Data technology and machine learning concept. Information data flow into artificial intelligence head with electronics circuit from bottom. Cogs and light bulb on background.
PHOTO: Shutterstock

Artificial intelligence (AI) is shaping the way we shop, surf social media and interact with our favorite brands. But that’s just the beginning. Gartner predicts that by 2020, 85 percent of our interactions will not be with humans, all thanks to AI. Accenture forecasts that AI technology will lead to 1.8 million job losses, yet that same study predicts AI technology will generate 2.3 million jobs because of the new skill sets necessitated by AI technologies. Another positive finding from the study shows that AI technology will increase business output 6.5 times.

In other words, AI is going to shift things around — a lot. Humans may no longer be at the forefront of customer interactions, and marketing calls may be based more on hard data rather than individual ideas discussed over a conference call.

With this technology in flux and advancing into mainstream, we asked industry experts and practitioners to weigh in on how AI is transforming the everyday customer experience.

Related Article: The Challenges of Delivering Personalized Customer Experiences

Personas at Scale, Hyper-personalization and Predictive Intent

Speaking at dotCMS’ Bootcamp 2018, Chief Sales Officer Stefan Schinkel explained how AI will help brands identify patterns, automatically generate personas and eventually learn how content is consumed at scale so that brands can automatically deliver the right kind of content — and at the right times. “[AI and] machine learning is really all about patterns and understanding those patterns, but also creating personas,” said Schinkel.

“Let’s say, [if I want to have] a couple of billion persona definitions, it is impossible to [create] that manually. So that’s where machine learning can help. [In addition], we also have content being consumed. There’s [a] correlation between how content is being consumed vs. the personas you want to define, and to do that at scale, again, it is something you cannot do by hand,” Schinkel explained.

Dee Smith, CTO at Linx Communications, commented on how AI is changing the way brands are interacting with their consumers by automatically adapting the experience to both consumer behaviors and segmentation. Smith continued by saying that AI is allowing companies to deliver, “...hyper-relevant, hyper-personalized content experiences through targeted creative, topic recommendations and predictive intent. For instance, if your customer is looking at a particular product or service on your website, AI takes that information and allows you to send emails or messages with additional information and offers related to that product or service,” said Smith.

“From topic affinity to engagement models, AI is making it possible to deliver truly dynamic onsite experiences, which allows the user to essentially create their own experience through self-discovery and personalized offers. Savvy marketers are now able to track, analyze and predict future behavior at scale using first and third party data and analytics. Through modeling and predictive analytics, these algorithms are now able to produce actionable insights so that businesses can apply them in real-time,” Smith continued.

Schinkel also added that “predictive analytics and predictive targeting is becoming the foundation of content targeting and personalization. [Predictive analytics and targeting] can deliver personalized experiences that are based on intent — and there is no way you can do that without predictive analytics.”

How Can AI Benefit Your Digital Experience in the IoT-era?

With consumers and enterprises adopting Internet of Things (IoT) devices at an exponential rate (it has been estimated that there will be over 75 billion devices worldwide in 2025), it is imperative for brands to carefully map their experiences for each device. But as Schinkel noted, this is a long and arduous process — unless you’re leveraging AI. “[With] the multitude of IoT devices and the number of touchpoints that will increase rapidly, if you want to map the right content to all those touchpoints by hand, that’s [not going to happen]. This is where you need [AI] algorithms to do this for you at scale,” Schinkel said.

Related Article: 11 Industries Being Disrupted By AI

How Should Brands Approach AI to Enhance Their Digital Experience?

When approaching AI technologies for content generation, Matthew Potter, programmatic content manager at PACIFIC, advised brands to start with content types that can be scaled easily. “AI content should be used for product descriptions, FAQ-style pages, and any other content [that informs the] reader,” Potter began. He noted that AI shines the brightest when it comes to those content categories.

“This is because the content is delivering information in an easy-to-read narrative. It is also easier to scale this type of content because, in many ways, it is formulaic. It is a challenge to translate AI content into different languages. Natural language generation and natural language processing (NLP) software still make poor choices when moving from one language to the other,” Potter said.

Eddy Swindell, Co-founder and CRO at London-based Fresh Relevance, added that instead of following the “common misconception” of the primary function of AI, which is predicting what the end-user will purchase, brands should provide their AI with product-focused data. “AI, if fed with appropriate product information and direction, is even more capable of surfacing products that need to be sold promptly to the appropriate [consumer] at the appropriate frequency to optimize the conversions of those specific products. Increasing the emphasis placed on certain products as opposed to focusing entirely on overall conversations leads to the emphasized product sales moving more quickly,” Swindell said.

Swindell continued by explaining how AI benefits from its ability to “learn rapidly from a small amount of data.” He gave an example of a brand selling seasonal products or fashion items. “Considering that with new seasons' products there will initially be limited data on these products and the people that purchase them. It is likely that clustering techniques derived from behavior around last season’s products can be rapidly applied to a new season's products if we know enough about [those last season] products,” said Swindell.