that accompany top management decisions. It is equally true, however, that from being a manager of 'collateral' activities such as preparing materials for executive committees or reading large volumes of data, AI will take on the role of cognitive partner of the C-suite in a relatively short time (one to three years in fact). As of now, 59 per cent of CXOs report a significant reduction in the time and costs involved in decision-making processes, 47 per cent point to an improvement in the quality of the choices made, while more than half of the executives note benefits in terms of efficiency, creativity and predictive ability, confirming that AI is beginning to affect not only the speed, but also the depth of choices.

Capgemini's report also reiterates a concept that has emerged several times in recent months and specifically tells us that 67% of CXOs indicate clearer governance and accountability as the essential condition for being able to integrate AI into decision-making processes with confidence. On the other hand, only 34% of the companies surveyed state that they have already implemented adequate frameworks in this area, highlighting a structural gap between potential and reality that is also reflected in terms of transparency. More than 70% of leaders, in fact, avoid publicly declaring the use of AI for strategic decisions, mainly due to fears linked to the reputational risk in the event of errors, legal implications and the difficulty of explaining choices influenced by algorithmic models. By contrast, only 11 per cent say they do so (or consider doing so) without reservation, and 'only' 41 per cent of CXOs declare an above-average level of confidence in using AI to support executive decisions. Of concern to the C-suite members are data security, the quality of basic information, the 'explainability' of models (the techniques that make the decision flows of AI and machine learning systems understandable to humans) and the absence of structured processes to validate the recommendations generated by the algorithms. All factors that always bring the issue of managerial responsibility back to the centre.