It’s hard to go far in the tech industry without running into the idea of super-smart devices taking over our lives. The terms tossed around vary – artificial intelligence, deep learning, machine learning, neural networks – and I’ll be the first to admit the predictions can get a little breathless.
So, let’s separate fact from fiction. What the heck is everyone really talking about when they use these terms, what are the practical applications, and, for investors, where is the money to be made?
It’s About Mimicking Your Brain
I like the term cognitive computing because it gets to the heart of the matter: it’s about using technology to mimic the way your brain works.
That’s a heavy undertaking. We acquire information in a number of ways, including thought, experience, and senses. Watching how quickly a young child picks up language skills, for example, is to behold an awesome amount of cognitive power.
As a technology investor with nearly two decades of experience, I see no program or device remotely close to matching the brain’s capabilities. What I do see, however, are real and substantial breakthroughs in technologies that can recognize patterns, draw inferences from them, and pose questions about what it has found.
Driving these breakthroughs are three things: more sophisticated algorithms, ever-faster processors, and, critically, enough data to test the software under true-to-life circumstances. While they won’t replace the brain, this triad will drive a quantum leap in digital assistance to help companies keep pace with the torrent of data flooding them every day.
More important, they will drive billions in investment deals and corporate growth.
Asking The Right Questions
We have storehouses of information whose size balloons daily with every spreadsheet, internal report, mathematical model, social media post, and video clip. So much data – so much we could learn. But what if we don’t know the right questions to ask?
That was the dilemma the founders of Cognitive Scale set out to tackle – and their early success illustrates the practical, and profitable, potential of cognitive computing.
One of the company’s early engagements was with a national retailer that had strong Web traffic but weak sales conversation. Customers would visit their site and then buy elsewhere – and no one knew why. Cognitive Scale began by making each customer the center of its attention. It looked at online clicks and external data to build hundreds of portraits, each with a nuanced set of buying habits. In other words, it treated customers as individuals – just like a brick-and-mortar store. And it worked: the retailer saw a 13% increase in conversions, a significant jump.
With the right algorithms and enough data, it’s easy to see this kind of personalization applied to everything from bank lending and portfolio management to education. Because when you understand precisely what a person needs, delivery becomes much easier problem to solve.
The second area of focus of Cognitive Scale – and where cognitive computing will play a major role – is operational processes. For instance, the company let loose its algorithms on the claims-processing operation of a major hospital. By learning what made a payment request likely to be rejected by a health insurance company and adjusting accordingly, the software cut the hospital’s processing time by two thirds.
That kind of power is astounding, achievable, and people will pay for it.
Sizing The Market
The predicted revenue chart for cognitive computing software is the hockey-stick shape that gets investors like me excited: from $109 million in 2015 to $10 billion by 2024, according to Quid.
This kind of growth is great for investors in two important ways. First, it legitimizes the market: questions of what cognitive computing is and what it can do are beginning to fall away. That means startups can have more meaningful sales discussions, more quickly, and with more potential customers. The distance to positive cash flow is far shorter than it was five years ago.
The second reason investors should be optimistic about cognitive computing is that there are big buyers with deep pockets looking for startup talent. The big five of high tech – Amazon, Apple, Facebook, Google, and Microsoft – are moving quickly to buy up expertise and personnel. Little wonder that there was $8.5 billion invested in the market in 2015, four times that of 2010.
All Pieces Finally In Place
People have been working on cognitive computing since the dawn of the industry. For a long time, the biggest barrier was technology itself: a lack of focused and affordable processing power; an inability to easily gather large data sets; and algorithms that could not be tested under real-world conditions.
The steady march of Moore’s Law and the rise of cloud computing have wiped away the first two obstacles. As a result, engineers are getting better at designing software that mimics the brain and teaches itself how to get better. It is hard science, to be sure. But it is science that is advancing steadily.
Expect it to get more attention, funding, and success in the year to come.