Machine learning hits the cloud -- and more businesses

The cloud makes business artificial intelligence cheap enough to incorporate into your business operations

Machine learning hits the cloud -- and more businesses
Credit: flickr/Seniju

"Computing that can think" has been an IT battle cry for the last 30 years, but the price of artificial intelligence has been out of reach for most enterprises. Enter the cloud -- along with its cost efficiency -- and perhaps a revised business case for thinking machines. Amazon Web Services, Google, and Microsoft are all on board, and they each have some kind of practical AI offering.

In the context of business, artificial intelligence is called "machine learning" -- that’s the best way to explain it in boardrooms. Machine learning can determine patterns, think about how those patterns are emerging, and even get better at predicting patterns before they emerge.

An example is an inventory control system. You know that the amount of inventory you keep on hand is determined by demand. You also know you can do a fairly good job of predicting demand based on what the demand was last year and the year before. You might be able to pull off 65 percent accuracy.

Machine learning takes this concept to the next level, in that you can change how the system needs to think via the patterns you detect. In the inventory example, you can also take into consideration economic data, weather data, changing demographics, the overall pattern of the business, and so on.

However, machine learning needs to do more than determine correlations, which you can already accomplish with data science. In fact, machine learning brings a certain value: The ability to remember how the correlations actually made productive calls, as to what to do with inventory or, to a great extent, what to do with the business -- that is, actually thinking and making decisions based upon experiences, not data alone.

Does it make sense for your enterprise to use cloud-based machine learning? I think the answer for most is yes. Most business processes can benefit from the use of automated decisions that work with machine learning and good data analytics.

If you told me this would cost several million dollars to implement, I would say the risk is too high. However, if you told me you could use a cloud service for a few hundred dollars a day, then I'd say we’re in business. Guess what? We're in business.