Ecosystem Intelligence for AI-based Assistant Platforms

Abstract

Digital assistants like Alexa, Google Assistant orSiri have seen a large adoption over the past years. Using artificial intelligence (AI) technologies, they provide a vocal interface to physical devices as well as to digital services and have spurred an entire new ecosystem. This comprises the big tech companies themselves, but also a strongly growing community of developers that make these functionalities available via digital platforms. At present, only few research is available to understand the structure and the value creation logic of these AI-based assistant platforms and their ecosystem. This research adopts ecosystem intelligence to shed light on their structure and dynamics. It combines existing data collection methods with an automated approach that proves useful in deriving a network-based conceptual model of Amazon's Alexa assistant platform and ecosystem. It shows that skills are a key unit of modularity in this ecosystem, which is linked to other elements such as service, data, and money flows. It also suggests that the topology of the Alexa ecosystem may be described using the criteria reflexivity, symmetry, variance, strength, and centrality of the skill coactivations. Finally, it identifies three ways to create and capture value on AIbased assistant platforms. Surprisingly only a few skills use a transactional business model by selling services and goods but many skills are complementary and provide information, configuration, and control services for other skill provider products and services. These findings provide new insights into the highly relevant ecosystems of AI-based assistant platforms, which might serve enterprises in developing their strategies in these ecosystems. They might also pave the way to a faster, data-driven approach for ecosystem intelligence.

Introduction to the Minitrack on IT Adoption, Diffusion, and Evaluation in Healthcare

“…Research by (Schmidt et al, 2021a) shows that AI-based assistants evolve towards digital platforms by creating rich ecosystems with skill providers and device vendors. Direct and indirect network effects within and between these groups may be observed (Schmidt et al, 2022), which opens further strategic venues for applications in practice and research in academia.…”

Impact of Telemedicine Technology Tools and Perceived Barriers on Satisfaction and Continued Usage Intention of Office-based Physicians

“…We use the re-engineered data model of assistant platforms (Schmidt, Alt, et al, 2022), which recognizes modules as a key element for assistant platform ecosystems. To organize the available modules, they are stored in a module registry (see Figure 3), which is an established solution in service-or component-oriented architectures.…”