Can machine learning save companies millions?

While it might sound like the stuff of science fiction, we’ve reached a level of technology where  computers are learning to solve complex processes on their own – and that could save businesses a lot of money. Tech gurus have always been able to wrangle impressive results from clever coding, but we’re entering a world in which we no longer tell computers what to do and how to do it – but rather build them to teach themselves.


That might sound a bit Star Trek, but just a few months ago, two Facebook chatbots decided that English was too inefficient to communicate with each other, and threw the rules of English out of the window to streamline the conversation. The results? While not evidence of the technological singularity, it does demonstrate the point: machines can learn to develop new solutions to solve complicated tasks, and companies are beginning to see the potential.

Outside of quirky tech experiments, could there be a future in which we trust computers to make critical decisions that could potentially save businesses millions? Some of the biggest tech companies seem to think so, and have already started implementing the technology.

Artificial Intelligence. Quietly clever.


Artificial Intelligence (AI) has been creeping into our lives over the past few years – sometimes without us even realising it. Take customer service, for example: live chat has become a staple on many websites, giving consumers the ability to talk to an advisor immediately, and get the support they need.

Now, you’re just as likely as not to be speaking to a so-called ‘chatbot’, rather than a real person – and it can be so convincing you can struggle to tell the difference. It’s an example of how companies are using AI to streamline departments, cutting out the need to employ additional staff.

Microsoft have quietly been turning machine learning expertise into practical solutions for years. It’s how Clutter on Office 365 can decide whether you want an email in your inbox; how OneNote can decipher your handwriting; and why the touch-screen keyboard on the latest Windows phone is so accurate.

“Everything we do now is influenced, one way or another, by machine learning.” – Peter Lee, Director of Research, Microsoft.

But it is also being implemented in their internal processes too. According to VP of Machine Learning, Joseph Sirosh, every transaction placed on Microsoft’s website is screened by machine learning models behind the scenes. “It has saved us millions every month because of its power to detect what’s a good and bad transaction”, he reports.

Learning from customers

At Feefo, we’ve long known that extracting data from customer feedback can lead to money-saving, actionable insight. That’s because reviews are loaded with some of the richest insight you can possibly receive – and for machines to learn, they need a lot of data.

One client, after analysing all their reviews with Insight Tag technology, recognised a trend in negative responses that identified a problematic supplier. They were able to correlate this information with specific products, and narrow the issue down to a conveyor belt in the factory that produced them. Now that’s deep insight.

Now imagine machine learning technology doing all of this intuitively: analysing everything your customers are saying, all of the time, and automatically identifying trends. Machine learning is helping businesses understand current customers, to acquire new ones. We now have the technology on hand to analyse historical sales data, behaviour patterns, purchase history, buying preferences and even social media activity on the fly.

It’s like having a super-powered squad of data analysts, efficiently pulling apart feedback, cross-checking product information, and presenting you with a report of steps to take to make crucial business decisions with confidence. And the best part? It all comes from what your customers are already saying about your business.

Net Promoter® and NPS® are registered trademarks of Bain & Company, Inc., Satmetrix Systems, Inc., and Fred Reichheld.