Value Creation in AI Start-Ups — A Framework of AI-based Business Models

There are six types of AI business models through which start-ups are currently exploiting the commercial value of AI.

1. Introduction

Recent years have seen a dramatic increase in the adoption of artificial intelligence (AI) applications in businesses (Brynjolfsson & Mcafee, 2017). This adoption has been driven by technological breakthroughs as well as better AI products and services. However, it is only through a company’s business model (BM) that it creates commercial value with a technological innovation like AI (Chesbrough & Rosenbloom, 2002). Only through carefully creating BMs that service the customers’ needs best, companies will be able to harness the full commercial potential of AI. Because AI will permeate almost all aspects of human lives and the economy, I was curious to answer the following question:

“What types of business models do start-ups employ where AI is paramount to their core offering?”

2. AI Business Models

Figure 1. There are 11 key dimensions for evaluating AI business models.

To analyze AI-based BMs in a structured way, I developed a framework that builds on the nine high-level dimensions key resources, key activities, value proposition, tangibility, revenue streams, channels, customer engagement, customer segments, and corporate partner. These can be seen in Figure 1. My framework describes the BMs of AI start-ups structurally and in detail using multiple characteristics in each dimension. The developed framework can be used to better understand the necessary components to create value through BM innovation. It can also be expanded or reduced to represent the dynamic nature of BMs and the commercial possibilities of AI.

Clustering the survey results of 104 AI start-ups along this framework, I recurringly found the following six BM “blueprints”: Data source enablers, automator, analytics-as-a-service, insights-as-a-service, AI researchers, and mass-customizers. These “blueprints”, as shown in Figure 2, describe how the identified start-ups create value through their BMs.

Figure 2. The six types of AI business models are Data Source Enabler, Insights-as-a-Service, Automator, AI Researcher, Analytics-as-a-Service, and Mass-Customizer.

Data source enablers are companies that provide access to a new form of data that customers could not access before like images, video, or speech. This allows them and their customers to amass rich data sets promising even higher innovation and insights potential.

Automizers automate their customers’ processes. Companies in this category can be split up into two main streams. While physical automizers predominantly use machine perception and robotics to automate for example manufacturing or construction, virtual automizers use expert systems and machine learning to automate bureaucratic tasks like approving insurance claims.

Analytics-as-a-service companies on the other hand use customer data and machine learning (ML) to offer analytics tools and process optimization to their clients, for example by reducing waste or increasing sales performance.

Insights-as-a-service companies gather data from different public sources to derive insights about a certain market or global feature. They use natural language processing and ML to filter data and process unstructured data formats like images or text.

AI research companies are characterized by their high propensity to conduct research on ML and to innovate. They use highly skilled talent to research current problems of AI and solve them by developing better ML models or hardware optimized for ML.

Mass-customizers are companies that use AI to adapt a certain product exactly to the needs of their customers. This is done by using ML and natural language processing. They further enrich their customers’ data with public and social data.

3. Key Findings

In parallel to the development of the framework and the derivation of AI blueprints, five key findings were observed:

1. The created framework captures new important aspects of start-ups showing important distinctions in modern, technology-driven, but particularly AI-based, companies. Dimensions like tangibility, data sources, and technology sourcing need to be included in the evaluation of any modern business.
2. There are four entirely new BMs, that evolved from the business potential of AI: data source enablers, automizers, AI researchers, and insights-as-a-service. With AI, mass-customizers can create offerings that adapt to customers in every given moment and with an accuracy not possible without self-learning technology like ML. Analytics-as-a-service companies can, like insights-as-a-service companies, improve their analyses by using ML and machine perception in their data acquisition and processing efforts.
3. Data source enablers are found to be the largest group in the existing AI landscape (39 of 104 analyzed AI start-ups) and a novel BM type that came only into existing through AI.
4. Automizers need to be differentiated into virtual and physical automizers which have different requirements to their BMs. However, in both the virtual and physical space, they create value by dramatically reducing the work required by humans.
5. AI start-ups currently create little value for other customer segments than businesses. This is however expected to change. Additionally, an increasing number of AI start-ups offering solutions to governments will likely emerge.

Figure 3. Only 27% of AI start-ups target either consumers or governments.

4. Conclusion

While over 59% of businesses expect AI to have a large impact on their industry in the next years, only 25% of companies have already adopted AI (Ransbotham, Kiron, Gerbert, & Reeves, 2017). Before these companies embrace AI in their businesses, they must understand how they can create value from it and how their BM should be built. My research addresses the research question, what the BMs of AI companies are and how they create value. As a result, the six business models data source enabler, automizer, analytics-as-a-service, insights-as-a-service, AI researcher, and mass-customizer were derived. These blueprints form a crucial step towards a common understanding of the current landscape of AI start-ups. Only through the careful design of BMs companies create sufficient business value to successfully take AI to the market. With constantly emerging technologies and shorter development cycles, those businesses which understand the BM implications of AI best will identify new opportunities quicker and will work more successfully with companies which have yet to embrace their AI empowerment.

Please let me know what you think and feel free to contact me on LinkedIn.

References

Akter, S., Michael, K., Uddin, M. R., McCarthy, G., & Rahman, M. (2020). Transforming business using digital innovations: The application of AI, blockchain, cloud and data analytics. Annals of Operations Research. https://doi.org/10.1007/s10479-020-03620-w

Brynjolfsson, E., & Mcafee, A. (2017). What’s driving the machine learning explosion? Harvard Business Review. Retrieved from https://hbr.org/2017/07/whats-driving-the-machine-learning-explosion

Chesbrough, H., & Rosenbloom, R. S. (2002). The role of the business model in capturing value from innovation: Evidence from Xerox Corporation’s technology spin-off companies. Industrial and Corporate Change, 11(3), 529–555. https://doi.org/10.1093/icc/11.3.529

Gassmann, O., Frankenberger, K., & Csik, M. (2014). The business model navigator: 55 models that will revolutionise your business. Pearson UK.

Herting, A. M., & Schmidt, A. L. (2020). A systematic analysis of how practitioners articulate business models across disruptive industries. Technology Innovation Management Review, 10(10), 29–42. https://doi.org/10.22215/TIMREVIEW/1394

Nickerson, R. C., Varshney, U., & Muntermann, J. (2013). A method for taxonomy development and its application in information systems. European Journal of Information Systems, 22(3), 336–359. https://doi.org/10.1057/ejis.2012.26

Osterwalder, A., & Pigneur, Y. (2010). Business model generation: A handbook for visionaries, game changers, and challengers.

Reim, W., Åström, J., & Eriksson, O. (2020). Implementation of artificial intelligence (AI): A roadmap for business model innovation. AI, 1(2), 180–191. https://doi.org/10.3390/ai1020011

Yin, R. K. (2014). Case study research and applications: Design and methods (6th ed.). SAGE Publications. Retrieved from http://journals.sagepub.com/doi/10.1177/10963480970210010