08.03.17

Expert Opinion

Artificial Intelligence at the heart of changes in companies

    Artificial Intelligence (AI), which is now distinguishing itself on a daily basis, is being swiftly implemented in most business sectors, shaking up in the process the human expertise existing within companies.

    The reality that is AI, together with its efficiency and versatility, are all undeniably demonstrated by the sheer variety of examples of applications using it. One can notably observe the growing importance of AI applications that classify, analyze, and imitate human genius. These same AI applications are poised to infiltrate every component part of companies so as to optimize how they function and their profitability.

    The associated transfers of expertise from the human being to the smart system represent challenges that company chiefs must address quickly. Those chiefs thus find themselves having to make new trade-offs. In shaking up the organizational model of companies, Artificial Intelligence is fundamentally changing the balances and driving forces in play. From now on, in considering the benefits that AI brings, company chiefs must view them along three major strategic lines: Performance, Organization, and Human Resources. None of these areas of focus can be considered individually without adapting the others or bringing them up to speed. Since they are interdependent, they have to be addressed as a whole, and in a way that takes account of economic and social constraints.

     

     

    AI is bringing about a revolution in the economic performance of companies

    Several studies undertaken in 2016 on the subject of AI's impact between now and 2035[1] assert that the economic growth of countries will no longer be assessed on the basis of their capital, but on their level of maturity in AI. This switch in reference criterion speaks volumes about the changes to come. In France, AI is set to increase national productivity by more that 20% between now and 2035 (this increase being 35% for American productivity).

    By looking at the current state of the art for AI's functional capabilities in terms of its new uses, we can draw up a non-exhaustive list of the potential benefits that AI can bring when incorporated into companies' information systems.

    AI can be (and is) used to produce, for example, more reliable and accurate performance indicators, and to implement systems for automatically recognizing products and determining production classifications. It additionally enables cost-controlled inventory management and planned flows (Order, delivery, transportation..) to be optimized. It is also turning out to be effective in detecting cases of fraud and accounting and financial anomalies; in identifying market crises and recoveries in advance; in assessing a customer's default risk etc... It can be used to study in-store behavior patterns, automate online customer support (interactive ChatBots), make personal product recommendations, set the sales prices of those products dynamically and optimize them…

    Its very broad scope of application is directly down to the Machine Learning and/or Deep Learning capabilities deployed on dedicated platforms.

     

    AI is causing new working and organizational methods to emerge

    Once it reaches maturity, Artificial Intelligence will change the very nature of work and human-machine relations.

    AI will take charge of all repetitive and codifiable tasks (by applying dedicated algorithms). In this way, it will shake up Fordist-type traditional models and develop new organizational structures within companies, causing the emergence of a growing collaborative human-machine relationship, like cobotics, or industrial robotics in the production field.

    Cobotics (or collaborative robotics) aims, for example, to produce robots that assist man, by automating part of his functions and work. In this way, it will reduce the arduousness of certain tasks by increasing the capability, autonomy, and overall performance of companies' employees.  Examples of this already abound, such as cobots installed on the assembly lines of car manufacturers, or the use of exoskeletons to provide muscular support to employees and thereby facilitate the performance of numerous tasks.

    In this way, AI will enable companies to become more flexible, less hierarchical, and more horizontal, notably by offering their employees greater autonomy and independence in their work.  The new jobs that will come with AI-induced change will essentially involve supervision and control functions, with managerial relations being completely reshaped throughout companies. Those companies will no doubt see new "profits" coming their way, such as a decline in absenteeism and accidents at work, to name but a few of them.

    In order to successfully integrate AI at the heart of its information system, a company will always need to invest in managing human talent and training employees. Conversely, talent management and human resources in the broad sense will gain in efficiency thanks to Artificial Intelligence. In the end, the central question concerns future changes in the job ratio:

    Number of new jobs created by AI/Number of jobs destroyed by AI

    Pessimistic forecasts are at odds with the most optimistic ones which predict replacement with no loss of current jobs. Actually, uncertainty, together with the lack of data and perspective on the subject, preclude any reliable forecast of this ratio.

     

    AI is redefining Human Resources policies

    AI is set to bring about fundamental changes in the way companies' employees are recruited, managed, and assessed. Companies are logically going to give preference to the profiles of candidates who are versatile, flexible, and open to changes, as compared to those with highly specialized, rigid expertise. AI is becoming a major player in the recruitment process, with results which, on that front too, are surpassing human judgment. A study published this year in the Harvard Business Review[2] asserts, with statistics to back it up, that AI remains 25 % more reliable than a human being in assessing the dossier of a job candidate. Led by a team of Canadian and American researchers, the study specifies that the more candidates there are for the same job, the more this difference in reliability increases in favour of AI. The human allows himself to be distracted by the details in dossiers and ends up producing a mental summary of the information which is imperfect, distorted by elements of cognitive bias and by too great a volume of information. 

    As a sourcing tool, AI is becoming more and more effective in processes to select candidates' dossiers. A solution like Riminder[3] manages to process 5,000 CVs effectively in 23 seconds! It provides HR departments with the combined powers of big data and deep learning to find the best profile for a given job, and vice versa. Riminder is a platform developed in partnership with the laboratories of Centrale Paris, the Ecole polytechnique, and the Ecole Normale Supérieure de Paris. By using this solution, it is possible to save considerable time in the selection process, notably by automating and optimizing administrative tasks linked to recruitment. The time saved can thus be devoted to the human aspect of assessing job applications.

    The rise in the power of Artificial Intelligence at the heart of companies and in the servicing of their needs will not happen without a fundamental rethink of their organization, decision-making chain, and culture. The transfers of expertise from humans to algorithms can be contentious if the performance gains are not sufficiently tangible. Any collaboration between man and the system requires at least a certain degree of consent at each level of the hierarchy and managerial staff. This consent is founded upon the trust that one places, or does not place, in AI which brings about changes to practices and jobs. Such changes pose questions for the decision-maker: how should one go about integrating the new distribution of tasks within the company and having it accepted, when it changes existing balances and eliminates certain jobs? How can performance, agility, and ethics be reconciled in the deployment of AI? The company chief has to take up this central question and make a trade-off between complex choices, and he must do so on good terms!

     

     

    [1] CIGREF 2016 Report (67 pages) "Artificial intelligence governance in enterprises": http://www.cigref.fr/wp/wp-content/uploads/2016/09/Gouvernance-IA-CIGREF-LEXING-2016.pdf

    [2] D.M. Klieger, N.R. Kuncel, D.S. Ones, "Algorithms that are stronger than instinct" – Harvard Business Review ) December – January 2015.

    http://www.hbrfrance.fr/magazine/2014/11/4824-les-algorithmes-plus-forts-que-linstinct/

    [3] Riminder: https://riminder.net/