 Hello everyone, and welcome to this week's Product School webinar. Thanks for joining us today. Just in case you didn't know, Product School teaches product management, coding, data analytics, digital marketing, and blockchain courses online and at our 15 campuses worldwide. On top of that, every week we offer some amazing local product management events and host online webinars, live streams, and ask me anything sessions. Head over to productschool.com after this webinar to check them out. Hello everyone, my name is Casey Phillips and thank you for joining me for this product school webinar on increasing chatbot user engagement with personalization. A little bit about me before we dive into the webinar. I'm currently at the chatbot product owner at Intuit. I've been working in the chatbot space for the last couple of years and I'm an avid blogger in my free time on chatbots and Alphins AI for chatbots magazine and the startup. So first let's start off with explaining what exactly user engagement means for a chatbot. The easy answer for that is it means everything. Although user engagement chatbots provide absolutely no value, think of a chatbot that introduces itself as a welcome message that's allowed to the user. The user does nothing. They don't message the chatbot back, they don't click a response button if the chatbot has one, but they don't take any form action that the chatbot provides. In these instances where there's no user engagement, there's no opportunity for the chatbot to provide any meaningful value to the user. The chatbot is not going to cut it in the long run. Generally you'll see higher quality, higher forms of user engagement. Usually deeper in the conversation, the deeper you go, the more likely it is that the user is actually getting the answer to the problem that they need or the user is actually getting the value that the people are using. For example, it's looking to use the chatbot to order pizza. The further the user is able to get into the conversation, the more likely they are to actually be able to order that pizza and get that value. Unless of course the user is stuck in an endless loop. It always says always something that you want to avoid when creating or designing a chatbot. You don't want users to end up stuck in and just going through repetitive motions with the chatbot. That would obviously get them further to the conversation, in terms of conversation steps, but certainly not a high quality of user engagement or any value being provided there. So definitely an exception to that rule. Right now this is a general negative connotation of chatbots that can definitely mean optical to their user engagement. The current public perception is chatbots is mixed at best, I would say. A lot of this is due to struggles of early chatbots. They create a lot of preconceived notions and just a lot of pessimism from users that can be very hard to overcome. There's a lot of users that lack confidence in the capabilities of chatbots now because of that. So chatbots really need to show their abilities early and often and wow users to really prove it to them that the chatbot can provide value and that it is worth the user's time and that it's better than the user picking up the phone and calling a human or whatever the user did before this chatbot existed. So moving on, what exactly is personalization when it comes to the chatbot? Really it just involves customizing the conversations and creating customers intimate conversations that appear tailored to each individual user and the situation. It helps boost confidence that users have in the chatbot, which leads to ultimately increased user engagement, a shortened path to value for the user. So we're talking about what that is, having the user, requiring the user to go through fewer conversation steps and back and forth with the chatbot so they get their answer or get the value that they're looking for. One thing that you will know is frequently with users depending on the space that the chatbot is in or exactly what the chatbot is doing or supporting. Users are kind of very infuses as to how much time and effort they're willing to spend on the chatbot until they get the value that they're looking for before they're going to get frustrated and try an alternative method of getting the value that they're looking for. So if you're making them go through too many conversation steps, you may back and forth with the chatbot or it's just taking too long in overall time duration, you definitely run a risk of the user's fuse running out and they'll either pick up the phone and call someone or find some other way outside of the chatbot getting the answer or value that they're looking for. So ultimately we're really trying to create an optimized chat experience for the user with personalization and really help to increase user engagement, just create overall a better chatbot and a better experience. So there's a few different types of chatbot personalization that I want to highlight and then they kind of start on the lower end in terms of how much influence and impact they can have and kind of increases to go through them. So the first most basic form of personalization that I think are kind of the wow factor stuff that would include addressing the user by their name or perhaps mentioning the weather. If the chatbot is able to know where the user is tagged from, where they're located, also just even changing the green based on the time of day so the chatbot knows what the time, what time of day this for that user, what their time zone is. So you know, good morning in the morning, good afternoon, good night, and etc. You can also randomly customize the way responses are phrased. So if you have to welcome us to areas of how it's phrased a little bit differently, you know, when the chatbots acknowledge the user don't always say okay, they can say for sure sometimes say they got it. Just customize it can say that so the user, so it doesn't just start to feel repetitive and dull the user. And it helps to read definitely forward with the chatbot in the conversation. But outside of kind of like helping with the initial user impression, it really does not provide a lot of value to the user. But it definitely is important. And especially early on the chatbot conversation in the chatbot's initial message like that welcome message, it can really help to get the user over the cliff and to commit to responding and interacting with the chatbot and going down the path of having a conversation with it. There's also historical user activity personalization. So this is most customizing the conversation based on any past activity of the user. Like past activity could be the chatbot or the product or perhaps a website that the chatbot supports. An example of this would be if you had a chatbot that helped you to order pizzas. When the user interacts with the chatbot, if you know that every Sunday there's like a game day special and the user always orders that and they and you know their delivery address, you know their payment information, instead of requiring the user to enter all of that every single time. If that user chats with their chatbot, it's Sunday, you know that game day special is going and you know that the user's maybe ordered that like four of the last five times on Sunday. The chatbot could say, hey, we've got our game day special going on. I see that you freaking or order this today. Is this what you'd like to order? No, the users are might of that special and now all they have to do is respond yes. And they've got their initial order, their initial order going just like that. The user doesn't have to type everything that they're looking for. They don't have to type all their toppings, like what drinks they want. Just one quick response and they're moving on. You could do the same thing with their delivery address. If you know the address that they get delivered to, then the chatbot can say, hey, I see that we have your delivery address as such and such. Is this where you want the pizza delivered to? The user says yes, they're moving forward. Same with their payment information. So in this case, now we're really providing a lot of value in addition to just wowing the user. So ultimately, the value we're trying to provide is to provide just a quicker, more seamless and easier experience for the user to order pizzas. Then they could order them online on the website or over the phone with a human by remembering what the user likes to order. We're making a lot easier for them. It's a lot quicker. The user doesn't have to type everything out. So we're just really creating that seamless experience that's faster for them. If the user had to type everything out, they would honestly probably be easier for them to just place the order online or to go ahead and just call someone over the phone. I also have to keep in mind if the user is on a mobile device and they're just trying to interact with the chatbot through an app or on their phone, then making them type all that out would really become or something. So we're really providing even more value for mobile users in this example. So again, we're impressing the users. We're going to be ultimately providing them with a lot of value that hopefully cause them to go from being a new user to being a frequent, recurrent user of the chatbot. The last type of personalization I like to talk about, the one that is my favorite and really the most powerful is the real-time user activity personalization. What this is all about is customizing the conversation based on the real-time activity of the user that's interacting with the product or the website that the chatbot supports or could even be based on the user's actual community customer based on the real-time conversation with the chatbot as well. A great example, this would be a support chatbot that lives in a product or lives in a website and it supports that. If you know, for example, that users are in a certain section of your product, if there's an overwhelming FAQ that users always seem to have, this is like a really frequent issue that always seems to apply to users at this part. In this certain part of section of your product, you can have the chatbot actually present that to the user early in the conversation and say, hey, I see that you're working on this such and such part. The product, is this the problem that you're having? Do you need help with this? So now not only are we allowing the user and we're really impressing them, but we're also overcoming one of the hardest things about chatbots and that's how do you handle when the user asks something in a way that the chatbot is not prepared for. In this case, we've been proactive and we've actually gotten ahead of that. So we don't have to worry about the user asking this question or how they could potentially try to ask this question or phrase it. We're presenting it directly to the user and if the user is having the issue, then they can just respond yes to the chatbot and you can start down the troubleshooting path. Another example of this would be, a very example would be a sales chatbot that could potentially change its sales recommendations or product recommendations and the options that present based on what the user is currently shopping for or looking for. So for example, if the user is currently looking for a certain type of hat and it opens the pop-up, the chat pop-up, then the chatbot early on the conversation could present relative content related to the hats that the user is looking for and not only does that show a lot of confidence to the user that the chatbot actually is able to provide value to them, but it also hopefully leads to better sales conversions for the chatbot as well. So to cast some of the real-time user utility personalization, it really creates a lot of value for the user and it really helps shorten the path to the answer. So the chatbot already has an idea what users want and need without needing to ask. So it really is great, especially cutting out a lot of the early steps in the conversation that will have to happen for the chatbot to get an idea for the user situation and what they're looking for. Instead, right when this conversation starts, the chatbot already has an idea what users are looking for and the conversation begins already tailored to that. So it's really great for impressing users, great in the wild factor, but also just really provides a lot of high value to the user that should create a great chatbot conversation experience and keep these users coming back for more. So in summary, there's really no such thing as too much personalization when it comes to chatbots. I definitely recommend mixing and matching the different types of personalization that I've mentioned in this webinar to really find the perfect fit to get your chatbot users engaged. There's really no perfect formula, something that you probably got to play with and see what would resonate to most of users and what really gets them to engage more and what really kind of makes or breaks their conversations and cause them to get the most value. I definitely don't fret if your chatbot is only capable of the basic wild factor personalization that could just be used by their name, mentioning the weather, customized breeding, you know, good morning, good afternoon, based on time of the day, stuff like that. If that's all your chatbots capable of doing, that's still fine. That's better than nothing. Just really try to do as much as you can, especially early on the conversation and that welcome message to really get users over the cliff and get them to commit to actually engaging with your chatbot and having a conversation. I definitely recommend those things a longer term goal of getting your chatbot to the point where it is capable of driving value through the activity based personalization that we talked about. So the historical user activity and then just even more importantly and even bear the real-time user activity if you can get to that point. Remember though about user engagement, chatbot really has no value. It all begins with user engagement. If the users don't engage with your chatbot, there's nothing. You don't have any opportunity on any chance to provide any value and to keep that user coming back for more. So user engagement is incredibly important with your chatbot and you've got to be innovative and you've got to be creative and find ways to really help drive it. Ultimately it comes down just really impressing and impressing and winning over those users that I mentioned that are a lot of them pessimistic towards chatbots and have doubts and have concerns. I think this is a waste, this chatbot is a waste of my time. It doesn't know the answers that I'm looking for. This chatbot can't help my personal situation. There's a million negative thoughts and connotations that go through a lot of users' minds when it comes to chatbots and again as I mentioned before a lot of it is definitely earned and it's definitely justified and it's based on a lot of past struggles from early chatbots that came out that just did not have a lot going on underneath the hood and chatbots that just really provide bad user experiences. But over the years chatbots are evolving especially right now. Chatbots are evolving at such a fast pace and chatbots right now have a lot of great capabilities. A lot of users don't know that and they're not aware of that and a lot of times they're very hesitant to commit and put forth their time to actually be able to realize and see that value. So you've got to impress them. You've got to impress them early. You've got to impress them often especially in that initial welcome message from the chatbot and you've really got to prove it to the user that your chatbot is worth it and that's going to be able to provide the value that the user is looking for. If you're not able to do that the user is just going to pick up the phone, call a human or find some other alternative to the chatbot to get the value that they're looking for. That concludes my webinar today on increasing chatbot user engagement with personalization. Thank you all so much for watching this webinar. I hope it was very valuable for all the chatbot builders out there and I hope that you're all able to take some lessons from this and some things that you can use as you build and design your own chatbots. I definitely encourage anyone if you'd like to connect with me on either Medium or LinkedIn. I love connecting with people in the chatbot space, sharing ideas, talking about chatbots, something that I'm very passionate about. It's a very exciting space right now that's moving at the speed of light and it's very exciting time to open a space. So definitely welcome anyone to connect with me and if you have any additional questions on the road or like to just talk a little chatbot. We definitely appreciate it. Thanks so much to CROC School for providing the opportunity to present with this online webinar. Thanks everyone. Have a great day.