In this interview, we chat with Wizeline’s head of growth, Matt Pasienski. As a PhD-educated data scientist and chatbot enthusiast, Matt has lots of interesting opinions on the emerging role of artificial intelligence in the worlds of business and social media. So if you’re curious about any of these topics, don’t be afraid to hit him up on Twitter.
Watch the interview to hear Matt’s thoughts on:
- How intelligent chatbots are simplifying and personalizing the customer experience
- Why chatbots are having a moment (hint: it’s a confluence of mobile technology and affordable, on-demand computing power)
- What steps a CMO should take to get started with chatbots
To learn more about Matt and our team’s work on chatbots, visit Wizeline’s solutions page.
Adam: Everyone, welcome to the program, I'm Adam, and today we're being joined by Matt Pasienski who runs the services team at Wizeline.
Matt: Thank you very much Adam, happy to be here.
Adam: Matt, what is chat bot.
Matt: A chat bot is a very simple way of interacting with the user. If you think about the change in the UX, the user experience that people expect from the web to mobile, a lot of things got simple, very easy, it's one click, kind of stuff. You see a similar simplification to messenger and to chat bot. Now, what you're doing, is you're bringing with chat bot, you're bringing your user experience for whatever app you might have to a very simple text interface. For a certain class of applications that means much lower development times and actually much happier customers because it's so quick and easy.
Adam: What does that mean for marketers now, who want to continue to provide new experiences for their existing audience, and then what does it mean for new entrants and people who are trying to aggregate audiences.
Matt: What's incredible about chat bot is you can get people anywhere. You can actually insert chat bots into a lot of different situations because they're so quick and simple. That means you can respond extraordinarily quickly, you can actually service a much higher rate of customers because you can have automated response. A lot of times, 90% of the questions you're going to get on your website are the same things. You can have pre-canned responses that happen instantly, because if you engage someone in the first 10 seconds they're on your site, that's a very different thing than getting back to them 6 weeks later or the day later with some email.
Adam: The contrarians would say this is only a millennial trend, where only millennials want to interact this way. How do you think about, you know, is this something that's just a flash in the pan, or do you think that this type of interactivity is something that's going to be the defacto rule?
Matt: I don't think that this trend is driven by consumer tastes as much it is about the enablement that's happened with a few technological trends. One is 8 or 9 years ago you'd hear about AJAX or responsive websites, that trend has continued to such a degree that you can instantly interact on any number of platforms at any time. The second reason is because the machinery that powers your website, Moore's law and the idea of these distributed data centers like Google and Facebook and Amazon, they almost practically free when you look at the cost of operating your website. You can throw a thousand computers at something very quickly now and very cheaply, so what's happening is, natural language processing is not a new thing, it's not a new technology, but what is new is that you can now through a thousand computers very cheaply at a website, and that means that you can have something that was not possible a few years ago other than the biggest research centers, just devoted to really any application. Instead of hiring additional staff, you just throw more computers at it, but the cost of computers is going down so quickly. I think that's what's driving this trend more than anything else.
Adam: Let's come back to the customer experience. You're a social media team for a big brand, you're not going to be able to have one-on-one conversations at scale, unless you have thousands and thousands of people. How do you see that relationship between, let's call it computer assisted or AI assisted customer experience evolving as chat bots become a bigger part of how customer bases are used to engaging with brands.
Matt: Yeah, any time a trend like this comes out, you have a lot of people who have a vested interest in saying it's going to completely change reality and, I think the reality is somewhere in the middle. There are going to be certain types of problems that are really well suited towards natural language processing. If you already have a call center, you already have FAQs, you're in a perfect position, and to very cheaply, because again, you're exchanging the price of these drastically decreasing in price computers. You can now automate away a huge majority of that work, again by just using all of these computers that you can put to work with machine learning.
Adam: I think that you also have to consider the better customer experience, too. I almost start there and say, If I'm calling in to a brand to solve one of my problems, there's already been something broken in the process.
Matt: There has been a tremendous amount of research ...
Adam: I don't want to talk to somebody.
Matt: There's been so many research reports that if you respond to someone within 10 seconds, a minute, the response rates that you get are so completely different.
Adam: The sentiment, all that's ...
Matt: Right, because everyone knows it's about addressing somebody when they have buying intent or when they have a certain intention, and again, the ability for computers now to roughly get that intent and provide them with either an answer or a set of very targeted answers, you get that person right when they want to know something, and an hour later they don't care. Or a day later, they absolutely don't care.
Adam: Right. Couple more questions. Where do you see bots and AI assisted customer interactions going in say the next 3 years.
Matt: Any time you use a really automated system, you'll eventually find, hey, here are the key things we're doing. I don't think natural language process is going to necessarily be the core of your experience down the line. I think over the next few years, you're going to see a great amount of applications coming out that are enabled by this very fast development process where you can iterate very quickly. You'll see many more new products come out, and then eventually you'll just kind of reach some steady state where there are certain types of things that are enabled by machine learning, but then a lot more that's driven by your own team and their intuition.
Adam: What advice would you have for a CMO at a Fortune 500 consumer brand?
Matt: The advice that I give, so I talk to people who are doing this at scale already at companies like Facebook and Google, they're advice is, small experiments, do a lot, because machine learning, it's learning. You learn by doing, you learn by getting in the field. You're cost of development and your speed of development is super fast. Get out there, start playing with it now, and because you're learning, you're developing that knowledge base and that kind of fluency with these types of application, and if you get that early, that could be a huge differentiator for your brand versus another one.
Adam: All right, well, Matt thanks so much for sitting down with us.
Matt: Awesome, man, thank you very much.
Adam: It was a pleasure, as always.
Adam: Stay tuned.
Matt: Thanks a lot.