The first HAiLife seminar discussed human data business opportunities
The data-based business is here
Our HAiLife project's first seminar invited two experts to share their perspectives about data economy and possibilities of utilizing human data in business. In the first keynote Mika Ruokonen, Industry Professor of digital business, shared his visions about innovation, new business creation, value creation, business models, growth, commercialization, and business renewal.
According to Ruokonen, hypergrowth in volumes of data enables various new business opportunities. Sensors are everywhere, and our GPS and internet connection are always on, which means that people produce enormous amounts of data while practicing their everyday duties. In addition, cheap data storage is available for everyone. In summary, the ”Lego bricks” to mix and match for data-based businesses already exist. Ruokonen showcased several existing business models and categorized six types of business opportunities for data and algorithms:
Product-oriented models
1. Selling Existing Offerings
2. Selling Data, Analyses, and Reports
3. Selling Data Platforms
4. Selling Ai Assets
Service-oriented models
5. Selling Data And Ai-Based Services
6. Selling Outcomes Or Capacity
In the future, value creation in companies occurs less in manufacturing traditional hardware, but it is increasingly built on connectivity and intelligence and lies in particular in ecosystems, services, and data.
Though a data-driven economy constantly evolves with opportunities, pitfalls may also exist. To avoid those, Ruokonen pointed out that company management should understand the opportunities realistically and invest in commercial and legal know-how. In the human data business, one must follow guidelines, rules, and standards right from the start. Seeking professional advice who also brings in culture & ways of working with data is essential as well as ensuring scalable technology and cloud services.
“Every company is a data company, whether they know it yet or not”.
By investing in commercial and legal know-how, companies can renew their businesses, although it may require company-internal transformation to scale up. Attention must also be paid to the company culture and leadership: the classic top-down hierarchical model may not suit the data & ai driven companies. Ruokonen suggested that companies carefully consider the pros and cons of centralized & decentralized organizational structures.
More and more, data can be a strategic priority of firms. Ruokonen concluded his eye-opening presentation by pointing out that: “Every company is a data company, whether they know it yet or not”.
Human data demands tools to deal with noise and ambiguity
The second expert introduction was provided by medical technology professor Tapio Seppänen, whose specialty is digital data and signal analysis technologies. He is a pioneer in the development of data analysis, algorithms, and applications related to human measurement
Seppänen started from the roots emphasizing that the pipeline from the raw data recorded to the actual customer often involves several procedures to make the data reliable and usable. For example, many of the current technologies used in medical contexts to monitor patients' vital functions must apply filters and algorithms, ensuring that the data artifacts do not bias the results. This step can become even more crucial in “noisy” everyday settings. Seppänen emphasized that these signal technologies demand extensive development and validation, which should be considered when the business is being developed.
Further, Seppänen introduced examples of data gathered from human behavior and interaction contexts to analyze affect and emotions. It is, for example, possible to trace affective states from audio signals or biosignals such as heart-rate variability or electrodermal activity, and several products have already been developed to utilize such technologies. Seppänen closed the presentation by introducing the latest turn to do the signal analysis multimodally by combining different data streams such as video, audio, and physiological signals. One of the remaining challenges is to decide which data channel to rely on the most if the signals remain divergent. An example could be a case where the facial expressions analyzed from the video during a discussion can indicate a particular affect, but the measured physiology “under the hood” can signal the total opposite.
Stay tuned for future events!
Enlightening thoughts from both presentations remained lingering in the air, kindling enthusiastic discussion in many LeaF round tables. We hope to continue these discussions together with researchers and scientists in our upcoming HAiLife - seminars. Stay tuned for more!
During the event, the audience got familiarized to LeaF premises, equipment and methodical know-how. Demonstrations about collecting multi-channel data in practice that are suitable for research and business were also presented to the onsite and remote attendants.