Using Chatbots to Capture Leads & Sell: Joseph Holguin



Hey, how's it going? So my name is Joseph Hogan going to be talking to you today about getting leads from a bot for the company. So I'm the co-founder smart Loop got a I wear a chatbot and live chat platform. We just launched live chat Monday. So now if the bot fails which it probably will in most cases for everybody now a live agent could come in and take over the conversation have ten years digital marketing experience and you might think this dude looks young. I'm 34. So that makes me 24 when I started and I was just think about it last night is like that's when Facebook pages started rolling out for businesses like around that time. So I got into it because my family they grow spices and New Mexico like red chili green chilly stuff like that. They're selling them locally.


I started doing these Facebook pages. And then we started selling chili like all over the country with like people internationally wanted to buy chili just because we started creating these communities and I was through Facebook pages and now it's now probably ten years ago five years product. We had a previous start up the crowdfunding space and we've been working on Smart Loop for about three years now. So smart Loop, like I said, we're end-to-end platform for Bots plus now live chat and the beauty of that is that you can easily create a bot. You don't have to write code. You don't have to do these crazy setups of like Azure or Amazon web services and then you can just click a button you can go to Facebook. So this is about that we do in National Geographic. These are some of our customers we work with Coca-Cola, AOL National Geographic some government agencies and University of people and that's who we're going to talk about today.


So they wanted to enroll more students using Facebook Messenger. So University of the people is the company that used our platform and this was recommended to them by the Facebook Education team. They reached out the Facebook Education team works at University of the people and University people is like hey, we want to do a Facebook Messenger bot. We have this problem. We need somebody to help us so they made a recommendation to us and we were able to help them create the bot. So University of people just a little bit of background on them. They're founded in 2009 their nonprofit tuition-free u.s. Accredit University and I think it's like really cheap to get a bachelor's I think you get a bachelor's degree in for like two thousand to three thousand dollars, which is really cheap compared to traditional universities and they have students.

They have around 20,000 students in 200 countries. So their goal was to enroll more students and when Facebook reached out to us. They said hey, we have this partner they have this problem, but we need to do this in three weeks. And if you think about rolling out a solution typically the not going to take it's gonna take a lot longer and maybe three months so we were tasked with helping them enroll more students and get it launched and working so that they can get some data and prove it out for their Facebook f8 conference. So we're able to do that in three weeks versus three months and I'll go into more and how we do that. So University people their goal was to enroll more students. The problem was since it's an online university. There may be having a touch point with somebody for the very first time there's a lot of Education that comes into it.


They have to answer the first part of like who they are. What is online degree? What degrees they offer? How much is the cost? What is the process of actually getting a accepted so it's a lot of communication that needs to happen there and this is the that they had they had people from all over the world coming to them on Facebook messenger and they would ask simple questions like what degree programs do you offer and that's a simple question right? But for the volume that they were getting being that it is a worldwide University. They were getting questions 24/7 something like this would take about three to four hours just to get a reply sometimes. So if you think about it, you have three questions by looking like 8 to 12 hours before you actually can get some type of response and some general consensus of what the university is about. Then they would try to push you to their website and then from there on the website, then you would have to poke around and find requirements before you could even set an application, right? So that was a major problem and I've experienced this firsthand when we started talking with them. I went to their Facebook page ask a question for hours later. I got a reply. It was just a long like canned response and then I had to go to the website try to figure it out.


So it was a very broken system. So our solution was a messenger chatbot to educate and qualify potential students in this case their leads. We don't like to say leads but their needs so we create a chatbot from you click get started. There's an introduction about the university that tells them okay. This is where online university we set the expectation right off the bat. This is a chatbot. You're not talking to a real person. You're talking to chat Bots and then we ask them.


Are you a student? Have you been a student or you're interested in being a student? So right off the bat we're already qualifying the user. We're actually setting these parameters around the user so that we can serve specific type of content to them if we know they're already an existing student and they have questions and we can route them. To a different agent inside of the university to answer their questions or if they're unused student and they getting get routed to admissions to answer questions. But as you can see from The League qualification right off the bat we start to ask these questions. Are you a student? Have you been a student? Are you interested in being a student? So at that point we're starting to build a profile on the user and we can use this dynamically inside of the bot to direct them to certain types of content or in the future. We can serve other types of marketing material or other types of content based on that. So the body has been live for about three months. Now, they've already had 40,000 students come into the Bots and between the Bots and the user.

There's already been over a million messages. So these are people coming to the bot learning about the university learning about their degree programs going through this lead qualification step, and I'll go into more what that is, but there's a lot of steps in there and that's where the Bots really helps to facilitate lead qualification.


So out of those 40,000 people that came into the bot 51 percent of them actually started an application and it was hard for them to tell us exactly what that was before the bot. But we're thinking it's fine. Probably three to five percent of the people that actually came into messenger without anything went 12 15 hours of conversation there and then went to an application now we're seeing 51% of the people actually go through the bot. They get through all the information The League qualification steps and they go to start filling out their out the application so that 51 percent 20 percent of the people actually complete an application.


So think of this this is 20 percent of those people are going to the bot learning about the university filling out the application to the university without even talking to a person at all. So just drastically removed any of the friction that there was before with learning about education and the requirements and getting qualified and this reduced Arab spent on Facebook for leads by 62% I think previously they were running Facebook ads and they would just push them to the website page and then they would poke around there and then you can go to application.



So this drastically reduce that and this was paired with the sent a messenger ads, which I'm sure most of you guys have seen a lot of companies aren't doing them correctly right now with connecting to a bot or even using a bot. But when you pair the sent a messenger ads with the Bots and you do it correctly, it's very seamless and you can just take people from the news feed straight into a bot without any restrictions at all. So let's talk about how we actually did this in two to three weeks. And that was that was a stressful time for us because a lot of us on our team. We have the platform anybody can sign up but at smart Loop dot Ai and use our platform. But we also provide some type of service sometimes where we help people understand how to build robots. So we worked with them to build a bot at this point. They manage everything themselves content changes flow changes everything, but initially we kind of help them do that.


So these are our steps to launch So first it was the user understanding the user flow and content. So for them, the user Journey was I'm a new student I come into the bar. I don't know anything about the university. So let's tell them about the university. The next step is okay. Now, you know about the university what type of applicator what type of degrees are you interested? So they offer business administration computer science and health services degrees. Okay. So tell us what type of stuff you're interested in.


Is it an Associates or Bachelors that you're interested in? Okay, so there's different content based on that and we're logging this and building the profile and then once you've told us okay, this is interested in a bachelor's in computer science. Then now the lead qualification part. So for them as most universities, you have to have a GED or high school diploma, you have to for them specifically since its worldwide you have to have like an English proficiency and you have to be a certain age and there's other requirements on top of that. So for us it was really defining this is the initial content. And then also what are the lead qualifications steps that we have to verify and Route a person based on who they are. So for the user Journey, we initially just started just pen and paper and we started mocking allocate. This is a start. This is the onboarding.


This is where we learn about them and then based on that if they're interested in the degree programs routing them to their so a lot of the first work was just really mapping out what that user experience was looking like and then from there we were able to come back and say okay. This is the type of content that we have to write for these sections. This is the type of supporting content from computer science. And then once we had that then getting into the lead qualification steps, so the league qualification steps, there was a lot of work that came into this where we really had to understand. What is the University or in anybody else's case like what are the steps or information that you need to collect? So for us it was like are you a student? Yes or no. That's very right off the bat. Okay, the second one. I was like, okay.

Are you interested in what degree computer science? Subset of that is are you interested in a scientist or bachelor's degree and based on that we were able to Route people to there's different pricings based on that. There's different requirements based on those types of degrees. And then once we were able to do the high level degree information, then we really got into the lead qualification. I know this some of this League qualification is kind of a little bit more complex than just a normal business where you would just collect like maybe a company size location, maybe some technical requirements, but essentially it would be the same. Thing depending on the business. You just have to really identify. What are the key lead qualification steps that you need to do? So based on this we are able to build inside of the platform able to build like a user profile based on this so we know that this person and I just pulled this directly last week.


I knew this person was interesting computer science, they want to associate and they were not a student so initially when they start the conversation they went through the bot. They told us this information and based on this information. It was able to create this kind of custom user experience to explain to them about the university the cost of the university and then initially get to the application process, but we can still use this in the future to reach out to them about upcoming programs. Maybe they got their Associates and now we can say hey it's now it's time to get your bachelor's and we can send my message or for example, there's a lot of students that come into the Bots and they are students. Ready? So all this new student information is irrelevant to them. So if they tell us like I'm already a student have questions about a class or I have something like this and we're able to Route them to that part of the university to answer those types of questions. So it's very key to kind of from the very beginning of your Bot is to understand. Okay one if it's legit lead generation.

Okay, what do I need to know for lead generation? But if other people are using the bot like existing students, how can we use that data to actually extend the experience and provide them maybe customer support or some kind of other detailed content? So for traffic the good thing about this was that they're already had a lot of people coming to their bot organically. Just I think I think this is just my insight, but they do a lot of education programs in third world countries or outside of the US so they get a lot of organic traffic right now probably at first when we started the bot it was all organic and people were just coming in asking questions. So that was really good initially to start the bottom. So for you guys when you think about your Bot think about there's I think there's two or a couple little hacks that we do right now is if you go into the page settings, there's a setting that says, I don't know specifically what it's called, but it's like basically pops open the messenger window when somebody visits your page and that draws in a lot of tension because if somebody just comes to your page for the first time they're going to see your messenger window pop up and then your Bots right there. So for sure turn that on and then also utilizing like your existing organic reach. So if You have a million followers or 500,000 or whatever then you can use that along with the messenger links to get people to come into your body and that's what we did initially and then you can add it to your website. So we're in the process of adding it to the website where they already get a lot of traffic. So that's another good way to get traffic and then the biggest thing is Facebook as right now.


So Facebook is really pushing the sent a messenger ads. And again, like I said a lot of you guys already seen them sometimes I just think because the way the platform is it's very easy to create send a messenger but adds right now, but they're not paired with bots. So you would click on this and to message it would come into messenger and then it has like the canned responses and then you have to wait for someone to reply to I think a lot of us have already seen that but if you do it correctly, you can click Send to messenger whatever platform you use if you're using smart Loop, we show you how to do it really easily. And it's almost seamless like you don't even know that you came into a bot. You're supposed to set the expectations and say it's a bot. But you just click on the button and boom you're in to Facebook Messenger and you can start going through this process. So for them, it was really useful because we're able to use a Facebook ads we can Target may be people that don't have an education or a higher education right now in certain countries, and we can just they click the ADD and we can Them into a specific part of the bot. So just to touch on the build process, like I said smart loop we're able to do this in two to three weeks because we just make it so simple you don't have to write any code to create these Bots dialogues and bought flows.


Yeah, we have a point-and-click builder. We have analytics and natural language processing built in its multi-channel. So you can deploy it on your website or Facebook Messenger one-click deployment. So you just click a button on Facebook and it goes you can do broadcast notifications. And like I said, we just launched live chat. So that's really useful in making sure your customers get the best service that they can and then use that data to improve it. So you don't have to answer the same questions or do the same things over and over again. So in conclusion, I think Facebook Bots or robots in general really good to do lead generation.

It's just clearly defining. Okay. What is the user Journey look like what lead qualification steps do I need and then you can just set it free and it will should work for you. All right. So you guys have any questions based on what you just talked about on maybe how to launch Bots or anything that we saw in the whole process will be the mic. I think it's really impressive that you've deployed that in 3 weeks. I'm wondering how you kept the business requirement Gathering process to a minimum because I know sometimes just figuring out the requirements for the flow takes a long time. Like were you doing it agile and coming up with the diagram showing them immediately getting sign off and then writing the conversation from there.


Yeah, so that makes it really useful so you can use like our friends at bot copy bought society and you can really kind of map that out and the good thing about using a tool like ours is that you don't have to go you can just make changes on the Fly. It's not a huge technical requirement to change some copy or to change a flow. If you see that something needs to be updated. You can just do it in a couple minutes versus a couple hours leather process. So that was really helpful for us. I also understand you will be giving a workshop tomorrow from like 925 that get in more depth and detail. Is that correct? That's correct. That's correct.

Another question pre Workshop question. Yes. Okay. Hi. So easy similar to many chatter fewer is a smart lupus similar product. I don't see any of those guys here. So yeah, we're bitter actually because we have no speak the truth. We have natural.


We have natural language processing built in so you don't have to connect it to any external service. It's not keyword match. Edge and we also give you the ability to write code inside of the box. You can build some custom features and pull from CN crms and external Services. Very easily plus Live Chat is really nice snow over there. All right, watch your feet Ginger coming through. So the examples you showed the all were closed entry Fields. I was just wondering why that was and what part like the buttons? Yeah, it was all UI responses.

Right? So our philosophy is that if I just open this up and I say what degree you're interested in that just opens it up to a whole plethora of Different flows that can happen so our ideology. Is that like if I have this question and I can show buttons then I will restrict you to those buttons because that's all I care about right now. There are some ways you can get out of it. But if you're in this then like specifically like this yes or no, like it's really not beneficial to know that you're from Bulgaria and you know English right? I just want to know yes or no, so we try to restrict people in forced people in a nice way to get to where we want them. Can I force anyone else in a nice way to ask one final question? There you go. So you told you have in build NLP. Did you create your own NLP or used one of the open source and LP's? Yep, so there is nice guys from rasa here.


They have open source that allow you to use that as well and we use some Google products as well. Yeah, any other burning questions? Okay, how about you? Do you have any kind of final comments? Yeah, so looks like you went for the mic there. There you go. If you're interested in a lead qualification template, you can sign up as smart Loop dot AI we have some templates there. If you're interested in more something like this then feel free to talk to me later and I can help you come up with something like this as well. That's all ladies and gentlemen, that's your cue to clap.

© 2017 by Chatbots Conferene TM.