Machine learning has come seemingly out of nowhere over the last four years. Things that were once science fiction, like a machine diagnosing cancer or playing Poker, are now a reality. And it’s only going to get more exciting and more extreme from here on out.

What’s so cool about this new technology is how accessible it is. In the past, if you wanted world-beating equipment, you had to pay world-beating prices. But because machine learning is essentially invisible, infinitely copyable algorithms, it’s something that is available to everybody at relatively low cost.

One of the most difficult areas for many companies and startups is the area of marketing. How to market and who to market to are difficult questions with highly technical and sophisticated answers. Fortunately machines – with their incredible data-collecting capacity – are perfectly suited for this sort of thing. And so now companies are exploring how machine learning can help them do marketing better.

Forecasting customer profitability

When customers use services, they leave a trail of information in their wake that companies can use to their advantage. Up until now, companies have had to rely on instinct to figure out whether a particular customer will prove profitable or not. But with the advent of machine learning, they no longer have to play guessing games.

Neural networks can be used to analyze patterns of data created by groups of people on the internet to discover who is most likely to make the company the most money. In effect, machine learning is allowing businesses to better target customers, not just on relevancy, but also on potential profitability too.

In the past, calculating profitability was tough, as businesses could only go on indirect measures, like a person’s income. But today, they can use their past behavior to figure out how they’re likely to spend in the present.

Predict content demand

According to The Data Incubator, the type of content people demand is in constant flux. One moment one topic is trending: the next something else has caught people’s attention.

There are businesses that already take a sophisticated approach to content creation, producing stuff that is related to trending topics on social media. But the problem is that this approach is fundamentally reactive, rather than proactive.

Now, though, machine learning promises to change all that by using data collected from social media to predict what type of content will be demanded by the community in the near future. Startups will be able to better compete in the SEO war if they know ahead of time which topics are likely to be popular, and which aren’t.

Enhance customer support

Finally, machine learning is being used to pair customers to call center staff with complementary personalities. It works something like this: A customer calls up their company with a complaint. A machine learning algorithm then listens to their tone of voice and their choice of words. The customer then gets assigned a particular personality. Next time they call, they are paired with an assistant who matches their personality and can best help them.