2.1: How AI Can Help You Have More Human Conversations (Lauren Creedon)
Sammi Reinstein: Hey, this is Sammi Reinstein and you're listening to Conversation Starters. On this show we talk all about bringing conversations back to B2B marketing and selling, because if there's one thing we know about doing business in the revenue era, it's that the best customer experience wins. Through the power of our own conversations with drifters, customers, and special guests we'll learn how to deliver a sales and marketing experience that puts the buyer first. Let's get into it. Welcome, everyone, to season two of Conversation Starters.
Speaker 2: I'm so excited to be back with you, Sammi. It's been a few months, lots been going on, and I'm excited to cover it all and more this season.
Sammi Reinstein: Yes. Lots has been going on between season one and season two at Drift. We had a website relaunch.
Speaker 2: Yes, total rebrand.
Sammi Reinstein: Total rebrand.
Speaker 2: New colors, new fonts.
Sammi Reinstein: While that was going on, we launched three new products and we also had our first in- person event in over two years.
Speaker 2: Yeah, and those three products all came under our new platform name, The Conversation Cloud.
Sammi Reinstein: Yes. So you could say there was a thing or two going on in between season one and season two.
Speaker 2: While this podcast was not happening, we were keeping busy. Fear not, everyone.
Sammi Reinstein: I promise we were still working.
Speaker 2: Yes, but speaking of those launches, I already mentioned Conversation Cloud and Conversation Cloud includes Conversational Marketing Service and Conversational Sales, but it's all powered by the underlying Conversational AI.
Sammi Reinstein: Yes, and Conversational AI and artificial intelligence in general is something that some people are super familiar with, and for some people it's a little bit of a fuzzy term. It can be somewhat of a black box. So I'm really excited for our first guest of season two, Lauren Creedon, and she is going to explain a little bit more about artificial intelligence, how marketers can use it, how sales people can use it, and how to move it from that black box to really understanding what it is, and how you can use it.
Speaker 2: Yeah. You covered it all. So let's just dive right in.
Sammi Reinstein: Lauren, thank you so much for coming on the podcast. You're the first guest of season two.
Lauren Creedon: The pleasure's all mine. Thank you for having me.
Sammi Reinstein: Before we got on this, Elizabeth and I were actually looking through your LinkedIn and we were just wildly impressed. We were just chatting about you for like 15 minutes before we got on. So for everyone who maybe hasn't connected with Lauren on LinkedIn, which you should, can you give us a little bit of background and what it is that you do at Drift?
Lauren Creedon: You're too kind. Thank you so much. LinkedIn, oh, my gosh, it's a little bit of a resume, but at Drift, I lead the AI product strategy, and if we get into the specifics of what that means, we, at Drift, have a number of applications for marketing, sales, and service roles. My group of teams identifies the problems that we're uniquely positioned to solve with AI and try to find solutions to layer onto our existing products in a way that makes someone's job easier, helps them in a business outcome. So at Drift, the example would be chat bots. We have a Drift product to engage site visitors and my team comes in and layers on a combo AI tool within chat bots that helps you answer a question better with AI or get them to the right outcome. So that's an example of a product that would solve a user problem, and that's all on the product side, but a day in the life for me, where I have the most fun is split between engineering teams, go to market teams, and customers. Going back to that point about democratizing AI, I can't stress how important it is as a product person to spend more of your time with those customer teams and users, because I've always been drawn to business challenges. I've already always been drawn to innovation, but I had the most success when you stopped having it be about the idea and having a great idea and innovating and having it more focused on the problem.
Sammi Reinstein: Yeah.
Lauren Creedon: Finding a really complex problem, finding ways to experiment, to solve people's problems better. So I got into the AIML space at my last company, Huddle, just because I was drawn toward the most complex problem and I wanted to help people stop throwing things over the wall at each other, between product and go to market and instead speak the same language or figure out how to simplify something and really focus, because, yes, there's the user problems, but there's also all of us trying to do our own job and figuring out how that whole system could come together to hit an outcome has always really drawn me to product. So, at Huddle, we were a sports video software company. I like to share the stat that on Friday nights we had so many high school football teams uploading video to Huddle that it was one third of big peak upload traffic of YouTube. It was one of those really crazy fun spaces to be in, but the outcome really mattered. Then we started to build our first hardware product. So I got into that AIML space working on that ML powered smart camera that meant tackling some pretty gnarly problems across ball tracking software, live streaming video applications, or scaling video processing infrastructure, developing mobile apps to get the cameras self installed on the wall by a user using an app. So those types of problems drew me to the space. At Drift, its Conversational AI powered by NLP, which is its own new set of hairy problems. Yeah, to bring it home for me, it's less about the solution and more the fun of diving in, looking at a complex problem, and with anything, half that problem comes back to communication, comes back to teams. Those complex problems, those turnaround projects, those untapped potential areas all come down to simplifying, focusing, and super charging the team, the whole team, to help understand the problem, help communicate about the problem, and feel empowered to do something about it. So something like AI is classic. When it was just hand wavy, oh, AI's magic, but nobody really knows what we're talking about. That was only a small little group of people that were empowered to do anything with that, and users weren't even powered to do anything about it. So if we talk about making something more human, talk about human type solutions, and making our jobs all feel more human, I like to always come back to making the technology itself more approachable and simplified.
Sammi Reinstein: Yeah. I think that's a great approach with anything really, especially with such a complex subject like automation. I feel like we should have you come back at some point and do the ABCs of this space. It's like ML, AI, imagine a bunch of other things. I think with that, AI can feel, to some people, scary. To some people, it can feel kind of fuzzy. I think it can take on sort of a world of its own and maybe Hollywood has an impact in that and how they talk about artificial intelligence.
Lauren Creedon: Robots.
Sammi Reinstein: Yeah. Robots and things taking over and all of that, but can you give your best definition of what artificial intelligence means specifically in the sales and marketing space?
Lauren Creedon: Sure. Yeah. So let's start by talking about why you'd ever want to hire AI to do a job over a human in this space.
Sammi Reinstein: Yeah.
Lauren Creedon: So you usually want a machine to do something that's either super mundane or super complex, and so mundane tasks like syncing data fields helps a human spend more time with other humans, or complex functions like understanding language at scale or making predictions or aggregating themes lets humans focus on making the decisions or taking actions on those themes, the things that humans actually like to do and are better at. So let's break it down a little bit more. Those types of complex functions I talk about, I like to think about it in two sides. There's the language side. When you hear NLP and NLU, natural language processing and understanding, and the structured data side, which is more predictive modeling and doing fun stuff with math to help you make a prediction, but on the language side you have what we do at Drift, which is conversational AI products. Help you understand what people say. So you often think about voice activated devices like Alexa or those phone systems where they recognize your voice, but at Drift, we have a chat bot and you want to understand, in this specific example, train a model to understand the ways your buyers ask about a topic and it can get better at recognizing other ways buyers could say that. Then let's get another specific example over on the structured data side. You might think about products that help you make predictions, something like a lead scoring product that's building a predictive model off a number of factors and then applying those factors back to new inputs to continue to learn. So to demystify what's going on with AIML, on either side, I just think about the term machine learning. A lot of people find that's a more demystifying term because what you're doing with ML is teaching a software application to become more accurate at predicting something, whether it's understanding what somebody said or the predictive outcome without that explicit rules based if then programming.
Sammi Reinstein: Right.
Lauren Creedon: Which scales a whole lot better.
Sammi Reinstein: Yeah. Yeah. You kind of have started to scratch the surface here, but a lot of what Drift is about is starting conversations and appealing to the human in other humans, right? We're not selling to just this corporate buttoned up business. We're selling to another human on the other line of that chat or email or video, and artificial intelligence is inherently not human. So can you help explain that dichotomy and how AI can appeal to that human and start more conversations?
Lauren Creedon: Yeah, and I appreciate you calling that out. I think we've learned even from launching AI products that were a little bit more in a black box, that weren't as clear how it worked, that marketers resisted it a little bit, as they should. If they didn't have control or visibility, it's hard to build trust. It's hard to put that AI out on the front lines with your customers when you don't know what it's going to say or don't feel like you can do something when you see a bad experience happen. It's really visceral when you see that. So while it's not human, I think any marketer could say those mundane tests or those complex scaling problems of helping buyers answer questions in the middle of the night, there's problems you might want to solve that humans aren't ready to solve. So if you can train AI to do that for you for a specific job that you need to get done and you have the control and visibility into that, then that's where you start to unlock the things that are more human. So if you think about being a human as part of a B2B buying interaction, it's kind of like the consumer space has trained us to always want to feel like a human. That means being in the driver's seat. Get what we need when we need it, even if it's at night after hours, after we put the kids to bed, and you want to just answer a question, because you're still in your consideration stage, not a decision stage yet, and you want to just answer a question. So to set a data point, I think it's 83% of customers are willing to share that data, willing to have conversations with touchpoints even before interacting with a human to enable future personalized experiences. I think what's critical about that is when we do, we now expect to get a better experience back. If you're giving that company data or you're sharing what it is you're looking for in the form of a free text question on a website, you expect that the company's going to then give you a personalized experience back, especially when we connect with a human. So there's a couple different examples there of just serving up a question when somebody needs it in that one- on- one example, but then also saving that context and serving it back up to the humans on your team at the right point so that you can really have a customer- centric experience through the full life cycle, not just that first personalization in the email.
Sammi Reinstein: Yeah. I think the key there, and something that you've always been very aware of, is that it is creating that more customer centric experience. It is keeping the buyer in mind and trying to appeal to where they are. It's a frustrating experience when you go somewhere and you're repeated something or they, I don't know, don't call you by your name, right? Just those little things. It is appealing to the customer side and that customer centricity is really important. Speaking of customer centricity, you talked a little bit about the black box and how we wanted to build a little bit more trust and we're trying to bring AI out of that black box. Can you tell us a little bit more about what you mean of bringing AI out of the black box and we can get a little Drifty here and talk about what that means at Drift.
Lauren Creedon: Absolutely. Yeah. So I think at first, let's just start with the language. So we've stopped talking about AI as a hand wavy magic and literally put the controls in the hands of our users to both train the model on what it means to understand its buyer saying and how to respond in its brand voice and what's going to happen once it understands that topic. So I can talk about that product in our chat bots. I can also talk about our insights product, but before I get too much into the product, it's also about stopping using complex terms that made it hard to understand, hard to sell, hard to help solve customer problems with. We've oversimplified our language and we've found that all helps anybody do their job better when you understand what's going on under the hood, what specific job a customer is hiring AI to do for them, and can measure and trust whether it's doing that job. So for example, our chat bots, we had an AI powered chat bot that was purely trained to handle the conversations unstructured on its own and the customer didn't have any visibility into what it was going to do next. They could submit how it would reply, but needed to trust the bot to make a lot of those decisions on its own. We found that just the right balance of, one, hybridizing where and when customers want to speak in open text and marketers can provide a good experience back with open text, helps marketers feel more in control of that experience and buyers actually get it, get the experience that the marketer's looking to help them get. Then, two, helping them decide what outcomes they want to direct people to so that they get to the right human on their team. So something as simple as when somebody types a question in free text, we can serve them up and answer because we've understood what they've said, but that we can also give the marketer the tool to say," Branch them actually over to this other team, because this is a sales related question. Their high intent. I don't want to send them through a lead qualification flow because they just asked me a question about pricing. They're in decision stage, let me connect them to a member of my sales team." Or if you're exploring a really in depth technical question, and you're already a customer, you also don't want to be sent down a lead qualification flow, asked what your email is. You want to be treated like a customer, knowing exactly where you are in the life cycle and be connected to your customer team. So we've given the tools to marketers to set what rules happen after we understand something and also given them the tools to identify what themes are coming up, how buyers are actually asking questions so that your model gets smarter at responding to that and you can add more of those signals to where your conversation paths would go.
Sammi Reinstein: Yeah, and a lot of that trust building exercise, like you said, it was finding the language that they're speaking and connecting it back to their goals and their priorities. I think that's a really important lesson in all, really, go to market strategies. When you're introducing something new, building that trust and making sure that you're connecting to the inaudible the what's in it for me, but also what's in it for their buyers and how you can create that experience.
Lauren Creedon: Well, trust us never works.
Sammi Reinstein: Right.
Lauren Creedon: At Drift, we lead with," Tell us what outcomes you want to measure by. We'll help you hit those outcomes." The whole customer engagement, ties back to those outcomes and helping them hit that and it's all about the data and the tools to adjust along the way. So AI has to fit into that too. It can't just be like," Trust us," because AI isn't better without that human input, without really learning the voice of your customer. We'd like to just help you get the data for the humans to make the decisions.
Sammi Reinstein: Yeah. I'm just chuckling to myself thinking of us going to AI and being like," Automation, trust us."
Lauren Creedon: Nope. It's not going to work.
Sammi Reinstein: Speaking of those goals, something that marketers are very obsessed with, and sales people, is getting insights into how can I make this better? How can I optimize this? What language are my customers speaking? How can I speak their language? Something that automation provides is insights. Can you tell us a little bit more about how marketers can use those insights and action on them?
Lauren Creedon: So if you think about marketers day to day, they have a million data sources and there's a lot of sophisticated tools out there to help them use that data, to provide a better buying experience, a better buyer journey, but one thing that's always been really hard to gather in aggregate is the voice of the customer in their own words. There's a lot of noise and signals that marketers pick up, forced to make assumptions about what customers want. Like," Okay, they opened my email. They clicked on the link. They're higher intent," but what if you're just somebody who's doing your own research about how somebody's marketing a company and you have no high intent at all? So we often see marketers assuming intent and what a customer is interested in, but what about what they've actually said in chat, in an email, in a meeting? That's where the real juicy insights are.
Sammi Reinstein: How can AI help marketers understand their customers better?
Lauren Creedon: As for those juicy insights I referenced, we like to call it exact intent rather than approximate intent. So if you want to take action on the voice of the customer, understand what someone's saying and get them what they need, you can do that one to one in a chat conversation. You can also do it to understand aggregate themes across all conversations. Without AI, you need to read through those conversations or guess at keywords to search and then read them or wait for someone to forward you an email with a juicy quote and then extrapolate like," All customers must feel this way," but with AI, it helps you be unbiased about things like what questions your buyers are actually asking the most, and use those insights to deliver back better experiences, or help them get educated on what they're actually looking for, make decisions faster. That can extend to the site experience or what help articles you need or any number of things through that buying journey. Ultimately, it's all about helping people make good decisions faster.
Sammi Reinstein: Yeah.
Lauren Creedon: Helping customers solve that pain point.
Sammi Reinstein: Yeah. Like you had mentioned, the bias can sort of creep in. We mentioned this a little earlier, Lauren and I were chatting. Ego can creep in a little bit. I know that sometimes I think I have an amazing idea and I've done this in Drift too. I think that this hook is going to work so great. Then I launch it and, the data tells me otherwise, and insight sort of helps you get ahead of that. It's telling you what your customers are saying, and you can use that to appeal to them, and that also goes with empathy, which we were talking about earlier. Lauren and I, if you can't tell, we were just chatting up a storm before we started this podcast, because I love to just pick her brain and learn a little bit more about where her mind's at, but the empathy portion too, marketers don't want to have a bad conversation. They don't want to provide a bad experience and everyone knows what that is like. So being able to remove that feels really good and you're also working towards your goal. So it's a win- win.
Lauren Creedon: I agree. I agree. When you see a bad conversation, you want to know why the AI model identified it wrong or what to do about it, or you want to, if you hear a lot of customers asking about a particular thing and AI's bubbling that up to you, like," Hey, a bunch of people are asking this question. You don't have a topic for it. Do you want to add one?" We're trying to think of those little granular jobs that AI can help do that just put a marketer's mind at ease because ultimately it is about providing that better buyer experience and helping free up the humans on your team to spend more of their time on those decision stage conversations where you really do need to speak to a human.
Sammi Reinstein: Yeah, and we've been talking a lot about marketing, but AI really is applicable to a lot of different types of roles. One of those being service, another being sales. How does AI help sales reps have better conversations?
Lauren Creedon: Great question. It all comes back to remembering context and a little bit of that reducing the mundane tasks, but if you think about a real pain point these days, that we all feel in our jobs, it's how many people are involved in both sides of a B2B deal. From the early stages with even the marketing materials and the value perhaps that have been created in that early product development, to then a customer speaking with a BDR, through to an account executive, and solutions consultants, and AM might take over the account after a decision, and eventually a CSM, or a renewal specialist, and roles are switching all the time. Then on the buyer side, there's just as many, if not more, stakeholders on that buying committee, making a decision, really hiring a product to get a job done. So in an age where personalization matters or getting those questions or driving alignment, having that clearer communication, it's important to have that context from past conversations at a glance summarized. So that in that little bit of context, switching, you have to prep for a meeting, you can pick right up where you left off and then extend it through the life cycle. Net retention, we were talking at our customer advisory board, net retention matters more as the metric that aligns a whole organization. That happens when a customer gets what they bought and gets more of what they bought and is provided the right maturity opportunities to grow with that product. So in an age where it's easier than ever to switch to a different solution if it solves your problem better and more internal champions lead for other roles, you need to make sure as a B2B business, you have that context from those early buying conversations to make sure that customer is always at the center, full life cycle, solving the pain points that matter to them in the beginning.
Sammi Reinstein: If you had a crystal ball and you could see the future of artificial intelligence and what that looks like for sales people and for marketers, what do you think that crystal ball would tell you?
Lauren Creedon: What would it tell me? It would be something that takes it out of the black box. So it helped me trust the data. It would help me know why the machine made decisions that it did and how I could improve it, whether I'm a marketer who wants to train a model to understand or respond differently, or if I'm a sales rep seeing content that I don't find relevant or a tag from a conversation that I feel like is inaccurate and I can give that feedback loop. It's helping everyone feel like they have a little bit of control and can customize the outcome. It's giving visibility. If you have an engagement score or lead score, it's like," Okay, why is it four flames? What are the factors that went into that?" Then I also envision a world where we can all ask more questions and get answers that help us go," Aha. I get it now I get what type," because then we can ask for the types of pain points we want to be solved with AI. I think today it's very hand wavy what we hope AI could do for us and if we got a little clearer on what it was, and we as marketers and people who are building the products themselves got clearer on what types of jobs AI does for us, then people feel more empowered to ask AI to do something for them and speeds up the whole product development process too.
Sammi Reinstein: Thank you so much for coming on the podcast. It's been a pleasure talking with you and I'm so glad that you are our first guest of season two, you're setting the bar very high. Where can people connect with you to learn a little bit more about artificial intelligence?
Lauren Creedon: Yeah. LinkedIn, Lauren Creedon on LinkedIn. Send me a message. We can talk about AI. We can talk about product. We can talk about anything, career stuff. I love to mentor women in product and help other people realize you don't have to know a ton about AI to start learning and even get a job in it. So yeah. Open door.
Sammi Reinstein: Thank you so much, Lauren.
Speaker 2: You know, working at Drift and on the content team, I hear a lot about AI. I know it to an extent, but I really appreciate the way Lauren just broke that down in very clear steps. Especially when she talks about how AI isn't replacing humans and how it can be human and actually make marketing more human.
Sammi Reinstein: Yeah. Yeah. I love the way that Lauren communicates and she talked about this a little bit on the episode, but communication and building trust is all about being able to break things down and speak simply about a topic. I think she did just that. So I hope you learned a lot and if you are looking to learn a little bit more about Drift and Drift's conversation AI, it is in the platform section on drift. com and we will link that in the show notes.
Speaker 2: We'll see you next week for another episode about how you can make sure that those first meeting conversations are personalized and efficient.
Sammi Reinstein: Yeah. Season one was all about starting the conversation. Season two, we're continuing it. So see you there. Thanks so much for listening to Conversation Starters. If you like this episode, please leave us a six star review by clicking the link in the show notes and hit subscribe so you never miss another one. You can connect with me on Twitter at Sammi Reinstein and follow all of our shows at Drift Podcast.
We talk a lot about making conversations between marketing and salespeople feel more human on this podcast. But, artificial intelligence (AI) is inherently not human. So, why then, are we hearing all this talk about the importance of integrating artificial intelligence into marketing and sales strategies?
To kick off Season 2 of Conversation Starters, Lauren Creedon, Drift's Group AI Product Lead, joins Sammi to explain the dichotomy of leveraging artificial intelligence to be more human in marketing and sales conversations.
- (2:49) What Lauren does at Drift and how she got into AI
- (7:43) What artificial intelligence means in the sales and marketing space
- (10:31) The dichotomy of using AI to be more human
- (13:49) What it means to bring AI out of the “black box”
- (18:24) How marketers can leverage insights from AI
- (19:21) How AI can help marketers understand their customers better
- (22:34) How AI can help sales reps have better conversations
- (24:40) Lauren’s predictions for the future of artificial intelligence
Like this episode? Let us know by leaving a review!