How exactly Artifical Intelligence could help managers ? To answer this question, you need to understand the limits of AI before sorting out how you can use it. Artificial intelligence starts, of course, with digitisation. Most interactions, hence management ones, are now trackable because they are digitised (verbal communication, mails, chats, social networks, collaborative softwares). You need then people to curate these reports and put all of them in the same format: there is no artificial intelligence without information architecture. This can be a considerable amount of work, although it is necessary to produce clean data. Human language processing is still a hard thing to do for AI. Logically, the first limit here is privacy : will managers accept to be “scanned” (or “spied” if you take the worst possible perception) even if vendors claim that collected data is anonymised? The second limit is the kind of AI available for now and the years to come: kill your dream of singularity (the existence of a general AI or Human level AI)! We live in the world of narrow AI (bound to a specific function). The tasks that these machines can accomplish can’t be broad enough to replace the entire management mission. Narrow AI has thus serious limits on assessing all management situations. There is indeed no such thing as a good practice in management. It’s a matter of context in which the manager acts which in turn changes the context which then changes the manager’s decision which … you get the picture. Narrow Artificial intelligence has no common sense. It can rely on very rare cases to assess most common situations. In others words, narrow AI available on the market, solves relatively isolated problems, these are standalone solutions not able to assess the global picture we just mentioned.
Where are the limits?
Knowing these AI limits is not enough : you need to be aware of the managers and organisation’s limits too. Let’s start with the organisation. The advices and recommendations AI offers are data driven, improves over time and with more data. Most cognitive technologies have obviously statistics at their core, and if your organisation has not done anything with statistics and analytics, it’s probably not going to do much with AI either. Big companies have difficulties to integrate AI into their businesses. They have a set of legacy IT systems and well developed business processes that drive and inform their activities. Integrating AI based decisions into them is not easy. For example, a company has developed a new set of sales propensity models to identify the most likely buyers of particular products and services, integrating the recommendations into customer relationship management (CRM) systems and processes is not likely to be easy. So include it into the behaviour of salespeople can be a daunting task. Have a look now on managers themselves. Using AI efficiently means that managers understand cognitive technologies and how they work. At best, these managers are able to pair which technologies match with what types of tasks and the strength and limitation of each. It is easy to make mistakes if you don’t understand the trade-offs behind each technology. This managers’ ignorance can – and will – lead to mistrust. However, three different scenarios can be envisioned, depending on different kinds of contexts. Different solutions are warranted depending of these scenarios. The common thread here is that AI helps managers place themselves in the right scenario so that they can pair the right approach with the corresponding situation to achieve optimal outcomes.
Context is simple: full automation
In this option, context is simple. Interactions are linear. Automating some structured and repetitive management processes via robotic process automation could be possible when managers know what they are doing and have seen it a thousand times before. This is the domain of legal structures, standard operating procedures, practices that are proven to work. This means that there are rules in place, the situation is stable, and the relationship between cause and effect is clear: if you do X, expect Y. AI would establish (sense making) the facts, categorise, then respond by following the rule or applying best practice. For example, AI automates the onboarding process. Or, AI collects all of the work products from emails to files shared to chat messages. AI draws the connection between them. It gives managers tools for collecting, understanding, and acting on information.
Context is complicated: decision support system
In this option, context is complicated. The relationship between cause and effect requires analysis or expertise; there are a range of right answers. AI helps to assess the facts, analyse, and apply the appropriate operating practice. AI structures the manager’s analysis toward a decision, but does so by leaving to this manager refined judgment and expertise. Artificial intelligence considers the management question as a complicated problem and helps managers to look at every possible sequence of decisions. Not only AI helps managers to make sense of the context they are in so that they could make better decisions, but also AI helps them avoid the problems that arise when their preferred management style causes them to make mistakes. For example, using data still means using it selectively as it comes from a single managerial perspective. Under conditions of uncertainty which often characterises decision making, managers prefer to use both data and decision making processes with which they are confortable. Not surprisingly, for example, finances guys rely on financial indicators. AI enlarges the scope of data to consider and the global perspective as a result. This “augmentation” of managers could also be the result of AI training the managers: AI can access all the information regarding a particular situation and links it to relevant managerial knowledge (with the manager’s personal cloud learning): machine learning allows AI to learn to spot management topics by ingesting all the management patterns shown by human managers on limited numbers of topics (start small, for example setting goals, delegating tasks,…)
Context is complex : leaders in full control
In this option, leaders (and not managers) act in complex context. Complexity occurs when understanding parts does not review the whole complex web of facts, networks, and relationships. It involves large numbers of nonlinear interacting elements. Cause and effect can only be deduced in retrospect, and there are no right answers. Take a company level problem when leaders have to factor missions and strategy x leadership x organisation and culture x policies and procedures x management practices x structures in the same equation. These are systems that are impervious to a reductionist, take-it-apart-and-see-how-it-works approach, because your very actions change the situation in unpredictable ways. AI (and its backbone here, machine learning) can’t help leaders in any automated way because it doesn’t have the ability to turn the data into information and usable knowledge. Leaders are in full control. They can tap into AI resources to simplify the context from complex to complicated. AI helps to find instructive patterns if these leaders conduct experiments that are safe to fail. AI make them move from incremental reasoning (that is, always more of the same thing) to experiential reasoning (let’s try something different). AI helps here to formulate the strategy differently because it gives deviant – or creative ;)- minds the power to challenge the dearest taboos of the company by conceiving alternative documented scenarios, easing correlation research, establishing causalities.
If artificial intelligence takes over all of these tasks, managers and leaders have to add value to the human side of management. And expectations on that side will be high. They will need to have an understanding of analytics and data structures, a lifelong learning capability and a strong ability to communicate effectively the outcomes of machine activities to employees…
Comments