John von Neumann conceived the notion of IT in June 1945, driven by the wartime necessity for intricate calculations - from ballistic trajectories to weather forecasting and atomic bomb development. This emphasis on data handling and complex computation has directed the evolution of the IT industry to date. The field has been characterized by three terms: code, data, and algorithms, which we simplify into data and programs (coded algorithms).

Data has been dubbed the new oil of this century, underscoring its economic influence on everything from industrial processes to consumer habits and political opinions. Like oil, data requires refining and curation and, through neural networks, enables significant achievements across multiple domains.

Another prevalent belief is the algorithmic nature of all things (Gartner: digital growth as algorithmic business). Adherents to this notion assert that society can be distilled, matrix-style, into an enormous collection of algorithms and APIs (connectivity between them using the Internet), capable of computing every facet of existence.

Data and programs form a closed ecosystem: data feeds programs and, conversely, programs produce data, all founded on a constantly evolving code language base. Currently, global corporate efforts in digitalization aim to increasingly govern all company activities via data and programs.

Yet, on closer inspection, this reductive view of the economic sphere appears overly simplistic. Consider your everyday work; it encompasses more than mere data manipulation or algorithm execution. Daily tasks involve exchanging information, exercising judgment, responding to unforeseen events, appraising scenarios, etc. You employ data and programs since they underpin your company's operations, but it is not true that they wholly define your work or your organization. We don’t take decisions based on data, but on information. Information is data with a purpose; information is data wrapped with intention and meaning or, in other words, information is data plus knowledge about data.

The only way to fully articulate the nature of your work is by introducing an additional concept: Knowledge. Knowledge is the indispensable third aspect necessary to explain a company's function: we employ data to record events, utilize programs to manipulate or analyze this data, and apply knowledge to steer the entire operation. Knowledge is complex and mutable, yet it forms the core of any human's interaction with the world.

When discussing the knowledge used in our daily tasks, we can differentiate between to two orientations. The first is the knowledge that is related with our work. For example, I know that a checking account with a zero balance cannot be overdrawn. This is a simple statement about a fact. Such facts, often based on precise data and programs, are integral to our daily operations.

Another form of knowledge is meta-knowledge, which involves understanding for what we know or should know. This type of knowledge prompts us to question established norms and to seek innovation. Our previous example was that overdrafts are not admitted. However, we can question: why not? Who decided that? It is because a legal rule? Is there any opportunity to change that rule to obtain a commercial advantage? What is the impact of this rule in the operations? etc. This simple example shows how meta-knowledge is the part of the human activity that challenges the comfort zone and innovates how our organizations are doing.

Knowledge is a slippery matter: it changes continuously, and its past is as important as its present because we need to know how it evolves. Sometimes it is precise and easy to communicate (e.g., a policy) and in other cases it is really fuzzy and difficult to express (e.g., how to design a system). Sometimes knowledge can be written in textual form, but in many cases, it must be depicted in the form of drawings, blueprints, or other alternative presentations. Part of the knowledge is concrete, concerning well-described facts, but most of it is abstract, describing concepts that are difficult to grasp and that often are subject to interpretation (e.g., the legal system).

Despite extensive research on knowledge management towards the end of the last century, its application in organizational enhancement has been minimal, possibly due to overreliance on the promise of artificial intelligence solving complex issues as if by magic.

Knowledge's recursive nature sets it apart from data and programs. We leverage knowledge to modify existing paradigms or to generate new ones, making it as our most potent tool for evolution and innovation. Should we have a dedicated Chief Knowledge Officer to oversee this domain?

Our industry heavily invests in the generation and processing of data through complex algorithms, yet we're just beginning to tap into the immense potential of knowledge. Data, programs, and knowledge comprise a dynamic trio that powers our organizations and societies. From this perspective we can visualize the whole building of IT as comprising three layers:

• The code layer: The foundation of the software industry is code itself - billions of lines that embody the software layer, including neural networks. This is the executional bedrock known to all.
• The model layer: Nowadays, we build systems that are not programmed but modeled. For instance, in relational databases, the 'data model' determines the behavior of the data engine: the data model is configured, not programmed. New technologies, referred to as 'low code,' also belong to the field of model-driven initiatives. In short, a model is structured information (technically referred to as ontologies) about the world that a platform can interpret and transform into the expected system. The practice of model-driven development allows us, today, to build the entire information systems of companies fully based on models, in a 'no-code' fashion.
• The knowledge layer: We now find another player in the field: Generative AI. This has been a significant surprise for all of us, not only due to its power but also due to the rapid pace of its development. In the few years since the publication of the foundational article on transformers ("Attention Is All You Need" by Ashish Vaswani et al., published in June 2017), an incredible new technology has been developed. We now envisage unexpected capabilities where knowledge can be leveraged as a computable asset for companies.

The model layer enables us to implement systems in a highly efficient and scalable manner, reducing or eliminating the need for coding. However, the knowledge layer is now changing two core aspects of the IT world. On one hand, the interaction between people and systems is becoming conversational. You can communicate with technology that understands your needs, creating a radically new user experience (which is unfavorable for chatbots). On the other hand, knowledge-driven engines can aid us in the development of code and models, even dynamically. Business requirements can be transformed into system behavior in real time.

The key to using GenAI in new applications (both on the business and user/citizen sides) is not their ability to generate text but to explore meanings and transform raw text into something knowledgeable and computable. This implies using combinations of model-driven solutions with GenAI-trained models to create new applications and experiences, thus fully exploiting the value of our lakes of knowledge.

All things considered, in the coming years, we will witness a profound transformation in the IT industry, technologies and competencies, as well as a multitude of new applications for information systems within companies, governments, and the public.