If you’ve ever asked your iPhone for driving directions, binge-watched a TV series, or searched the world wide web, you’ve interacted with AI — perhaps without even knowing it. Apple’s Siri, Amazon’s Alexa, and Netflix’s magical movie recommendation tool are just a few of the products that use machine learning to make your life easier.At ��������, we’re using machine learning to create better experiences for our partners, customers, and our employees, too. John Thomas (�������� VP of Data Science and generally brilliant individual) shares more on just how machine learning is helping us build our business below: 

What even is artificial intelligence?

AI is all about making machines smarter than they are today. We often categorize a system’s level of intelligence based on capability and sophistication, ranging from “performs tasks very well” to “could beat John Cena in hand-to-hand combat.” Here are the three levels at which AI has the capacity to learn, create ideas, and respond within complex environments:

Artificial Super Intelligence

Thanks to Hollywood and science fiction novels, artificial intelligence has garnered the reputation of something along the lines of “robot overlord takeover.” In reality, this idea refers to artificial super intelligence — a system with capabilities far surpassing those of humans. Right now, ASI isn’t possible. But we shouldn’t rule it out of our near future. Elon Musk has been in his belief that while deep AI has great potential to help humanity, it could also be our greatest threat.

We were all thinking it.

Artificial General Intelligence

A step down from artificial super intelligence, artificial general intelligence is any system able to complete tasks capable of being executed by a human. This technology isn’t projected to take off until the 2040s — a timeline slightly closer than that of ASI.

Artificial Narrow Intelligence

As the most common version of artificial intelligence, artificial narrow intelligence, or “weak AI” is simply able to complete one task. All artificial intelligence that we see today fits into this category, and it’s currently being implemented at ��������.

Cool. What does all that mean for ��������?

As transactions are going digital, customer behavior is becoming more and more complex. One way we’re using machine learning to solve problems in the home services space is through intelligent call routing. By employing voice recognition technologies (a type of artificial narrow intelligence) across several businesses, we’re able to collect information and use it to ensure we’re getting callers to the right place.

(Without having to do this.)

On the marketing side, our data scientists and engineers developed a product in-house which allows us to automate/optimize near-real-time bids across paid search, social media, and other paid media. Using machine learning, we can predict customer behaviors and outcomes with efficiency and precision. (Read all about that here.)Machine learning also allows us to segment audiences and launch hyper-personalized website experiences for each unique visitor. This way, we can place different creative assets within different experiences to drive different outcomes. It’s powerful technology that relies on a data science model running in the background to identify audience segments, and it’s compounded by our customer experience analysts, designers, copywriters, and engineers who are constantly testing new assets to improve the experience and create more value.But for all these great improvements, there’s a laundry list of things that machines just aren’t good at. Algorithms don’t have empathy. The ability to relate to a customer and ease their hesitations is uniquely human. And you can’t ask a robot to “be creative” or come up with a solution that isn’t in a fixed set.

Exhibits A-Z.

The bottom line: we’re building some really powerful technology at ��������, but our people will always be our greatest problem solvers. With machine learning in our toolbox, it simply means we’re able to learn and solve problems — faster.