How Does Artificial Intelligence Change the Business Models of the Financial Industry? (Part 2)

In order to position itself in the FinTech market with an AI-based application, a company needs a deep understanding of the AI FinTech market, i.e. both the existing applications and the business models that have developed around these applications. In the first part of this blog post, we used the AI Application Taxonomy, a tool to describe the properties of an AI application in more detail, to analyze a total of 79 AI applications from 75 FinTechs and to identify four overarching application archetypes: (1) Alphanumeric Prediction, (2) Interaction-Based Recommendation, (3) Image-Based Categorization and (4) Environmental Data-Based Reaction (Figure 1).

Figure 1: Archetypes of AI applications in the FinTech sector

Analyzing the archetypes of AI-FinTech business models

In order to not only perform an analysis of the AI-based applications, the FinTech business model taxonomy of Eickhoff et al. (2017) in Figure 2 was applied to the 75 FinTechs that operate the applications.

Figure 2: FinTech business model taxonomy according to Eickhoff et al. (2017)

The FinTech business model taxonomy is closely based on the business model understanding of Osterwalder et al (2005), some elements of the taxonomy can also be found in the well-known Business Model Canvas. By means of a statistical cluster analysis seven AI-FinTech business model archetypes could be derived from the data, which are briefly described below and arranged along the vertical in figure 3:

Figure 3: AI-based strategic positioning options for financial institutions

Archetype (I) Personal Assistance is an app- or web-based personal assistant, which usually either addresses financial topics at the customer interface (e.g. planning customer interaction with ABAKA) or provides support in handling classic banking tasks (e.g. accounting processes through Candis). Archetype (II) is also an old acquaintance in the financial industry, offering B2B and B2C customers support in making investment decisions. Archetype (III) Open Banking, on the other hand, is based on a new, more open paradigm and is in many places a result of government regulation, such as in the European Union, where the PSD2 regulation requires the financial industry to exchange data with each other. Nevertheless, it offers many new opportunities for strategic positioning. This enables AI FinTechs such as or FabricAI to integrate the customer data they need to provide their B2B accounting services automatically via an API interface. Conversely, in the new open banking world, financial institutions can also integrate external services into their portfolios via APIs, such as the regulatory services of Apiax. In addition, financial institutions can also obtain external services via B2B appstores such as those of Family Finances (ff) and offer them to their end customers, or offer their own services to other providers for integration or even establish themselves as B2B appstores themselves, which is in line with the strategy of “ff”, for example. The AI-FinTech business model archetype (IV) Fraud Detection is used for user identification (e.g. using the voice analysis of the AI FinTech Spitch) and for the verification of payments (e.g. using the live face recognition of the AI-FinTech saffe) and in further variations beyond that, while the archetype (V) Decentralized, Intelligent Intermediation has so far largely been used for working capital financing (e.g. Finturi). The special feature of this business model archetype is that here AI is usually combined with distributed ledger technology (DLT), also known as blockchain technology, as in the case of the crypto-asset platform HeapX. The archetype (VI) Co-Creation represents cross-company business models that are based on the consolidation of data, such as the insurance ecosystem around’s Rideshur platform, which offers fleet insurance based on the driving behaviour of all drivers. The archetype (VII) Trading on the other hand allows AI-based and therefore convenient trading with different asset classes. This is primarily based on pattern and trend recognition, an example being win f(x).

Combination of the application and business model perspectives

Combining the two perspectives considered above in a matrix as in Figure 3, it is easy to see which AI applications enable which types of business models. The observable combinations give an overview of and orientation on the current FinTech market. The strategic management could now examine which of the observable positioning possibilities are already exhausted because many suppliers are active in this segment, or whether it might be worthwhile to enter. It is also conceivable that a larger financial institution could enter into a partnership with a FinTech or acquire it on the basis of a particularly good strategic positioning.

New strategic positioning opportunities for the financial industry

Beyond the business model perspective, the management can also carry out an IT strategic analysis and ask itself whether the development of new applications, e.g. within archetypes (3) and (4), is worthwhile, since obviously hardly any providers are active in this area and accordingly, few business models exist. First of all, one must question whether some of these white spots are not unoccupied for good reason, as they simply do not make sense or only address a niche market. In any case, after a process of elimination one can ultimately identify those combinations where potential for innovative strategic positioning arises due to the use of AI-based applications (Figure 4).

Figure 4: Innovative, AI-based positioning options for financial institutions

For example, the combination of (4) Environmental Data-Based Reaction and (II) Transaction-Based Decision Support offers potential for strategic positioning, which is supported by use cases in the field of sensor-based investments. In concrete terms, this use case means that more and more different types of data are available – and due to the Internet of Things (IoT) more and more sensor data – which describe the environment more and more completely. These new data can be used for decision making in the investment area. For example, early attention to the COVID-19 infection figures in China at the beginning of 2020 would have been a warning signal for all investors, and the movement data of construction machinery could be used to evaluate the efficiency of the construction company in question.

The combination of (2) Interaction-Based Recommendation and (III) Open Banking highlights the strategic positioning opportunities beyond the API-based information aggregation and payment services driven by the PSD2 regulation: The information collected by open banking services allows conclusions to be drawn about interactions, on the basis of which a wide range of recommendations could be generated in the ecosystem that go far beyond the banking industry. These could be recommendations for insurance or mobility services, for example.

Innovative strategic positioning options are not only made possible by the use of artificial intelligence alone, but also by the combination of AI and DLT/blockchain. The combination of both technologies promises a particularly high innovation potential, as their capabilities are to a certain extent complementary to each other: While AI can use artificial neural networks to recognize new patterns and relationships in data, the results of self-learning systems in particular are difficult to trace, which virtually prohibits the processing of sensitive data that could lead to discrimination by the AI. The DLT, on the other hand, can exploit its advantages here as an unchangeable infrastructure and document the AI’s decision making process, thus increasing confidence in the analysis results of AI-based applications. The decentralized nature of the DLT also allows the data of many different actors to be combined in a data protection-compliant and forgery-proof manner, which in turn is advantageous for AI models. In combination (2) and (V), marketplaces can be created where AI-based recommendations can be traded – again, of course, via DLT. In addition, the conclusion of standardised financial products could be automated based on defined interactions in the DLT network. A similar scenario is conceivable in combination (4) and (V), e.g. for corporate financing based on environmental data. While interactions arise from financial transactions or communication with other market participants, environmental data is generated by sensors in the environment or in machines. The potential for convergence of the two technologies could go beyond the improvement and automation of trade and financing transactions and promote the emergence of new services, business models and ecosystems (Dietzmann et al., 2020).


Dietzmann, C., Heines, R., & Alt, R. (2020). The Convergence of Distributed Ledger Technology and Artificial Intelligence: An end-to-end Reference Lending Process for Financial Services. Proceedings of the 28th European Conference on Information Systems (ECIS).

Osterwalder, A., Pigneur, Y. & Tucci, C. L. (2005). Clarifying business models: Origins, present, and future of the concept. Communications of the association for Information Systems, 16(1), 1.

Eickhoff, M., Muntermann, J. & Weinrich, T. (2017). What do FinTechs actually do? A Taxonomy of FinTech Business Models. In Y. J. Kim, R. Agarwal & J. K. Lee (Hrsg.). ICIS 2017 Proceedings. 22. Abgerufen von

Christian Dietzmann

Was sind Deine Erfahrungen mit dem Thema? (Kommentieren geht auch ohne Anmeldung oder Einloggen; einfach kommentieren, auf Freigabe warten und fertig!)