This article is based on one of our thought experiments we had in the Motius ThinkTank. At Motius we have the privilege of being close to the newest technologies and at the same time working with leading companies around the world on developing the future of different industries. Such thought experiments are based on our observations, discussions with our partners and Friday-mornings coffee-fuelled debates combined with a bit of analysis.
A new market segment of digital butlers
Virtual assistants have lately been popping up around the digital landscapes and gaining more and more attention. Cortana by Microsoft, Alexa by Amazon, Google Assistant, Bixby by Samsung and others are meant to be our personal digital butlers: learn about our needs and make sure we have what we need and want in the digital world. If that vision does become reality, those digital butlers would become a new market segment for digital products. They would need to buy us groceries, take care of our transportation, energy bill, banking, insurance, etc. With companies now scrambling to roll out digital versions of those products (or digitize the channels to them), they have to consider designing the products for the digital butlers rather than their owners. That is a fundamentally different task to the “classic” digital product design/development.
Landscape of digital ecosystems: Few wide-ranging dominating ones and many smaller specialized ones
The idea of a digital butler gains a lot credibility and plausibility from the trend of expanding ecosystems like Google, Amazon, Facebook & Co. Such companies mastered the art of building a digital ecosystem of users by trapping them in with convenience. They cover more and more aspects of their user’s lifes: so much so that they are almost unrecognizable from the companies they were at the beginning. This wide coverage of their users is what is going to make it possible for them to offer digital butlers. The digital landscape might be heading into a combination of few huge overlapping wide-ranging ecosystems and many smaller and specialized products / ecosystems. The digital butlers will probably be developed by the wide-ranging ecosystems to allow their users to navigate that diverse landscape.
This could do for businesses what search engines have been doing for businesses in the last decade. It could be a massive new opportunity that requires new skills to seize. Similar to mastering Search Engine Optimization, digital companies would have to master the skill of digital product development / design for the segment of virtual assistants. But what does that skill actually consist of?
The customer journey of a digital butler
A “customer journey” of a digital butler would be for it to identify a need / a wish of its owner, find products that suit them, like one of them and buy it. Identifying the needs and finding the suitable products is something companies like Google have been doing for years. The “liking” and “buying” parts are new aspects that come with the butler concept and pose new questions for both the designers of the butlers and of the products for them. Digital butlers –as real ones- would be expected to function with minimal involvement from the owner. That would require the butler to “like” according to its owner’s taste on a highly subjective level without his/her interference. Certainly machine learning techniques would allow the designers of the butlers to recognize the taste of the owner on a highly subjective level. A good product design would make it probable for the digital butlers to “see” high-level properties in the product to subjectively like.
Making brands understandable for digital butlers is the new design challenge
Brand is one of those subjective and high-level properties of products. The challenge for the butler lays in understanding its owner’s personality and being able -technically- to understand a brand. To understand the correct brand is the challenge of the product designers. In my case, my digital butler would easily know I am a 28 years-old male and would want to buy new sneakers. It could also find out my budget, preferred colour and size without disturbing me too much. It could even find out what sort of sneakers I bought in the past and what I have been liking on Instagram. However, how would it know whether I am now a Nike or Adidas person? More to the point, how would it know whether I would be a Sbibeeto person, which is a fictive newly founded niche brand?
Product design for digital butlers is training an AI-system
In the language of machine learning, this question could be reformulated as follows: what data should the product developers offer the digital butler in order for it to classify its brand and the owner in the same class? So if I was going to buy or rent my first car, my digital butler would have to infer from all sort of parameters that I would rather enjoy driving a sports car than a GT luxury one. It would also have to go through marketing material, reviews, movies, etc. of car makers and conclude for example that a BMW car would best suit me. Therefore it could say that product design for digital butlers would fundamentally be a training task in the sense of machine learning. That approach would probably result in design principles that are in contradiction to the current ones which assume scarcity of attention and time of human users.
Finally, for the digital butlers to be able to buy products, the developers of those products would have to reconsider their “digitization strategies”. This is a further front of the competition between existing IT ecosystems and companies in the process of digitizing their products. It forces the latter ones to face the question again: whether to embrace the expansion of Google, Amazon & Co. into their markets and use it to grow, or fight it head-on and risk isolation.