The Gap Between AI Opportunity and Reality
How to Increase the Reliability of Your IT Infrastructure Using Predictive Analytics REGISTER >
Most businesses aren't putting artificial intelligence (AI) to work in any meaningful way. There's a gap, if not a chasm, between opportunity and reality.
There's a growing paradox related to artificial intelligence (AI). On one hand, speech processing, image recognition and other algorithms are advancing at a furious rate. They're fueling autonomous vehicles, robots, drones, augmented reality (AR), virtual reality (VR) and much more.
On the other hand, commercial software providers and most businesses aren't putting AI to work in any meaningful way. There's a gap, if not a chasm, between opportunity and reality.
Consider this: Siri is now built into the Mac OS Sierra desktop interface, as well as in iPhones. It works reasonably well for basic functions, such as web searches, arithmetic, time zone questions, commands and a variety of other basic tasks. Yet, if I ask Siri to search my computer for certain content—say files, emails and other documents containing the term "cyber-security consultant"—and find people that match the query, it comes up completely empty.
Websites are equally dumb. Amazon is among the best at using AI and predictive analytics to suggest products, but it still has a long way to go.
Travel sites, as I have written about in the past, are utterly pathetic. There's virtually no attempt to understand and apply AI to a customer's preferences, past purchases and general behavioral patterns. When I conduct a search, I simply wind up with a data dump of flights, hotels or activities. I'm left to sift through the seemingly endless tangle of results and pinpoint the possibilities.
Customer service and support also fail in too many cases. Knowledge base functionality is often crude and frequently useless, and customer service bots usually get things wrong. In fact, speech recognition systems sometimes can't understand basic words and phrases and force me to use different language or route me to the wrong representative.
Recently, I said "correct" instead of "yes," and a system didn't know what to do. This isn't even an AI problem; it's a basic programming problem. But you can't get to AI if you don't have the basic programming and business rules figured out.
Fraud detection algorithms are often lacking too. Recently, while booking a hotel, my credit card provider blocked the overseas transaction. What was irksome was that this has happened before—despite the fact that I regularly make travel purchases through the web.
Worse: A text alert that asked me to authorize the purchase arrived two hours later. What happened to real-time commerce? By that time, the transaction had failed, and I had to send the hotel an email and ask them to resubmit it.
You get the idea. Amid all the talk about innovation and disruption, it's critical to remember that basic blocking and tackling is what often wins games—and builds a successful business. As a CIO, the effective use of AI can help you score big in many small but critical ways.