Knowledge Management of the Future: Merging Human and Artificial Intelligence

The phrase “knowledge is power” formulated by Francis Bacon in the 16th century has never played a greater role than in the digital age. After all, in today’s fast-moving era, knowledge that can be accessed at the right time determines whether the necessary business decisions can be made in a timely manner, whether customers remain loyal to the company because they can get the answers they need right here, or whether time and money are saved because, for example, machine parts can be replaced with the help of predictive maintenance before an expensive production shutdown occurs.

The same applies to other areas: Intelligent pattern recognition in healthcare can save lives or lead to new insights in research.

At the same time, in view of the rapidly growing mountains of data and increasing complexity, it has never been so difficult to acquire and apply precisely the knowledge that is needed in a particular situation. Human receptivity and time resources are known to be limited.

Machines and people

It therefore needs machine support, which on the one hand ensures that human experience is enriched with the missing information. On the other hand, intelligent systems should ensure that humans are freed from tedious routine work – from “monkey business” – so that they can better deploy their limited resources where they create value or significant differences to market competitors.

One of the prerequisites for this world, where human and artificial intelligence are combined, is to take the data we already have from the myriad silos where it is stored, put it into relationships with each other to generate information and, at the end of the day, knowledge – a process I call “Connecting the Dots.”

In view of the exponentially increasing mountains of data, this requires more than ever intelligent systems that have already reached a high degree of maturity and go by the names of “Enterprise Search,” “Insight Engine,” “Cognitive Search” or “AI Search.

Understanding knowledge management systems

The common denominator of these knowledge management systems is that they combine a variety of technologies under one roof: AI, machine learning, deep learning, natural language processing (NLP), natural language question answering (NLQA), and semantic content preparation that enable natural human-machine interaction.

The combination of all these technologies enables users to get exactly the information he or she needs at the right time and place to complete a specific task. This could be the address of a customer, the technical description of a component, the sales figures for a product in a particular region, or the indication for a particular drug.

One of the great advantages of enterprise search or insight engine solutions is that there is no need to break down data silos. The data stays where it is. Using so-called connectors, the smart systems retrieve all the required data – regardless of whether it is structured in a database or unstructured in mails and other documents – combine it and use it to form information that is needed at a specific time and for a specific task in a defined role.

To ensure that users are not overwhelmed by countless results as they would be with an ordinary search engine – i.e., that they only receive the information they really need – enterprise search and related solutions are now beginning to observe the behavior of users (assuming their consent, of course) and learn from it. Under the name “Behavioural Model for Information Retrieval System Design”, the following factors, among others, are analyzed: the role of the activity, actions taken in the past in connection with certain information, specific search behavior or even the emotions that users associate with information – a topic closely related to “Customer Experience” or “Experience Economy”. The goal, then, is to personalize the relevance of a piece of information. At the end of this journey is an intelligent assistance system with which the user can interact as naturally as possible and which provides him with the information he needs with pinpoint accuracy. This results in

These systems are constantly being developed further. One of the current trends in smart knowledge management is the addition of multimedia sources such as photos and videos, for example, to optimize the supply chain or support medical diagnosis. Gartner calls this “X Analytics.”

“Weak Supervision” as the Future

Another trend is “Weak Supervision” in AI: Until now, it took a lot of manual work and preparation to train an intelligent system. With Weak Supervision, a management system learns almost on its own, with performance continuously improving with use. At the same time, decisions made by the smart system become increasingly transparent – keyword “Explainable AI” (XAI).

Last but not least, forward-looking knowledge management systems can be used to optimize traditional business processes in the direction of flexibility and agility – the Corona crisis has clearly shown how essential these capabilities are.

Conclusion on knowledge management with machines

All the aspects mentioned point in the same direction: thanks to intelligent digital transformation, we now have the opportunity to combine our knowledge and experience with artificial intelligence. While the machine takes over tedious routines or makes needed information available anytime and anywhere, humans can play to their strengths such as social interaction, creativity and tact.

Read more: 5 steps to implement AI in your company

Daniel Fallmann

Daniel Fallmann beschäftigt sich seit frühester Jugend mit den Themen Künstliche Intelligenz, Machine Learning und Deep Learning. Im Jahr 2005 gründete er im Alter von 23 Jahren das Unternehmen Mindbreeze. Dieses zählt heute zu den führenden internationalen Anbietern im Bereich angewandte künstliche Intelligenz und Wissensmanagement mit tausenden Kunden weltweit.