Video: "Intercom + Fin Setup Workshop" | Duration: 2160s | Summary: "Intercom + Fin Setup Workshop" | Chapters: Workshop Introduction (6.43s), Messenger Setup (117.07s), Inbox and Routing Setup (533.045s), Fin AI Setup (1055.97s), Q&A and Closing (2075.105s)
Transcript for ""Intercom + Fin Setup Workshop"": Hello. Hello, everyone. Hello? Hello. Hi, guys. Welcome to our workshop on setting up thing and intercom. I hope you're all having a great day. So my name is Elena. I am Kendall. And I'm pushing. Yeah. So we are all from the, sales assist slash product specialist team of Fintech and Intercom. We are very happy to see you. And today, we are going to, work you through how to get everything up and running, yes, inside your, Intercom slash Fin, workspace, so you can start delivering better, faster, smarter customer service, yes, from day one since you start using us. So before we begin, just a quick note. Yes. We know that everyone works at a very different pace, and that is completely fine. Yeah. So just the goal of this workshop is to show you how simple is it, yes, to set things up, inside, FIN and inside Intercom and guide you through some basic, steps, yeah, some sections, step by step. So if you already have a trial account and you are able to follow along, that's great. But if not, not... Like, it's not a problem. You're welcome to just watch, yes, and take it in because, after the workshop, we will send you the videos on, like, absolutely every step that we are going to show to you. And I believe we are ready to start. Yep. Right? Let's go ahead and get started. Go for it. Alright, everyone. Let's go over our agenda today. So we'll be setting up... We're showing you how to set up the intercom messenger for both visitors and logged in users, myself. We'll walk through how to set up email forwarding and then how to add a teammate. Then we'll tag in my colleague, Christian, and he'll show you how to create a team inbox, route conversations to that inbox, and how to create articles. Then we'll pass over to Elena to explain what we call the Fin Flywheel. Gonna add content to FIN, set FIN Live, and then test FIN. So let's go ahead and get started. Now the first thing you do when you log into your workspace is you come into this little section here, which we call the inbox. Now we're gonna go to settings, and then we're gonna go down to Messenger, and then install. Now this is just gonna be a basic installation, a basic JavaScript installation. So we're gonna show you how to do it for visitors and users. Let's start with visitors first because it's a little bit easier. So I'm gonna click on visitors without logins, click on basic JavaScript, and then copy our code. Now this does require some knowledge of coding, so if you're not comfortable going into the back end on your side, this would be great to pass over to your engineering team, and they'll be able to help you. Now this is my test space called Visual Code Studio. This is what your back end will look like when you go in. Here, we have a little cheat code here. So your code goes before the closing body tag. Encoding, if we can scroll down and find one, your closing body tag will look pretty much similar to this, but instead of script, it says body. And that's the file that we want to add before. So what we're gonna do is we're gonna copy and paste our code. As we see, script script, make sure everything's inside the script, and this is the base settings that we wanna pull. So for myself, I have to save it. Now when we go ahead and open up our Messenger now... There we go. And open this with Google Chrome. And as we can see in our test dashboard, we have the Messenger here now. I can click it. I can see everything I set up for my logged in or visitors without logins. Send a message, our help center, etcetera, which is brilliant. Now we're gonna go back into our workspace and set up for users without logins. Now this is a little bit different because with users, you need a different type of code. So we're going in users with logins, back to JavaScript, and as we see, it's a little bit more information because we're pulling specific data. So we're gonna copy that code, go back to our Visual Code Studio. We're gonna add this before our closing body tag, but as we can see, some of these variables are orange and some of them are blue. What you wanna do with every variable that you're looking to pull, whether it's username, user ID, we're gonna add parenthesis around them. What this does is essentially tells our back end system we are trying to pull this data from the customer. You have to make sure that everything is enabled so that nothing is being missed. And once we set this up correctly, when we save it and reopen our website, we should be able to see the Messenger for logged in users, which is a little bit different. As we go to Google Chrome. Brilliant. Our test dashboard now has a Messenger for users. As we see, a couple different things have changed. We took away the help center, but you can still send messages. These little things are customizable as well within your workspace when you're setting things up. Just brilliant. That's how you set up the Messenger for visitors and users with logins. The next thing I wanna show you is email forwarding. So we're gonna go back to our settings tab. We're gonna go to channels and then email and then new email address. Here, you wanna add the email address that you want to use with intercom, so I'll say spicy taco at, you can select the brand that you'd like to select. If you have multi brand set up, you can toggle between the two, if it's available on your plan. And you wanna make sure that you have automatically forward emails from this address toggled on. Start the setup. They won't let me use spicytaco@kendall.com. So, essentially, having a little technical difficulties, but once this happens, you'll be able to start the setup for this address. Once you go there, you'll be able to see if it is authenticated or not authenticated. Here, we can edit this one and see. We can toggle on outbound messaging, any display names that we want, and then set up automatic forwarding. Once we get to this space, you're gonna copy your workspace email into your email provider settings, whether it's Google, Microsoft, etcetera. Copy that into the settings, and then verify automatic forwarding. So test messages get sent to you. Email providers give about a two to three minute window to pull back emails, so we won't be able to go into an email to see this in action. But once that test message is sent to you, you can open that message, and then your automatic forwarding will be toggle to green. You can also enable us for outbound messaging if you want. You'll see that it's correctly authenticated. If it's unauthenticated, you can click on this to verify your automatic forwarding and outbound messaging, and then click finish setup to go and complete that. The last thing I wanna show you before passing over to Christian is going over how to add a teammate. So we're gonna go to our settings again, our teammates, and new teammate, and click add new teammate. Here, we can say Tony Hawk intercom dot com. You can add their emails here. You can add a comma to add multiple teammates as well. We're gonna go over inboxes with Christian, which you can select any inboxes that you wanna add them to and then continue and set permissions. Each person that logs into your workspace will need a seat, whether that's a full seat with complete access or a live seat with limited access. You can set a role if you have any roles for them, Copilot if you want to add from Copilot to them. And here, you can set up conversation access. Here, you can see... They wanna see all conversations, anything that's assigned to them, assigned to the teams they're in, or any exceptions. Going down, you can set all of their permissions to take or add permissions, giving them access to the things that you want them to have within the inbox. Once done, you can click send invite and confirm. Now our teammate will get an email to the workspace. Once they click join this workspace, they'll be able to access the platform fully. So those are the steps that I would like to show, and now I'm gonna pass over to Christian to show us a little bit more. I think you can now. Alright. Perfect. So moving on, now we can create a team inbox. So here on the inbox, from here, to make things easier, you can use the search bar and look for help desk team inboxes right right here. Right here. Perfect. So in here, we click the... This area right here, new team inbox. Okay. Perfect. And then here, we are going to create it from our team members, Anna. You can see it right here. And here on the add team member... Teammate tab, in here, you can also add, or invite team members. And then you can sue... Can choose an assignment method. As you can see right here, we have three assigned methods. The first one is gonna be the manual with our conversations that must be manually assigned to your teammates. The other way is gonna be the round robin conversations are automatically assigned to teammates in sequential order. And finally, we have the balance that are conversations that are automatically assigned to teammate with a few open open conversations. Now here underneath, you can also set team office hours, and we have also, you can share a team reply time right here. Now if your team has a specific reply time, like, for instance, if you have a team in Europe and another one in The States, you can separate them. Okay? So next, I'm going to show you, conversation routing. Now, when conversations come, you can route them to specific places like Teams, views, and so forth. In order to do that, we will go to Fing AI agent right here. Okay. And we are going to create a workflow in this area. Now we're going to go to right here in a new workspace and create it from scratch. Okay. And we're gonna choose the trigger when customer sends the first message. Now typically, when we create a new workflow is to say the channel that is on, and we can choose, to work right here. As you can see, we can work from web, iOS, Android, WhatsApp, Facebook, Instagram, SMS, email, and it's like... And a little bit more. We also make sure over here to set up our audiences set, and you can have, like, custom audience like user, visitors, and leads, and we can have seen them, like, integration on the messenger. It's very well, and this is customized. K. One one that is set, we're going to make it entitled it routing. Okay. And the way this works is that when a customer, sends the first message through any of the channels that we selected, the next steps happen, as you can see right here. And you can see here all the channels that we have created on. Okay. Here, we can add a step right here, and we can assign conversations to team that you prefer. And in this case, I will choose customer support. We can also assign this to a specific teammate in here, and I'm just going to assign it to myself. Perfect. Now I have to be in the customer support team if I want to assign this to myself, but this can be done separately as well. We can add, any routing steps if we want any specific attributes to trigger. For instance, if somebody writes and wants a refund, this workflow can trigger specifically of the word refund. And once we finish this, we can set this live right here. And then when a customer send the first message, it is going to be assigned to customer support team, and it's gonna be to assign to myself. K? Now the next, I'm going to show you, how to create articles. In here, we are going to go to the knowledge hub that is located right here. Perfect. And here, you can create public articles. Here, public articles, internal articles, conversations, and macros coming from website, and also snippets and documents. Okay? Now this can be all be used for public health center and for thing as well. Now here, we're going to create a public article. K. Now we are gonna put a title, then describe it, And then we can start writing. Hello. In here on the right side, you can find information such as, type, status, article ID, the language that you prefer, created, created by, written by, last updated, and last updated by. Okay? You can also show the version history in here so you can see the changes to the article now in theme right here. You can see here what is enabled for, like service, sales, ecommerce, and compile. You can set up the audience right here. If you want a specific audience, like, if you have VIP customers, and you only want them to see the article that can be set up here. Now we have the help center right under here. And this part of this is important, because each articles needs to be in the collection for it to show up on on your public help center. So here, if you don't want, if you don't have a collection, we are gonna create one. Click right here. We are gonna create a collection right here. You can put the title and some description, and you can create the collection. Okay. Once we create the... That collection, we can add that article to the test collection, and on my website, it will be under that specific collection. We can also see any reports run under here. They can tell us the views, the conversations, their... How they have reacted, the free resolutions, and the feed involvements, and, also, we can see tags as well. Now, once we are have done with all of that, we can publish our article right here, and we can choose from publish now or schedule it for later. Once it's published, it will be available for the audience that we set, and that will be all for this part. And now I'm gonna pass it to Alina that will tell us a little bit more about Perfect. Thank you so much, Christian. So I'm going to show you now how to set up Fin AI just before we start. Very fast, I'm gonna show you how it works. Yes. So we're gonna go to our imaginary website. This is the banking app. Yes. And here we have our Fin AI agent in the service role. And what we can do, I'm... As a customer, can ask a question, and Fin is going to respond to me based on the knowledge center articles that we have inside Intercom, the way Christian showed to you now. Yeah? So as we see, Finn generated an answer for me. What it means for you? Yes. You do not have to deal with this kind of queries anymore. Yes. Finn can handle them for you. And our average resolution rate right now is close to 80% by FIND. So, let's go inside our test workspace, and I'm gonna show you how easy it is to set it up. So, for that, we just click on Fin AI agent section here. Yes. And as you can see, Fin is set up in so called Fin flywheel of train, test, deploy, and analyze. So, let's go through this section... Through every section one by one. Yeah. So first of all, you can train Fin to resolve, yes, even the most complex queries with your procedures, with your knowledge, with your policies. Why we need the training section before going live? Obviously, we want to ensure that Fin understands your business deeply. Yes. And it can deliver policy aligned answers. Yes. And resolve, customer issues autonomously without any human assistance. Then in the test section, you can run fully simulated customer conversations from the very beginning to the very end, and it will let you validate, yeah, Fin's behavior in realistic scenario before you launch it for your customers. So you can maybe identify some gaps. Yes. In your knowledge center, you can refine some responses, and then you can deploy with confidence. So in the deploy section, obviously, yes, you just, fix some final setups for Fin, and you can, deploy it through different channels, like, chat, social media, email, phone calls. We have, like, a lot of standard channels where things can be embedded. And finally, analyze section. Yeah. So here you can use AI powered insights to analyze and improve Fin's performance and deliver, better customer experience with time. So, basically, what it means for you, yes, the more you use Fin, the better it gets with time. So let's go to the train section, and here I want to show you the content section first. Yes. We already set up, an article inside Intercom with Christian, and now you can use this article for Finn. So Finn can scan it and answer. So, we have different options, yes, to feed Finn with different kinds of information. So here, yes, we have an option to add public article, internal article, some snippets, that we've just done. Or maybe, you do not have a knowledge center prepared yet. You don't have time maybe to write articles right now, but you have a bunch of information on your website, so you can just put your website here, yes, and it's going to be synced, yes, with Finn, and Finn is going to provide answers based on your website. Maybe, yes, you store your knowledge center in a different software, for example, Notion. Yes. So you can also connect Notion, and Finn is going to read the information from there. Maybe you have a bunch of documents, yes, where you have some frequent last question sections. You can just upload, it here. Yeah. So as you can see, we have different options here. No matter where you store the information, Finn can, read it. And after you sync Finn with this information, you can test it here in the preview mode and, check how it is performing. Yes? Then, I wanna show you the guidance section. So, here you can actually, shape the logic, yes, of how Finn is responding even more. Basically, you kinda, create a personality for Finn and, give him some maybe weird scenarios or some guidance of what Finn shouldn't respond. Yeah? So, for example, let's look at the communication style guidance. Yes. We can just click add here. We can, name it. Let's say how to communicate. Yes. We can set up the audience. So Finn is going to use this guidance for everyone. We can define in what channels do you want Finn to use this guidance. Yes. And even schedule when do you want this guidance to be applied. And, you can just describe in a very natural language how you want Finn to communicate with customers. So here, I just write that it should maintain a friendly tone, yes, and use positive language, and avoid very formal phrases. So, when we click enable, yes, this guidance for Finn will be enabled, and Finn is going to use it. So, here, we can also add some context and clarification, basically, some, customized guidance on the follow-up questions that Finn should ask. Maybe you want Finn to collect, like, the email of a customer or some additional information. So you can, create the new guidance here the same way as I just showed you, yes, on, follow-up questions. Yeah? Spam filter. Yeah. So maybe you don't want Finn to, talk to any potential customers with just Gmail, email address. Yeah? So you can set it up. Maybe you don't want to serve customers from specific countries because there are, like, a lot of spam there. Yeah? You can also set it up here and any other guidance that you have in your mind. For example, financial institution, they always set up the guidance, do not give any financial advice. So here, you can just custom the guidance according, to to your business. Yeah? And then after we set up the guidance... I don't know why it doesn't work, anyway. We are going to go to the test, test section. So in the test section, yes, you can add, real questions from customers and see how Finn is going to respond after you set up the guidance and after you set up the content for Finn. So if you maybe used another solution in the past and you have a bunch of frequently asked questions, yes, from your customers, you can even upload it as CSV and test it, yes, as a batch test, Or you can also add it manually. I'm going to add it manually just to show you how it works. You can type your question here, save, and now, Finn is going to respond based on the setup that we've just done. And after that, you can evaluate, Finn's answer. Yes. If you are happy or not happy with it. Yeah. So here Finn's summarized. Yes. What kind of different plans does my company have? It says that it used, article, yes, and three different guidance that I set up earlier. And here, we can rate the response. Yes. If it was good, well, wonderful. Finn will continue the way it responded here. If it was maybe bad, you are not happy. Yes. You can select the reason, why you are not happy with this answer. Yes. We have been messed up. And after that, a thing should, improve, yes, based on your market. Then we will go to the deploy section. Yes? So after we tested Fynd and we're sure that it's going to provide good answers, yes, to our customers, we can deploy through different channels. I'm gonna show, a simple deploy via chat now. So, here, we just finished the setup of Fin, basically. Yes. So if we go to this section, which is define what kind of audience can see thing. Yeah. So maybe you want thing to be enabled only for your customers and you don't want thing to be enabled for, website visitors. You can do that. Yeah. So you can, select the audience here, and after you select it, you can click save. Then, yes, you, define in what channels do you want to use FIM. Yeah. So when it comes to the chat, it can be on a web chat the way I showed you in the beginning. It can be inside the chart of your application. If you connected your WhatsApp, Facebook, Instagram, it can be in your social, and if this is a popular channel, yes, for your customers, and some other channels as well, like SMS, Slack, Discord, Telegram, and Microsoft Teams. So after we define the channels, we can click save here again. And then we just want to write how FAIN introduces itself. We have a couple of default messages here, but you can delete them, and you can add your own message if you want. After that, yes, we just check... We double check, yes, what kind of content does Finn use. We see that it's using our help center articles and what kind of guidance does Finn follow. If we do not like something here, we can manage it from this tab as well. Then we can define, what happens if Finn cannot resolve the question of a customer. It might happen with some complex cases. Yes? So, normally, what we recommend that Fyng just seamlessly, hand, the conversation over to a team, yes, in the inbox the way, Christian showed, yes. You can set up different team inboxes depending on the topic, and you can escalate the conversation into a certain team inbox or even into a certain person inside the inbox, and they can pick it up there and respond to a customer. Or, this is not the very popular setup, but you can, ask Finn to just close the conversation, and Finn will offer another way, yes, to get help. For example, maybe you can write, hey. Sorry. Like, I can't help you, but you can send a message to this email address, or you can give us a call if you want. Yeah? This is also possible. Here we can, ask Finn to send a CSAT, yes, to a customer to evaluate how how good Finn or how bad Finn is, so you can see it in your reports. And finally, follow ups. Yes. So, if a customer does not respond within four minutes, yes, after Finn answers, what do you want Finn to do? It can confirm, yes, and ask a customer, hey. Do you need more assistance? You know? It can offer an escalation to a human, or, maybe there is no overlap, yes, if it's not necessary. And, also, if it can close, yes, abandoned chats. So if a customer left the chat and after certain time, it can be days, it can be hours, it can be minutes. Yes. Finn will close the conversation, and you can add a custom message here, like, sorry. You were inactive. I'm going to close this conversation now. And that's pretty much it, yes, when it comes to deployment. After you, checked, yes, all these settings, you can test it one more time in the preview mode here and then send it in live, and it's going to go live to your customers. And finally, what I want to show you is the analyze section. We're gonna go to another workspace for that. Yes? So, if you only started, yes, your trial, this space will be completely empty. But here, you can analyze the performance. Yeah. So here, we can see automation rate of the FIN, involvement rate, yes, and resolution rate. So probably, resolution rate is the most important. Yeah. We want it to be as high as possible, so it means that FIN is resolving most of, the questions, yes, from your customers. Here we can see customer experience score, yes, how happy customers are with Finn at the moment. Here we can see performance channel where we want... Where where we have FIN. Yeah. So we have it in the chat. We have it in the phone. So, here we can see that FIN for services involved, yes, in the chat conversation. So it's basically the funnel, yes, of how thing is working for you, how many conversations it is assigning to human, etcetera. Yes. And here we have the pro section. So let's go to recommendation. It's empty now. Yes. But, basically, usually, in the recommendation tab, you can see recommendations to improve your knowledge center articles. Yeah. So, basically, it can review, yes, the conversations that Finn cannot resolve and and, the conversation that Finn is sending to your team. And then, our pro add on can provide weekly recommendations, yes, to fix gaps in your content. So maybe your customers are asking about, I don't know, a card replacement quite often, and you don't have enough information here in your knowledge center. So that's why Finn cannot resolve it. So maybe you can add an article, yes, about card replacement, into your knowledge center. So here is a topics explorer. Yeah. So, topics explorer uses AI to automatically group your support conversations into different topics and subtopics, and it shows basically what your customers are asking, yes, and how these conversations are performing. So here, as you can see, yeah, we have a lot of conversations about current replacement. We have a lot of conversations about update and contact address information. Yes. So these are, like, the most problematic topics. Here, we have trends, very easy. Yeah. So it just shows what is going on inside your workspace, yes, with your support operations. So here we can see that, fan handled fewer, but more resolvable... Okay. I'm not gonna pronounce this word. Card replacement, yes, during the last week. And finally, monitors. Yeah. So, monitors, it's actually our QA tool, yes, that is evaluating things, things conversations. And that is pretty much it, yes, when it comes to Fin AI and Intercom workspace. So if you have any questions, feel free to, type them now in the q and a section, on the right side from your screen. Any questions? Any questions, guys, while we're here? Alright. It doesn't seem like we have any questions, which is good news. We hope that everything is clear. So, yeah, thank you so much for joining this workshop. If you don't have a free trial yet, feel free to sign up. It's, for free for fourteen days. You can set up everything. If you run into any issues, feel free to send a message to Finn. It will Hi, respond. If not, it will reassign the conversation into our support team, and we will send you, the videos, yes, on, all the steps that we just showed you. Thank you so much. Have a good day. Thank you. Bye.