Why Small Businesses Need to Pay Attention to What Machine Learning Is

By Kathryn Casna

What is machine learning? Artificial intelligence? Deep learning? Ask a group of small business owners what those terms mean, and most can put together a working definition. The results you get from such an informal survey, however, could be as varied as, “They’re in the CRM I use,” to apocalyptic images, inspired by movies like The Terminator, of robots taking over the world. Either way, most people would likely say that artificial intelligence (AI), machine learning, and deep learning are basically synonymous. The truth is machine learning is related to both AI and deep learning, but this technical concept is distinct.

Machine learning is one of those concepts, along with social media marketing and publishing a blog, that many business owners hear they need to invest in, but aren’t really sure why except they’ve heard that “it’s the future.” (Machine learning really is the future, but that doesn’t mean it has to be as confusing as Skynet seems.) As Greg Corrado, a senior research scientist at Google, told The Guardian, “It’s not magic. It’s just a tool. But it’s a really important tool.” Corrado also believes machine learning will soon be something “everybody can do a little of.”

Even so, few small business owners truly understand machine learning, and only the most tech-savvy have a detailed plan for implementation.

Here’s what machine learning is, what makes it different from AI, and why small businesses should keep an eye on this technology.

Machine Learning vs. AI vs. Deep Learning: What’s the Difference?

While sometimes used interchangeably, machine learning, artificial intelligence, and deep learning are not the same thing. Calum McClelland writes that AI “[involves] machines that accomplish tasks normally associated with human intelligence.” Those tasks usually require planning, recognizing patterns, understanding language, or problem solving.

The term “artificial intelligence” isn’t as new as the hype would have us believe. In fact, John McCarthy, a computer scientist known as “The Father of AI,” first coined the term in 1956. At the time, computers weren’t fast or powerful enough to accomplish tasks anywhere near what humans could do, but that’s changed in the last 60 years.

Today’s developers can achieve AI in several ways. One is to write every single bit of code necessary for a machine to accomplish a specific task. For simple tasks, such as finding webpages containing the precise keyword typed in by a user types (as in the early days of Google), this is a great way to create AI. As user expectations increase, however, search engines such as Google must code more efficiently in order to accomplish a growing list of more complex tasks, such as returning search results using synonyms, spoken language, or similar keywords. Otherwise, it would simply take too much time.

Instead, Google and other companies are writing code designed to allow computers to learn how to pick up on language cues, recognize patterns, and draw conclusions about what users are looking for. Enabling computers to process the information necessary to come to their own conclusions is, in a nutshell, machine learning.

There are several routes to machine learning, and most are analogous to how humans learn.


In this process, humans complete a task, such as identifying photos of certain plants, while the computer observes and learns to identify different species. The result is an app like FlowerChecker.

“Read a Book”

Developers can upload existing machine learning data sets, such as a list of plant characteristics by species, to give an application a knowledge base from which to make decisions.

Deep Learning

Deep learning involves creating complex structures of decision-making trees, clustering information and tasks together, and reinforcing learning. Developers break down a task into “layers,” such as identifying the curves of a leaf, the length of a stem, and the colour of a flower, to allow applications to conclude what species is represented in the image.

Machine Learning Examples

While machine learning has received a lot of hype in the last few years, you’re probably been using this technology in your daily life already without even realizing it.


Next time you ask Siri to guide you to a new location, Apple Maps will use data from other drivers to study traffic patterns and find the fastest route. Similarly, Google Maps helps users plan what time to leave for a trip, and Waze gathers community data to keep users updated on what’s happening on their chosen route. Users input the date and time they’d like to arrive or depart, and artificial intelligence looks at past traffic patterns to predict how long it will likely take them to arrive later today, tomorrow morning during rush hour, or a week from now.

Fraud Prevention

Your credit card company, bank, and other financial institutions use machine learning to study the behaviour of their customers to get better at automatically detecting fraud. This includes recognizing patterns of spending, travel, and more.

Self-Driving Vehicles

If you’ve seen self-driving vehicles around your city, you’ve seen machine learning in action. These vehicles have cameras, sensors, and a “brain” that allow the vehicles to drive themselves. Every time the vehicle makes a trip, it’s not only processing information for the current drive, but also collecting and analyzing data to identify patterns. Machine learning is helping the technology in these vehicles to better identify obstacles and dangers to avoid.

Other Everyday Uses

Other machine learning examples may include thermostats that automatically adjust the room temperature as more people enter and heat up the space, and heat-sensing cameras that help retail stores determine which items and displays get the most attention from customers.

Machine Learning in Business

There are plenty of ways machine learning is already affecting you and your business. Meanwhile, more and more businesses are harnessing this technology to breeze past competitors, making machine learning more important for small business than ever.

How CRM Platforms Use Machine Learning

Businesses of all sizes use customer relationship management (CRM) systems to track interactions with their customers, and improvements in forecasting, marketing, personalized experiences, and more is all possible with machine learning. The best CRM platforms collect data using techniques such as A/B testing, website visits, service calls, and purchase history, then analyze that information for patterns. Insights come when teams pore over the data — or when your CRM platform uses machine learning to analyze the data and predict future customer behaviour.


Many small businesses struggle to offer the “anytime, anywhere” customer service that giants such as Amazon and Walmart are able to. Chatbots can help. In fact, you may have conversed with a chatbot that’s popped onto the screen while you’re browsing a webpage: These applications can ask preliminary questions about a problem, direct the ticket to the right person on the company’s team, and even resolve simple issues on their own. The more a chatbot learns from interactions with customers, and the more complex its deep learning decision-making trees, the more customer service tasks it can help with or take off a business owner’s plate.

Product Recommendations

Seventy percent of shows watched on Netflix are the result of a personalized recommendation by the platform. Those recommendations are possible because of machine learning. For Netflix and other companies that recommend products and services, an algorithm tracks individual customer behaviour, as well as patterns among many similar customers. This algorithm helps businesses personalize their web shopping experience as well as increase revenue per transaction.

With that said, sometimes machine learning doesn’t get it right. Personalized recommendations can be hit or miss, as Netflix and Amazon have both shown. This is just one aspect of machine learning that still needs tweaking — and more human guidance.

How Small Businesses Can Get Started with Machine Learning

If all this sounds too complicated, don’t worry. Until Greg Corrado’s vision of everyone creating a bit of machine learning technology for their own needs becomes a reality, tech companies are creating ready-made applications small businesses can start using right away. Investigate which business applications, from your CRM platform to the software you use to manage your sales efforts, use machine learning and AI, and how they can help your company keep ahead of your competition.

Share this Image On Your Site

About the author:


Kathryn Casna is a digital marketing and travel writer from San Diego, California.

Customer-facing retail, hospitality, and event production make up her professional roots. Today, she runs her own writing business from whatever new locale she happens to be exploring.