Insights from the Chatbot Conference Featuring Trace Hansen
Podcast: Digital Conversations with Billy Bateman
Host: Billy Bateman – Billy Bateman is the co-founder and VP of Operations for ChatFunnels, the digital conversation analytics, and optimization solutions provider. He has a background in digital marketing, operations, and entrepreneurship. Billy grew up in Idaho and graduated from Brigham Young University and has a Masters of Business Administration from Boise State University. While at BSU, he completed Boise State’s Venture College Entrepreneurship Program and became a mentor and entrepreneurial lead on a technology commercialization project. He is an avid outdoorsman and enjoys fishing and hunting. He currently lives in Springville, Utah, with his wife, Chelsea.
Guest: Trace Hansen – Trace Hansen is the Engagement Director at ChatFunnels, the digital conversation analytics, and optimization solutions provider. Trace has experience in operations, marketing and entrepreneurship having started and run several businesses. He has a Bachelors of Science Degree in Communications and a Masters in Technology and Education from Utah State University.
Overview: Trace shares some of his takeaways from the San Francisco Chatbot Conference held September 25-26. He gives insights on the importance of monitoring and improving bots and some tips on how to improve copy.
Billy: Alright everyone welcome to digital conversations. Today I am joined by our very own Trace Hansen who runs our consulting team. He just got back from the chatbot conference in San Francisco and just wanted to talk to him about a few of his takeaways and things that he saw in the industry while he was there, Trace thanks for hopping on today.
Trace: No problem at all thanks a bunch Billy. This was an awesome trip, absolutely loved it, great info, learned a ton, it was awesome to see the different companies that represented the industry, I mean we had huge companies like Adobe that was there. Sams Club was actually there believe it or not, Rasa, Rulai, other great companies that are leading the industry in the conversational IA landscape so it was really cool.
Billy: Awesome man. So coming out of there what were your one to two big takeaways after the whole thing was done?
Trace: Sure. I think one of the biggest things to me was that chatbots and AI are so pervasive. I mean they can be implemented in hundreds or thousands of industries, not just tech companies. Initially the first few companies were you know tech companies that were really excited about using chatbots and utilizing those. They can be implemented in hundreds of different industries. We saw examples of our hospitality industry we see utilizing chatbots, retail companies, B2B sales companies. One specific example is a university that was actually using chatbots to solicit qualified students to apply for their school. The problem that they were having, that they were trying to solve was two fold, number one they wanted to get qualified applicants and number two they wanted to reduce the universities operational cost per application that was submitted and so the bot creation project took about 2 weeks to put together and then they tested it over a three month period to see what ended up happening and how things changed. When they were able to get it up and running, 3 months later it reduced their cost per application by about 62% for that university which is just incredible. So I honestly feel like you’d be hard pressed to find any industry that can’t be positively impacted by chatbots and AI.
Billy: Yeah, I mean if you’ve got universities, which are usually slow to adopt any new technology, using it it’s gotta be everywhere which is what were seeing.
Trace: Yeah its incredible all of the different areas and spaces that chatbots and AI are getting into so it’s definitely fun to see that.
Billy: One other thing I wanted to ask about, so whenever I go to one of these conferences these always two of three people that are just doing things kind of on the cutting edge that I didn’t know about before going there. So are there any new technologies or companies that you saw doing interesting things you didn’t know about?
Trace: Yeah of course there were several, it was really neat. So Smartloop was actually one of the ones that impressed me quite a bit. Its a company that’s taking a great approach to utilizing facebook messenger based bots. So their software helps you to build your own chatbots. There kinda like the square space for chatbots. They have point and click bots so you really don’t have to know any code at all to be able to deploy these, its one click to deployment. They offer push notifications through facebook messengers, and the companies can also utilize their software to transition from a bot user experience to live chat as well. So they are really pioneering the facebook messenger space and they are doing a great job with it, so it was fun to watch them. Another company called VOIQ is a company that utilizes IA voice bots for business calls. This was a space that is growing quite rapidly, they use voice recognition in IA instead of the old IVP system where, I mean we’ve all been on these calls where you call into a utility company and they say “press one for this, press two for this, press three for this” and you get lost in the labyrinth of all of the different options that are there and it’s kind of hard to recover from that. You end up calling in 5 or 6 times. This company is really done a great job utilizing AI so that the user experience is a pleasant one instead of just an annoying one. Most of the time when there are options for press 1, press 2, press 3 people usually press 0 anyway just to talk to a human being. And so streamlining that has been a really pivotal part for VOIQ. They’ve done a great job with it.
Billy: Awesome man. Yeah nobody loves those like press one press two but if you can make it easier that’s gonna change the game a little bit. So what else did you learn while you were there? Anything really interesting like big takeaways for how people that are using bots should be, what approaches they should take to their process?
Trace: Yeah absolutely. So one of the critical takeaways that I found was something that we’ve utilized at our company, Chatfunnels quite a bit but it was cool to see it reiterated within the industry as well, is how critical it is to monitor an improve bots on your website on a regular basis. There’s a lot of important steps in accomplishing this and so just a few of these steps are really important for us to keep in mind as were monitoring these. First of all is to really evaluate the real conversations that you are your customers are having with your current bot configuration. So on a regular basis get into those conversations and see how things are going, see if the conversation flow is going well. Step two, go back to the bot and adjust based on any flag issues that you saw as you were going through those conversations. Some of those things to look for or some of those flags could be attrition points where you see a lot of users bailing on the conversation at certain points, it might be that the bot is simply saying something in a way that doesn’t seem natural and so the person would just rather go to the website and find the information that their looking for orthey just bail completely. Other things to look for are tags that aren’t working functions that aren’t functioning as they should on capturing data. So thats step two. Step three you wanna make sure to test the configuration model after you’ve made some adjustments. Always always always test things before you hop into them and set them alive. The next part would be to reevaluate don’t just set it live and let it go. Is there anything else that can be improved before we actually deploy this, is an important part. And then deploy the new model if it passes your set metrics and scores that you’re looking for. And so that was a really important thing to keep in mind, is the need to continually maintain bots. I think the feeling in a lot of companies that have employed chatbots just set it and forget it. They want it to be the slow cooker where you just don’t have to worry about it anymore for a while and that continual maintenance is critical to maintaining a consistent voice and message and being able to actually speak in an effective way to your target audience. So another big takeaway for me was something again that we stress pretty intensely at chatfunnels. Its to test test test again. We need to make sure that we are running experiments consistently. Performance metrics whether the metrics you’re looking for are emails captured, meetings booked people enrolled in a subscription plan . Whatever it may be, we want to make sure we are always testing and using experiments to provide data for what we are trying to accomplish with these bots. It may be that there is some copy optimization that you could do. If you change the very first phrase the bot says you might have a lot more people engage with the bot. Experiment and iterating on these things is critical because it may be working fine but if you change one little thing it can completely revolutionize how the bot interacts with the user. And create those key performance indicators that you really want to have happen. So the pointers there, obviously run experiments track and clean the experiment data so that is has meaning. You want to make sure that you’re doing copy optimization and then continuous improvement.You want customized contextualized complimentary experience driven by continuous improvement. So there’s some critical things there as far as making sure you’re testing on a regular basis as well as improving the bot. One other interesting thing that was pointed out in this as far a copy goes. I think that’s one of the things that a lot of people hesitate on is being comfortable putting in copy that sounds a little informal in some cases. We want these conversations to sound as humanlike as possible. So a couple of the tips that were given as far as writing copy, they called it the holy trinity of copy writing and I love that phrase. But the holy trinity of copy writing was be clear, be concise and be upbeat. Those three things are pivotal. It has to be absolutely clear to the end user what they’re trying to do or what’s expected of them,what they’re going to get. Keep it simple: you don’t have to spend 8 or 10 lines on a sentence that could be 5 or 6 words and then be up beat. Make sure that a happy positive experience for that end user. And then that’s again the holy trinity of copy writing is to make sure that those three things are there. Be clear, be concise and be upbeat.
Billy: Let me ask you one more thing. Right after you came back you told me about an exercise to make sure your copy made sense and was good. That you were going to have our team do that someone suggested where you sit back to back. Could you explain that a little bit?
Trace: Yeah for sure. So I think a lot of times when we’re creating chatbot conversations we sometimes get too much into the weeds in the conversation and we think its perfect. When in reality an end user hoping into that conversation it feels disjointed and a little odd. And so what we were encouraged to do, and I love this idea, is to actually as were testing out the language as we deploy the bot is, have two different people sit back to back so they can’t see each other. There’s no voice inflection. Because we’re trying to simulate the actual conversation experience. And so what were creating is, you can’t see their face you can’t, as much as possible anyway, you’re trying to take out the human aspect of this bot. And so you go back to back sitting on chairs and one person is the bot and one person is the end user and you simply go through the dialog as if that person were responding to a bot. And make sure that the conversation actually flows and that there’s no disjointed thoughts, being very radically candid with each other in that process is critical.
Trace: One other thing too that was interesting to learn there were spending in America. We’re spending 200 billion dollars a year on human support for call centers. I mean that’s a lot of money, 200 billion dollars a year. And were handing conversations over to humans a little too early in some cases. A lot of companies in almost any industry can save a lot of time and resources by utilizing chatbots and then making that transition, that handoff to the support team, the human team a little bit later. Where they don’t need to necessarily capture as much data because the bot can do that for them as long as the bots are designed appropriately. So one of the things that is critical in that process is number one getting the bots to flow like a natural human conversation but also one of the new technology pieces that are coming out and emerging is natural language processing and natural language understanding. Where bots are able to utilize those conversations like a regular, like you were talking to a human. So I think one of the big takeaways in the end is, man this industry is huge, this customer support industry. And how can we utilize bots to make that more efficient for each of these companies. So that that hand off is smooth and so that there is time and resources saved by each of these companies.
Billy: Awesome man. Well thanks for hopping on for just a few minutes. When you were telling me about everything you learned when you went I was like we’ve got to share this with everyone. So really good stuff. And thanks again Trace.
Trace: Sure thing.
Billy: Ok guys that’s everything for today. Hope you enjoyed this episode! Remember to rate, subscribe and review. Until next time. Talk to you later.
Click here if you liked this podcast and would like to subscribe to the “Digital Conversations with Billy Bateman” podcast, or here to learn more about ChatFunnels and how we can help optimize your chatbot.