On what resources do management strategists need to focus, if they want to leverage Artificial Intelligence and create Competitive Advantage? To provide an answer from business academia, a framework for leveraging Artificial Intelligence in a business context is presented in this 5-part article series. In Part 2, the relevance of data, or even Big Data, is elaborated.
For the application of AI, the first and very basic raw material is data. ML especially, is by definition (inductive) learning from data by machines, thus inevitably depends on data. But also, AI in general needs data to be of use. For example, expert systems need data in form of formalized facts and rules in a knowledge database, in order to use an inference engine to deduct new conclusions.
Even in reinforcement learning, an artificial agent needs data or information in the form of feedback, to learn when certain actions should be rewarded or punished. Note that this data need not necessarily come from the real-world. In simple environments like chess or Go, the rules of the game can be easily formalized. Hence, the AI is able to learn from simulations of games, for example against another instance of itself.
However, for most real-world problems, the environment in which the AI acts and the “rules of the game” are very complex, in fact too complex to formalize. Consequently, you need real-world data to catch this complexity, in the hope that machines can detect and generalize the inherent rules and structures.
In business, data is indisputably one reason why AI and ML are successful and popular. Data is the new oil. is a famous phrase (first coined by Clive Humby in 2006), widely used to emphasize the immense value of information extracted out of data. Especially the concept of Big Data triggered a literal hype in the business world and related research fields. But what is Big Data actually? Several definitions exist, often simply listing vague characteristics of the data. Here, the consensual definition, proposed by Mauro et al. (2015) is used:
“Big Data represents the Information assets characterized by such a High Volume, Velocity and Variety to require specific Technology and Analytical Methods for its transformation into Value”.
The three “V’s”, namely Volume, Velocity and Variety, are the most common characteristics stated in definitions of Big Data. Volume reflects the challenging large size of the datasets for storing and processing. Velocity represents the high speed at which the data is generated, collected and analyzed. Finally, Variety refers to the unstructured nature of Big Data. Besides these three “V’s”, the stated definition also specifies the concept of Big Data by emphasizing the significant technological and analytical requirements, and by highlighting the economic impact through generating insights.
This transformation from Big Data to Value is exactly the purpose of BDA and can be conceptualized in six steps with the DIKW hierarchy (Lamba & Dubey, 2015) and the Information Value Chain (Abbasi et al., 2016):
The presented conceptual transformation can be applied for several purposes in practice. It can support human decision-making by revealing hidden problems, patterns or variability, automate processes for more efficiency, or discover new needs and create innovations like new business models, products and services (Manyika et al., 2011).
Two socio-technical features of Big Data also influence this value creation: portability represents the possibility of reusing existing data in another context (Günther et al., 2017). On the other side, interconnectivity refers to the possibility of creating synergy effects by aggregating multiple, heterogeneous data sources, to extract more valuable insights from their combination. Another chance to create more value is simply to take more data. Junqué de Fortuny et al. (2013) showed that even for already large datasets, increasing the sample size to a massive scale can make predictive modeling substantially more accurate. Note that bigger can be better – but need not be.
Consequently, one can state:
“Access to Big Data, and the ability to handle it can potentially result in Competitive Advantage.”
This fact challenges the theory of RBV, since data itself is not rare and thus does not meet the VRIN criteria (Braganza et al., 2017). Nevertheless, it has been shown that Big Data capabilities has significant positive effects on market and operational performance (Gupta & George, 2016). Research showed that “the more systematic the analysis, utilization and management of Big Data, the more extensive the utilization of Big Data in the organization”, and “the higher the top management understanding of the importance of Big Data use, the higher the contribution of Big Data to organizational competitive advantage” (Kamioka & Tapanainen, 2014).
Ransbotham and Kiron (2017) claim that due to this potential for CA, ownership of valuable data is altering power relationships within industries and even within companies, which is why data governance will be an increasingly important organizational capability. They especially encourage data sharing or trading between organizations, even between competitors, since it still yields win-win-situations.
However, data trading can be especially difficult since data as information asset is naturally nonrivalrous and only partially exclusive, which means it can be used by multiple parties simultaneously and intellectual property rights may be defined but only incompletely enforced (Pantelis & Aija, 2013). Finding a fair price is already hard, because as experience goods, it is hard to justify the value of the data to a potential buyer without revealing the intrinsic information, but once it is revealed, the value is significantly reduced because of nonrivalry and partial exclusivity.
Additionally, data is often expensive (financially and time-wise) to collect, but nearly cheap to copy or disseminate. All in all, these are the reasons why data privacy, security and ownership remain major challenges for data governance and management in general (Sivarajah et al., 2017). But despite the intimidating challenges with Big Data, Brynjolfsson & McAfee (2017) point out that for realizing significant performance improvements with analytics, you may not need all that much data to start with. Also Ross et al. (2013) suggest, to rather use “little” data more effectively throughout the organization, instead of solely focusing on Big Data.
In the third part of this 5-part series, the relevance of IT infrastructure is elaborated. Stay tuned!
Keesiu Wong is Co-Founder & CEO of Design AI, a start-up focusing on agile AI development and use case identification through Design Thinking. He is a trained Data Scientist with an academic background in Mathematics, Management and Data Engineering at the Technical University of Munich. In addition to 5+ years of experience with AI projects, he has entrepreneurial experience in 4 start-ups, as well as experience in Design Thinking, top management consulting and as a start-up coach at UnternehmerTUM.Weitere Beiträge
A Strategic Framework for Applying Artificial Intelligence in Business - Part 5