Fans of “Westworld”, HBO’s dystopian science fiction drama, recognize the understated yet outsized role of Lee Sizemore’s character as crucial to the series plot. As head of the Narrative and Design division of Westworld’s Mesa Hub, Lee is responsible for the storylines of the “hosts”. The scripts he creates for them ultimately determine the experience of Westworld’s guests, and whether or not they’ll become repeat customers. Lee Sizemore’s futuristic-sounding job is somewhat paralleled in the present-day by people who architect conversations for AI chatbots. It’s a field of growing importance that’s crucial to the outcomes generated by conversational AI, and ultimately, digital transformations.
For insights on the current state of this craft, we turn to one of its more notable practitioners, Rebecca Clyde, CEO & Co-Founder of Botco.ai. Rebecca shares many of her unique insights with us, and along the way we learn how early adopters of conversational AI are experiencing triple-digit percentage improvements in their sales conversion rates, how technology can make AI-driven chat conversations more engaging for people, and why taking the ontology learning approach to natural language processing might ultimately be better long term than machine learning.
Guy Nadivi:Welcome, everyone. My name is Guy Nadivi and I’m the host of Intelligent Automation Radio. Our guest on today’s episode is Rebecca Clyde, CEO and Co-Founder of Botco.ai, a conversational AI platform enabling meaningful and intelligent conversations between businesses and their customers. Prior to co-founding Botco.ai, Rebecca was with Intel for many years. In July of 2018, the Phoenix Business Journal chose her as one of that city’s most admired leaders, and in 2016, the Greater Phoenix Chamber of Commerce honored Rebecca as its Athena Businesswoman of the Year.
Guy Nadivi: Rebecca is not only very accomplished, but she also has a lot of insight to share about conversational AI, which is becoming increasingly important to organizations worldwide in light of the global COVID-19 pandemic. Rebecca is also the latest woman we’re featuring for our Women in Tech series, which profiles some of the outstanding women helping to shape the technology industry. Rebecca, welcome to Intelligent Automation Radio.
Rebecca Clyde: Thank you, Guy. It’s a pleasure to be here with you today and to participate on this podcast.
Guy Nadivi: Rebecca, at the Fuel conference in Belgium a couple of years ago, you gave a presentation listing some of the many advanced ways organizations were already using chatbots, automation, and AI, to empower their customers with an always-on channel for self-service, and a way for them to interface with the organization. For organizations that have already implemented conversational AI for their customers, how has that given them a competitive advantage today?
Rebecca Clyde: That’s a great question. Thank you, Guy. One of the things we’re seeing our customers benefiting from intelligent chat nurturing, is that they’re able to shorten their buying cycle, the buying cycle for their customers, because they’re cutting back this typical back and forth that goes on, phone tag, that is so common in the buying process. So by being able to eliminate that, our customers are noticing that they can close a sale much more quickly, they’re doing so in a much more effective manner, and they’re also converting more of their customers that either are coming to their website or coming to their different digital channels.
Rebecca Clyde: Our customers are reporting even 100 to 103% improvement in those conversion rates, which of course means that their top line is growing as a result. So that’s one of the biggest advantages that we’re seeing. Next, we’re also noticing that our customers using this technology, because they have analytics and they can dive into the types of questions that customers are asking directly on these chat channels. They’re able to glean really, really helpful insights that are useful to inform business strategy. And we’ve even had customers that have introduced entire new product lines based on that information. So not only are they helping customers faster and making them happier and selling more, but now they’re also understanding their customers better and knowing what those customers want to help them inform new products and new experiences but again, go back to driving that bottom line. And then the last thing, this one might sound a little bit more fun, but a lot of our customers are also getting awards and recognition from their peers and from their industry groups. And that’s also great for those businesses, right? So they’re doing good for their customers and they’re also helping to elevate their own brand and their companies within their industries.
Guy Nadivi: There are some concerns about biases creeping into the AI that powers things like chatbots. When you’re using content or some database to train the AI and machine learning that will power a chatbot, how can you ensure that biases don’t contaminate the curriculum, so to speak, and end up corrupting the chatbot?
Rebecca Clyde: Yeah, that’s a very common question and it’s definitely something that businesses should look at closely. We always recommend that our customers closely examine all of their content sources. So is this the latest information, is the data correct, and does it actually reflect the best foot forward that your business has? And then we also recommend testing. So when we have these inputs, what do the outputs look like, and is that really what you want? So this whole idea of just blindly throwing in data and then seeing what comes out on the other end, you really can’t do that. You have to test it and then if something is coming out on the other side that you don’t like, well maybe there was an incorrect piece of data or some source content or material that’s getting confused or creating this problem.
Rebecca Clyde: And so we always say look carefully at your content sources, look carefully at your data sources and bring in a lot of different points of view from the organization. So don’t just have marketing look at it. Have somebody that’s representing the customer look at it, have somebody that is representing the industry look at it, have somebody from legal. Have a lot of those different points of view. And I know that it takes a little bit longer to do that, but in the end the result is going to be better.
Rebecca Clyde: And this is also where organizations that have diversity built in to the team can also bring those diverse points of view to make sure that you’re not missing something important in this process. Yeah, so that’s really what we recommend. It’s like you have to be really crisp and clear and accurate in terms of what you’re putting in as your data sources, and if those are not right then you can’t possibly expect the output to be great.
Guy Nadivi: As sort of a reality check in order to root out bias, Rebecca, do you think AI algorithms should be audited the same way that financial statements are for publicly-traded firms?
Rebecca Clyde: Yeah. You know, it took us a while I think to get to that point in the accounting world, and in order for that to work you would need to have some kind of a standard that everybody is aware of and is compliant to. You would need some kind of standards body that enforces that standard. And part of the challenge we have right now is that no such standard exists and no such body exists. And so without that it becomes very difficult to audit in a way that’s consistent. I’m always a big fan of transparency for my own business, but not all companies operate from that point of view.
Rebecca Clyde: And then of course the other thing you’d have to think about is would Google and Facebook allow that auditing to take place? And I often think of, it would require Congressional regulation to get companies like that to comply with any kind of public auditing. So it’s a much bigger question that I think as an industry have to ask ourselves, and if we are all in agreement that this needs to happen then we need to get even the Googles, the Facebooks, the Amazons to agree and participate. Otherwise it’s not going to work.
Guy Nadivi: You’ve stated that in addition to computer scientists, programmers, and other technology people, the rush to build AI-driven chatbots requires poets, writers, artists, and others that can infuse a humanist element into the chatbots. Now, not to cast aspersions on people who majored in liberal arts degrees during college, but will the occupation of “Chatbot Conversational Architect” perhaps be a higher-paying career for them over the long run?
Rebecca Clyde: Yeah. Actually I do see a lot of job postings for that role nowadays and I’m sure they’re getting paid well. Writing a conversation for a virtual agent is a lot like writing a screenplay. You have to keep it interesting, you have to keep it engaging, and what makes the conversation really intelligent is not just the immediacy or the correctness of the answer, but it’s also the cleverness of the answer or the empathy that is conveyed in the response. That’s what keeps the conversation engaging. And so without humans to bring in that extra color or dimension, these conversations would just fall flat and be very quickly exited. So we do like to think about how we can introduce that human component to some degree in every conversational interaction.
Guy Nadivi: Now, speaking of that, you’ve talked about building emotional intelligence into chatbots, which of course will move them a step closer to being conversationally indistinguishable from humans. Do we really want that?
Rebecca Clyde: As long as we’re always disclosing to the parties involved that this conversation is being handled by a virtual agent, I think that’s fine. We always advise our customers to make that clear at the beginning of the conversation, in the middle, and at the end. So the person is always very aware of the fact that they are conversing with a virtual entity, not with another human. But as long as that’s happening, I mean, doesn’t everyone want to feel understood and listened to in a conversation? There’s some really great ways that technology can do this just as well or even in some cases better than a person. And that’s by using techniques like mirroring or contextual references, being able to reference something that had been mentioned previously and even providing an appropriate level of personalization. So those are the mix of things that you can introduce to an intelligent chat conversation that really make it more engaging, make it flow more nicely, and then ultimately are more helpful to getting the results that the person is wanting.
Guy Nadivi: Good point. In broad terms, what are some of the lowest-hanging fruit that are best suited for conversational AI applications within an organization?
Rebecca Clyde: I’m always a big proponent of customer engagement because I come from a marketing background, and so from my point of view the brand and that interaction with a new prospective customer is always one of the most important ones. So how can you use intelligent chat to really draw in a new customer, help them find what they’re needing quickly, answer their questions in a helpful manner, do all of those things? That’s always my favorite place to start. But there’s a lot of other low-hanging fruit opportunities out there, and some of those might be around handling questions about employee benefits for large companies or maybe, tech support is another really great one, whether it’s for internal IT support or for your product. So there’s a lot of different places where conversational agents can be very helpful. My favorites, of course, always have to do with engaging new customers, though.
Guy Nadivi: Rebecca, you argued that our on-demand economy has driven the shift towards customers wanting to use chatbots rather than go visit an organization’s website. Now in 2017, Gartner predicted that, “By 2020,” this year, “40% of all mobile interactions will be via virtual assistance.” I don’t think we’ve quite hit that forecast. So let me ask you, when do you predict we’ll reach peak app, as it were, and start transitioning to conversational AI as the predominant user interface going forward?
Rebecca Clyde: Well I can tell you that in my household it already is the predominant user interface, and it’s not just because of me. I’m actually not the one driving it, it’s my kids. So my children that are in their teens right now, that is their number one way of interacting. They’re more likely to ask Alexa or Siri or Google Home a question than to go to a website any day. And this generation, which we call Generation Z, they’re already starting to graduate from college, they’re entering the workforce, they’re becoming a legitimate part of the economy, and I really see this practice, this behavior coming with that wave. And as they emerge into the workforce and become a bigger part of the buying power of the economy, we will see businesses really start to adapt more to that favored way of interaction. And so it’s coming. This generation, within the next five years are going to be well into their twenties, making a decent amount of money and contributing in a big way to the economy. So we better be paying attention if we want to be selling them things like cars and homes and mortgages, all of those things. We better be paying attention.
Guy Nadivi: Interesting. Since you’ve got your crystal ball out, what are some of your predictions for conversational AI over the next three to five years?
Rebecca Clyde: I really see intelligent chat becoming essential to how businesses function. Just as websites became ubiquitous, social media became ubiquitous for businesses, we’ll really go from asking should I have this to why don’t we already have this?
Guy Nadivi: Rebecca, I’m sure many of our female listeners would love to hear from you about what your experience has been like as a woman and an executive in the technology field.
Rebecca Clyde: Yeah. So it’s brought its challenges with it. When I first entered the work force in the late, late nineties right before the dot-com boom there were a lot of women that were emerging in technology and so I had some role models during that time. But what I noticed is that it was kind of like there were a few at the top and then there was not a lot in the middle. Within a big company, it was often sometimes difficult to find female mentors that could really show me the way forward in this kind of new generation. They had older techniques that had worked for them, maybe when women were really trying to emulate men a lot more, but if you were a woman that didn’t want to emulate men but you still wanted to move ahead in your career, there weren’t a lot of acceptable ways to do that. And so I think for a lot of women in my generation kind of trying to navigate those waters has been challenging at times, but we have figured out our way and for many of us, including myself, it really meant leaving a lot of those corporate structures and starting our own businesses in order to create a workplace that really worked for us. I know that that happened to me and to a lot of women that are in my kind of age group.
Rebecca Clyde: And so now what I tell a lot of younger women that are graduating from college now is I say, “Look, go see if corporate America has the answers for you, and if they’re willing to support you the way that you need to, but if they don’t, don’t stay there. Go on your own journey because it’s not worth giving up your entire career if the corporate structure isn’t going to support you the way that you need to be supported.”
Guy Nadivi: Do you think there are any advantages to being a woman in the technology field?
Rebecca Clyde: Oh, huge advantages. Absolutely. I think the biggest one is just simply having a different point of view and a different perspective than everybody else in the room. I’ll often think about how a product would impact my children or would impact… Even form factor is important in technology. When I worked at Intel, I was always pushing for smaller devices or for mobile because women don’t like to carry heavy briefcases. We like to carry small purses or have things that fit in our back pocket. And so we’re always trying to drive the technology to be smaller because we have smaller things that we want to carry around. We don’t want to carry big bulky computers and big bulky phones around. So you have that point of view, you have kind of, I think we tend to think, at least I do, because I have kids, I always think about things in terms of how they would affect my own children or how they will affect even simple things like their developmental brain, their brain is currently developing.
Rebecca Clyde: So I might think about things differently and I also might think about it in terms of lifestyle. Like how does this affect a woman like me who maybe has a different schedule or a different kind of lifestyle that I’m balancing in terms of work and family and home and cooking? I’m just as likely to be on a conference call and running an investor meeting as I am making dinner and baking cookies. Those two things are not mutually exclusive in my life. And so I just think about things more holistically probably, and I think a lot of women are that way too. So yeah, for me it’s been a huge advantage and it allows me to fuse my projects and my work with a lot of creativity and the stuff of life, you know?
Guy Nadivi: Yeah. I like that, a more holistic approach. Very interesting. Rebecca, for enterprise IT managers who have never dealt with conversational AI, what should they know before deploying it?
Rebecca Clyde: So the first thing I always tell people is understand what is the business objective and what is the goal from an organizational standpoint? Because if that is not perfectly clear, then chasing a technology, no matter how great the technology is, won’t solve anything. So I always try to back up, because even my customers will come to me and say like, “Oh, this is so cool. We want to do this.” And I’m like, “Okay, that’s great, but help me understand what business challenge you’re solving. How is this going to help your business and what is really the measure of success here to your bottom line?” And if we can’t articulate that together, then I just say, “Don’t move this forward, because otherwise we’re all going to fail together.” So first and foremost, it’s that.
Rebecca Clyde: And then the second thing is I always find that a lot of organizations, they may not have the expertise internally for this technology, but they think that they should still build it themselves. And I always say that that’s a little bit of a mistake. So if you don’t already have expertise within your organization and it’s not core to your IP, then you shouldn’t be building it in-house. Really, partner with an expert, bring somebody from the outside in that can really show you the way you can bring you lessons learned from other projects and other failures. And that way you can build on what’s already been learned out there and not start from scratch. So those are two pieces of advice I would always give.
Guy Nadivi: And, Rebecca, for the CIOs, CTOs and other IT executives listening in, what is the one big must-have piece of advice you’d like them to take away from our discussion with regards to implementing conversational AI at their enterprise?
Rebecca Clyde: Yeah. So I would say spend some time with the people who actually understand this topic. I’m a big fan of the ontology natural language processing approach because it yields better, longer-term results. And that’s just not my own finding, it’s actually kind of widely accepted in the research and the academic community as well. We find that a lot of businesses will go down a more traditional kind of machine learning path, but that path has its limitations, and even though it’s easier to go down that path at the beginning, it also corners you into some tighter spaces at the end. And we see that a lot of businesses struggle to get long-term results that way. So from a purely technology point of view, I would say explore that ontology-based approach. There’s lots of really good literature on that. In fact, my Chief Science Advisor for our business, Deborah McGuinness, is one of the leading authors on this topic and she has a lot of great content that she’s produced that’s available online. She has a book out on this topic. You can actually learn about this from the experts, and I would say really evaluate that approach. At least give it some thought before moving forward, because I think it could save you a lot of headache in the end.
Guy Nadivi: Great advice. All right, looks like that’s all the time we have for on this episode of Intelligent Automation Radio. Rebecca, there is so much going on in the world right now with the pandemic, the demand for conversational AI solutions, and of course you’re also still a wife and a busy mother and a homeschooling teacher now. So given all that, we greatly appreciate you coming onto the show and providing some really valuable insights for us.
Rebecca Clyde: Thanks, Guy, for having me. I really appreciate it and it’s been a great afternoon conversing with you. Thank you.
Guy Nadivi: Rebecca Clyde, CEO and Co-Founder of Botco.ai. Thank you for listening, everyone, and remember, don’t hesitate, automate.