Machine learning isn’t just something that will happen in the future. It’s happening now. And it will only get more advanced and pervasive in the future.

So, just what is machine learning, or ML? It’s applying artificial intelligence (AI) to enable a computer (or machine) to learn for itself—or, more correctly, to recognize patterns and then apply the patterns to future input. (Learn more about the interrelationship between machine learning, AI, and deep learning.) This is done with programming that is completed before the machine learning takes place. The programming enables the machine to access data and use it to learn by itself over time without requiring additional coding. The machine is programmed upfront with an algorithm that can recognize and analyze patterns and then act in ways that take advantage of those patterns.

While that all sounds fine and dandy, what does it really mean? Let’s unravel the answer by looking at some real-word examples of machine learning.

Everyday Examples of Machine Learning

At the simplest, best-known level, we all use, or at least encounter, machine learning every day. If you use Siri or Cortana on your smartphone, you’re using machine learning. These apps use ML algorithms to find patterns in your questions and the way you ask them and then deliver more targeted answers over time.

Sites with recommendation engines, such as Netflix, use machine learning to power user-specific recommendations. In the Netflix example, ML algorithms compare a user’s behavior with similar users’ behaviors worldwide to predict which videos to recommend for a user.

Google Maps also uses machine learning. The app pulls location data from smartphones in a given area. That data is then mapped in real-time to determine traffic speed and provide commuters with route suggestions and warnings.

These are examples of enterprises incorporating machine learning into their products or services. But how are enterprises using machine learning for their own benefit outside of customer interaction, retention, and acquisition?

Enterprise Internal Uses of Machine Learning

One way an enterprise can use machine learning is to facilitate its own digital transformation. Digital transformation is driven by data. If an enterprise’s data has been accumulating for a long time, manually getting it in order to mine its value can be an arduous task. To solve the problem and get enterprise data in order faster, machine learning can be used for low-level repetitive data-preparation tasks. Machine learning algorithms can quickly and accurately unify and clean massive quantities of data. What’s more, with machine learning, speed and accuracy automatically increase with time. Human data experts are then free to do higher-level, higher-value tasks.

After its merger in 2008, Thomson Reuters made use of machine learning to prepare large quantities of data with the help of Tamr, an enterprise data-unification company.[1] Tamr helped Thomson Reuters use machine learning to unify more than three million data points with an accuracy rate of 95 percent. The process reduced the time needed to unify the data manually by several months and reduced manual labor by an estimated 40 percent.

GlaxoSmithKline used machine learning to develop information targeted toward parents’ concerns about vaccines. ML algorithms were used to sift through parents’ feedback and inquiries about vaccinations on online message boards and forums. GlaxoSmithKline then developed informational content to specifically address parents’ concerns and questions. Machine learning was key because regulations restrict pharmaceutical companies from directly contacting consumers. Even without those regulatory barriers, having a human search through all those message boards could have taken months.

Fraud detection and risk management are other ways enterprises can use, and are using, machine learning. PayPal uses three types of machine learning algorithms to detect fraud and protect itself, retailers, and customers. Its algorithms quickly determine if a new customer is a risk or might be a risk. “Risk-free” customers are approved virtually immediately while additional algorithms are used for more in-depth validation for other customers as needed.

Allstate uses an internal virtual assistant, the Allstate Business Insurance Expert, or ABIe, to enable agents to find information on commercial products, a critical need after it switched its primary focus from selling personal insurance products to commercial ones. ABIe is an avatar that uses ML to give agents specific information quickly. Much like Siri, it learns with each question how to predict future questions and provide faster, more targeted answers.

Potential Added Uses of Machine Learning in the Enterprise

Despite these real-world use cases, machine learning is still largely in its infancy. Future possibilities are as broad as developers’ and businesses’ creativity and ability to apply machine learning to customer and business needs.

Read More

For more information on machine learning and related topics, my colleague, Glenn Touger, explains machine learning, AI, and deep learning and their interrelationship in his June 30, 2017, post, “What’s the Difference Between Artificial Intelligence (AI), Machine Learning, and Deep Learning?” You might also be interested in his article, “Four Ways Artificial Intelligence Is Changing Lives.” To find out about future posts on machine learning, follow us on our blog, Twitter, and LinkedIn.

[1] Dataconomy. “What Is Revolutionizing Machine Learning for the Enterprise?” June 2017.


Share this: