Increasingly, human beings are delegating decision making to algorithms and bots. This transfer of power is seemingly everywhere. From self-driving cars to robo-trading, people are growing more comfortable with software making critical day-to-day decisions. Technology companies are doubling down on this shift with heavy investment in artificial intelligence, meaning that as AI technology becomes more readily available to consumers, algorithms will become responsible for more everyday purchasing.
The delegation of purchasing power signifies a massive shift in retail, marketing and advertising behaviors. To put it another way, the AI robot overlords are coming, and they will have credit cards. As marketers, we need to shift, too, refocusing our strategy on a new relationship: business-to-algorithm marketing. We’re going to need to change our mindset to sell to both humans and AI enabled autonomous agents.
While it may be too soon to accurately answer how marketers should do this, we can evaluate the birth and life of this new consumer.
The best way to understand its potential behaviors—and the corresponding potential size of this new market segment—is to look at technologies already in market that demonstrate AI’s capabilities. The ability to see, a sense of taste, a sense of responsibility, and a voice—these are human arenas in which AI with massive amounts of power is already playing.
Algorithms that See: Search Engines and Comparison Shopping Engines
Search engines and Comparison Shopping Engines (e.g. NextTag, Shopzilla and Google Shopping) marked a milestone in the marketing and advertising industry. With their introduction, marketers pivoted from creating content and messaging for consumers and instead focused on creating content for search algorithms.
Doing this successfully required learning how to communicate effectively with algorithms. Both search engines and CSEs harness data such as URL structures and microdata in order to make decisions about products. This same data is irrelevant to humans, who look at photos and read descriptions to help guide their online shopping. In order to capture wider market share, savvy marketers adjusted their strategies accordingly, accommodating both AI and human reading and decision making styles.
These shifts in marketing and advertising foreshadow how we may need to reach new AI consumers—not only through websites and apps, but also through structured content and APIs. The shifts also indicate the potential scale of this new industry. Marketing budgets are being siphoned off to “sell” to search engines. $65 billion dollars were allocated to this effort in 2016, with spending expected to grow to $79 billion by 2018¹. If that is the money spent trying to appeal to algorithms that don’t even have direct influence over purchasing, imagine what may happen when those algorithms actually have the ability to transact.
Algorithms that Set Taste: Recommendation Engines
As marketers, it’s our job to make algorithms “want” our products. This introduces an entirely new logic of influencer marketing targeted at AI. Understanding Recommendation Engines—aka recommenders—sheds some light on how to approach this issue.
Recommenders give AI consumers personality and opinions. They first hit mainstream commerce with Amazon’s “You may also like…” feature and have matured and proliferated rapidly with the growth of big data. These are learning algorithms that can filter down massive possibilities to just a few correct options.
Recommenders use different approaches to model an individual’s taste and link it to content and products. Take Collaborative Filtering, which is deployed in systems like Netflix and Amazon. This approach uses the buying or browsing behavior of thousands of individuals to determine recommendations, meaning that the recommendation isn’t influenced by what you are selling or individual interactions with the product. Recommenders are ambivalent to SEO-optimized snappy marketing copy and product descriptions. Instead, they make their recommendations based on aggregated data of buyer behaviors. Essentially, they curate taste for consumers based on data, so understanding how to sculpt their preferences through data will be crucial to marketers in the future.
Algorithms that Model Responsibility: Robo-Trading and Robo-Advisors
Marketers need to prepare to build purchasing trust between humans and AI that can rapidly scale in order to stay at the forefront of commerce. The application of AI in the financial services industry exemplifies how this kind of thinking has quickly made a huge impact.
Robo-trading algorithms and robo-advisors, their more consumer facing cousins, represent a significant step forward in the emergence of the AI consumer because they directly test human ability to trust algorithms with money…A LOT of money. From 2008 to 2011, as much as 2/3rds of all financial trades were made by high-frequency trading algorithms. Algorithmic trades decreased to 50% after the 2010 crash, but in 2012 these trades still equated to $1.6 billion dollars per day².
Building off of the trust in trading algorithms, robo-advisors have also been widely adopted. In 2015, they managed approximately $60 billion worth of investments, and it is estimated that they will manage up to 10% of global assets, or approximately $8 trillion dollars, by 2020³. This booming growth points to the growing purchasing trust humans are willing to place in AI.
Algorithms that Have a Voice: Virtual Personal Assistants
As time advances, more AI consumers will buy products and services for more human owners without oversight. This purchasing behavior is already modeled by Virtual Personal Assistants— autonomous applications (including Alexa, Siri and Google Home) that perform the functions of a human assistant. Known as VPAs, this realm of technology is rapidly proliferating as a pure play voice enabled channel.
Chatbots laid the foundation for the adoption of VPAs. Like bots, VPAs have the power to simplify complex searches through large sets of options. They combine this power with the ability to process natural language, meaning that users can interact with them through plain speech. As VPAs continue to earn consumer trust, they’re extending their reach into more types of tasks and gaining more power as purchasing proxies. This adoption of voice-based AI induces heavy branding and marketing implications. The typical arsenal of visual tools that differentiate a brand, such as product, packaging and identity systems, needs to be reimagined in order to better appeal to these autonomous agents.
The pieces of the new AI consumers described above are not emerging technologies—they’re already here, and marketers need to adapt. Human consumers live in a world of physical stores and digital websites, navigating with steps and clicks. The new AI consumer lives in a world of APIs and data, navigating by parsing and processing. Today’s marketers are focused on using big data and AI to sell. But it’s time to go one step further and look at how those same systems can be used to buy. Today’s designers need to broaden their human-centered approach to incorporate empathy for algorithms, better understanding how they work in order to appeal to their growing purchasing power.
By Cory Clarke, Partner Technology Lead, VSA Partners
Cory Clarke has led a career that bridges the technology and design fields, and he continues to drive innovation as head of VSA Partners’ Technology Practice. His experience ranges from working with major clients including Google, McDonald’s and IBM, to investing in and incubating startups, such as Paddle8. He has also taught at Columbia University and Pratt Institute and has spoken at events such as SXSW Interactive and SEGD Xlab.
This article was originally published here