The AI craze is officially upon us with the majority of marketing and consumer technology providers promoting some form of advanced machine learning, and some of the leading minds in business and academia like Elon Musk and Stephen Hawking warning of an imminent computer revolution right out of James Cameron’s Terminator film-series, or perhaps more topically after the season finale last night and my post last week, Jonathan Nolan and Lisa’s Joy’s Westworld.
One of the latest entries in the rapidly evolving space is Salesforce’s Einstein, an AI that the leading CRM and cloud solution provider claims is “built into the core” of the Salesforce platform. Einstein promises to be your own personal data scientist and assist with discovering insights, predicting outcomes, determining next best steps, and automating tasks. The language used to describe Einstein is very human, a far cry from the antiseptic modeling language of earlier incarnations of similar software, and he even has a cute little graphic representation. Is the revolution already upon us?
Computers are barely scratching the surface of a small portion of human talent…
As usual in the marketing world, one needs to separate fact from fiction, and the truth lies somewhere in between. It’s a fact that big data, ridiculously fast processing power, and increasingly sophisticated algorithms are completely changing our perception of what computers are capable of, and that some of the latest and greatest software and devices can seem intelligent at times. It’s also a fact that this new technology has real benefits for marketers as they seek to engage in truly personalized interactions with their visitors.
Thanks to solutions like Einstein, the same types of technologies that power a brief exchange with Siri that resembles a real conversation or Google’s predictions about where you’re headed on a road trip, can be applied to your marketing campaigns. This offers your organization the benefit of ever more sophisticated and targeted models to uncover new relations among your customers and prospects, and the ability to improve engagement by personalizing interactions in real time.
These are truly amazing developments in the marketing world, and they will change the way we plan campaigns, segment customers, and measure ROI, but, getting back to the original topic, is any of this truly intelligence in the human sense of the term? Or, more dramatically, do we need to be worried about the future of the human race if we set Einstein loose on our target audiences?
For example, a Presidential election forecasting model that…well, there’s part of the problem in a nutshell…
That’s a far more difficult question to answer, partially because intelligence itself is very difficult to define, and there are a lot of behaviors — memory, creativity, intuition, reasoning, etc. — that are usually blended together when we’re talking about the general subject of human intelligence. No offense to Mr. Musk, Mr. Hawking, Mr. Cameron, or Mr. Salesforce, but I believe computers are barely scratching the surface of an exceedingly small portion of human talent, and that a lot of what we are seeing with the AI craze is anthropomorphic language over-selling a certain type of data-driven analysis.
At its most extreme, I think it is fair to say that current technology is capable of (very) limited forms of what we would traditionally call deductive and inductive reasoning, though even that is a stretch given that the algorithms that actually draw conclusions from the data are written in advance by human programmers. In the deductive case, the appropriate algorithm and associated databases are designed to draw conclusions based on multiple premises. For example, I often go to my mother’s house on the weekend, therefore if I am traveling on the road I usually take to her house and it is a Saturday afternoon, Google assumes that’s my destination; Google might even give me the time to her house in advance every Saturday morning.
The inductive case is a bit more complicated. A large amount of premises that are either true or mostly true are turned into specific predictions. This is usually done with the assistance of a living and breathing data scientist, where the computer crunches the numbers and the human reviews the output. For example, a Presidential election forecasting model that…well, there’s part of the problem in a nutshell. Models are nice and neat; the real world is a lot more messy and difficult to capture in terms of the discreet data points that feed both artificial intelligence approaches. While a certain marking persona has the propensity to be a new customer, that doesn’t mean that an actual person will be one.
It’s dealing with this very messiness via the human capacity for intuition and ingenuity that has yet to be replicated by a machine, and — in my opinion — remains an essential part of true marketing intelligence, but before we return to the marketing world let’s consider the real Albert Einstein, usually known as one of the most brilliant minds to ever walk the Earth. How does his Salesforce-branded AI mascot stack up?
While a certain marking persona has the propensity to be a new customer, that doesn’t mean that an actual person will be one…
I think most scientific historians would acknowledge that Einstein’s greatest gift was the ability to make incredible intuitive leaps, to combine completely unconnected data in entirely new ways that ultimately invalidate (some of) the old assumptions and data. In Special Relativity, it was assuming that the ether didn’t exist at all and there was no privileged frame of reference. In General Relativity, otherwise known as gravity, it was the idea that acceleration and gravity are the same.
Furthermore, the key insights often came from simple thought experiments. For General Relativity, Einstein imagined an individual in outer space in a windowless elevator, if the elevator was being pulled upward at the same acceleration as Earth’s gravity exerts downward, the individual would have no means to determine if he or she was accelerating or in a gravitational field. He also imagined that if you were to jump from a very high building along with some office supplies, everything would fall down at the same rate and — for a while at least — it would seem that you weren’t subject to the effects of gravity and that you were in free fall in outer space, therefore acceleration can also cancel gravity’s effects.
This kind of insight and imagination — or dare we call it storytelling — isn’t subject to reason or modeling until after a living, breathing Einstein makes the mental leap. In Einstein’s case, these leaps lead to assumptions that lead to some of the most powerful ideas in human history.
Identifying and engaging customers without us worrying about the fate of the world…
In short, this isn’t the kind of conclusion Salesforce’s Einstein is going to make anytime soon. While the revolution will likely be televised, it’s still several generational leaps of technology in the future. In the meantime, we can safely enjoy the benefits of big data and machine learning as they make our personal and professional lives easier.
In other words, let Google keep helping us avoid traffic, and the new Salesforce Einstein identifying and engaging customers without us worrying about the fate of the world. Let’s also be sure to remember that marketing is both art and science; machines can help round out the science, but humans are needed to develop truly engaging stories and exercise creative judgment on how best to connect with other humans. This is true for marketing topics large and small; we can discuss further over my next post, Google vs Google: Do You Need Best Practices for Your Best Practices?