Fighting Knowledge Drain through AI and Automation
Studies suggest that around half a million people will retire - undoubtedly well-deserved - every year since 2020. What to do when experienced colleagues retire and take their valuable knowledge with them? Considering that every third employee in Germany comes from the baby boomers of the 50s and 60s, it doesn't take much imagination to visualize the dramatic impact their departure will have on the knowledge available in companies.
"You can't teach or buy experience, you can only bring it with you" is the generally accepted credo in many boardrooms. This experience is therefore irretrievably lost when experienced specialists leave the company after 30 years or more. In concrete terms, this means that younger employees no longer have anyone they can ask when a problem arises: "How did you always solve this?" Or: "What does this system message mean and what do I have to do?"
So what can be done to preserve the valuable knowledge held in the heads of these highly competent and, above all, experienced employees for their successors - and for the company as a whole? Can modern technology, including artificial intelligence, perhaps help here?
Scenarios in which the loss of knowledge is often particularly virulent can be found in the IT sector, among others. A typical example is self-developed software solutions that have been expanded and adapted to changing needs over the years. However, these solutions are often inadequately documented or, in extreme cases, not documented at all. They therefore represent a classic "knowledge silo" - a "black box" that works, but nobody can say exactly how, let alone develop it further. When their developers retire, it is not uncommon for the only viable path in these cases to lead to standard solutions - some might even say back to standard solutions. And indeed, many software providers report customers who have practically no choice but to replace their systems, which have worked well for years, with standard systems, even if these may not offer all the refinements they have come to appreciate. In the example of in-house developed software, the simplest method is probably to "force" the developers to document their software in detail in good time so that their successors can continue to use and, above all, maintain it. Although this is a possible approach, it probably falls short. It would mean that less experienced colleagues would have to laboriously pore over the documentation for every problem in order to find a solution. Wouldn't it be more efficient if the knowledge were preserved in another "easily digestible" form?
Documenting systems and software is certainly one way of saving the knowledge of "veteran" employees from disappearing. However, it would make much more sense to process this knowledge in such a way that it can also be used directly by less experienced or specialized personnel. This could be, for example, the addition of suitable context in order to obtain more meaningful alarm or status messages. The next stage would be specific instructions that tell less technically trained users what to do in a particular case. In both cases, artificial, learning intelligence can provide valuable support, as it is precisely capable of recognizing correlations and generating corresponding action statements. It thus forms the link between the experience of older colleagues and the support of their successors.
And if a system is able to issue context-based instructions, recommendations and even specific instructions for action, the step towards automation is not far away. This is because these instructions can often be executed either without any human interaction or directly in the respective system after verification and confirmation by the user. For example, the AI could automatically decide to shut down a threatened system or carry out a specific update, taking several conditions into account.
This is where software and system providers are called upon: On the one hand, standard solutions are needed to replace the "home-made" - and usually poorly documented - programs used by many companies. But above all, smart solutions are urgently needed: Solutions that can record and access the existing knowledge in companies. Ideally, they should also be able to place the knowledge in the right context using AI and automate as many procedures as possible. In the IT sector, AIOPs platforms, i.e. platforms that can automate and secure the operation of entire system landscapes using artificial intelligence, are ideal for this.
"The situation is there," said the first German Chancellor, Konrad Adenauer. And that is exactly how it is today: The valuable knowledge of long-standing, experienced employees must now be saved from being lost. Apart from the option of retaining these employees for longer, the only option is to transfer this wealth of knowledge into intelligent systems.