Example document docx

Speaker 1 (00:00)

Most of the things we hear about today in marketing or just in our general lives are really uses of machine learning, which you could probably define better than me. But the definition we always go with is making predictions of future outcomes based on historical data up. Hey, it's Paul Racer, founder and CEO of Tr 2020, HubSpot's first partner agency. Back in seven, we started the partner program and also more recently, the founder and CEO of the Marketing Artificial Intelligence Institute. So I'm here with Kevin Walsh, today product manager for HubSpot machine learning product management for Machine Learning HubSpot, to talk about the applications of artificial intelligence and digital marketing. So be sure to like or subscribe to the HubSpot Academy YouTube channel. And with that, throw it over to Kevin.

Speaker 2 (00:48)

Let's start at a really high level. Can you tell us what is artificial intelligence and what is machine learning? Yes.

Speaker 1 (00:54)

To me, the thing I would say about artificial intelligence is if you search on Google for what is AI, you're going to get ten different definitions from ten different experts. So what I always try and talk about is just at a very high level, it's kind of the umbrella term to encompass technologies and algorithms designed to make machines smart. And then within that, machine learning being the primary subset, most of the things we hear about today in marketing or just in our general lives are really uses of machine learning, which you could probably define better than me. But the definition we always go with is making predictions of future outcomes based on historical data. But the key with machine learning is the machines get smarter and the predictions get better without human intervention necessarily thinking again about.

Speaker 2 (01:38)

Do you have any sort of specific consumer examples that people might be familiar with that are using machine learning day to day in their lives? Yes.

Speaker 1 (01:48)

The best example I always give of AI and machine learning in particular is Google Maps. So when we try and make it less abstract for people, we just talk about the fact that you're using it dozens if not hundreds of times every day in your life. So if you think about the apps on your iPhone, from Facebook to Instagram to Google Maps to Amazon to Spotify, every one of them is using data to try and predict your behaviors, predict what you're likely to do next or what you're going to be interested in or what you're going to buy. And so Google Maps to me, is one of the best examples of what we're trying to get to with marketing. So Google Maps makes predictions about the most efficient way to get from point A to point B. But the human at the end of the day still makes the decision. And I think that's where we're trying to go with marketing software is to make it smarter, to make it predict a little better, what might happen. Or the best route to take, but the human still makes the final decision. So there's a book called Prediction Machines that I really like that's written by three economists, and they look at the future of jobs and say that it'll basically be telling machines what predictions to make and then figuring out what to do with those predictions, like using human judgment on those predictions.

Speaker 2 (02:58)

And so kind of what is the state of artificial intelligence then within marketing, sales and services as we think of them?

Speaker 1 (03:04)

I look at it and think it's very beginner level. Now, again, you may have as someone who's overseeing the building of some of this stuff within HubSpot's platform may look at a little differently, but I generally think that the technology is at the very early stages of development and application to marketing. Now, there are other industries where it's raced forward. So you could look at just Wall Street, for example, in finance, this Stuff's 30 plus years old. They've been trying to do this stuff to figure out when to make trades and how much to invest. So looking at the applications to marketing, it's just early. The money has started to move in. We track about 1100 AI powered sales and marketing technologies, and about 500 of them have funding combined for over 5 billion. But a lot of that money is funding companies that when you dig deeper, they honestly don't really have much proof that it works. So they talk about a and they talk about machine learning, and they claim to have been building smarter technology. But if you press for case studies of how it works, it's often you find that they don't really have them, and that really the machine learning and ads is part of a roadmap, and they're just very beginning stages of testing.

Speaker 1 (04:12)

So our experience has been very early, but that's actually a good thing because it means there are opportunities for both the tech companies and for the marketers that figure out what this stuff is and are Proactive in trying to find smarter, more efficient ways to do things.

Speaker 2 (04:28)

Yeah, I tend to agree. Here at HubSpot, we've been doing machine learning seriously, I would say for about two years, a little more than two years, and we've grown that team tremendously. It's now a full fledged group here within product. And we do take a lot of inspiration from those bigger brands like Google and Spotify and Netflix that have pretty successfully implemented, recommender and predict their systems that people know and love, like Netflix. I mean, who hasn't been injured Netflix over and over again, just because you would like this is getting almost spookily accurate.

Speaker 1 (05:01)

Yes. I think the challenge that we face is you're trying to as brands, even if you're a B to B brand. Consumer experiences have changed, and their expectation of the simplicity and personalization has changed. And so whether it is like a Spotify or a Netflix or Amazon, they're so used to that predictive capability. That when you start getting into having to buy and deal with chatbots that aren't intelligent, that are human based logic and going through what we try and manufacture as the customer journey, as marketers, there's a lot of I know the big SAS world loves the word friction. There's lots of friction in the buying process today that doesn't exist on the consumer side. And so I think that's the challenge a lot of big brands have and especially on the BW side is trying to find a way to remove that friction by building more intelligent solutions.