This is a writing sample from Scripted writer Mark Tweedy
In trying to crack the toughest marketing challenges, two heads are certainly better than one — but how about a million heads? Some marketing problems have grown so complex that humans can't or just don't have the time to untangle them. These problems can exceed the processing capacity of not just a human mind, but even an entire global community of minds.
Artificial brains think differently, structurally and functionally. While they can easily breeze through calculations many times faster than human brains, they will never be able to do things like Hilbert's 10th Problem. (Don't worry, AI, that one's unsolvable, even theoretically.) However, even tasks that are relatively simple for humans, such as recognizing a single individual in different contexts across channels, remains enormously difficult if not empirically impossible for AI as it exists today.
So the next question a practical professional must ask is, "What can AI do for marketers right now, in 2018?" The answer is inspiring and goes right to the heart of a new definition for marketing itself.
AI's IQ IRL
In the book "After Thought: The Computer Challenge to Human Intelligence," James Bailey explains in detail how the gap between human brains and AI is not one of degree (such as faster processing) but of kind (AI operates along wholly original logic systems). All the many forms of AI being developed around the world will never think in the same way humans do, but their strengths and weaknesses can be complementary to our minds in achieving specific goals.
The right way to go about applying AI is to start with an understanding of what AI can actually do, not what you want it to do. Professor Andrew Ng, former chief scientist at Chinese search engine Baidu, often consults on AI projects like Google Brain and the Stanford AI Laboratory. He observed that:
"Surprisingly, despite AI's breadth of impact, the types of it being deployed are still extremely limited. Almost all of AI's recent progress is through one type, in which some input data (A) is used to quickly generate some simple response (B)…. So what can A→B do? Here's one rule of thumb that speaks to its disruptiveness: If a typical person can do a mental task with less than one second of thought, we can probably automate it using AI either now or in the near future."
That's exactly what's happening in the marketing world. Here's a sample of the types of tasks AI is handling for marketers in 2018.
6 Current Wins for AI in Marketing
Salesforce's State of Marketing report included the fact that 60 percent of marketers are counting on AI implementations like Salesforce Einstein to have a "substantial or transformational" impact on programmatic and media buying for the year ahead and at least through 2022.
A CapGemini survey of 1,000 organizations found that 83 percent agree AI has already created new jobs and 75 percent have seen a 10 percent uplift in revenues directly tied to AI initiates. Meanwhile, 73 percent said that AI has boosted customer satisfaction scores and 65 percent said AI-based decisions have reduced customer churn rates.
Deloitte's 2017 Global Contact Center survey found that 56 percent of technology, multimedia, and telecommunications companies will be investing in AI this year, especially for "sentiment analysis" of social media data. Sentiment analysis uses AI algorithms to surface clues of underlying opinions in posts, social video, and comments. The sentiments can then be categorized and segmented based on audience types for more accurate targeting.
Voice analytics are helping marketers be more collaborative and persuasive in both internal and external meetings. Voice patterns, unexpected pauses, and previously unintelligible crosstalk can be unraveled by programs incorporating AI and machine learning. An example is a company that discovered their most persuasive speakers instinctively used a talk-to-listen ratio of 43:57. In their most successful interactions, they spoke 43 minutes for every 57 minutes of listening.
Oracle researchers reported that 36 percent of brands have already implemented AI chatbots for customer interactions and market research. A full 80 percent of brands said that they either have chatbots now or intend on implementing them by 2020.
Mass customization sounds like an oxymoron, but AI is making relevant content work at scale. Google is constantly refining its many AI projects but has been particularly successful at dialing up viewable impressions for premium placements. In advertising for Google's own Pixel phone, they used AI to boost impressions on premium inventory by 3X, while reducing the viewable cost per thousand (CPM) by 34 percent.
"In a modern marketing operation, the sheer complexity and variety of signals are way outside the grasp of a human mind. In the next five to 10 years, there will be a machine learning company whose platform sits on top of Google Analytics and derives valuable strategic insight from the underlying data, including directional guidance on which channels to tune, which to divest and which to accelerate."
That's the goal marketers are focused on currently, although AI's potential is functionally only limited by the imagination of innovators.
For many years, the big stories about AI were all about potential. Now that's changed. AI is actually having a measurable impact on revenue and profitability, and in the business world, nothing has ever been possible before it became practical. With measurable performance enhancements in marketing budgets, AI enters a new era as a tool instead of a concept. The next step is in recognizing the limitations of AI and matching it up efficiently with human intelligence to create a new kind of marketing that can deliver exactly what buyers want as soon as they want it.
Data storyteller -- Mark T is a professional writer on Chicago's Data Lakefront, using hard data and scientific research to tell data stories about innovation, startup culture, emerging global trends, and next-gen tech.