Service Hub: Scaling Support with AI

Foreign.

So welcome to another session of the HubSpot Admins Hug. Today we're going to be talking about scaling support with AI. So normally I would be the one sharing this housekeeping, but I've got some friends in the chat going to be doing that for me today. Let's see, we've got a whole bunch of new episodes that are coming up. We've got HubSpot data modeling on May 20th. On the 27th we're going to keep keep the service hub going and driving innovative retention with the team there. And then on June 3rd we've got the prospecting agents. That's going to be a great one. I can't wait to see all of the amazing prompts that are going to come out of that June. We've got multi account management and Sales workspace feedback session so make sure you come to that with a lot of your questions. I know we've got a lot of those and there's been a lot of feedback that's been happening in other sessions on the sales workspace, so bring those with you. Everyone's favorite workflows is coming again on June 24th so definitely RSVP for that one. It's gonna be a big session on July 1. There's not going to be a hug, so no we won't see you on Tuesday. That's the Global Week of rest. So go ahead and spend some time with all the other things that you've got on your plate for July 1st. We do have a bunch of other sessions that are scheduled for the end of July and into August, so go ahead and check those out when you get a chance and make sure you're RSVP putting this on your schedule. If you miss a session, make sure you are checking out the YouTube channel. There are a ton of videos on there, lots of content that you can take and bring back to your teams and digest whenever you get a chance. We do a ton of these. We want to see all of the amazing that your team and that yourself are creating for the HubSpot. Go ahead and fill out the show and tell let us know as soon as we get five or seven of those. We'll go through them. Make sure we get a session for everyone to share their amazingly awesomeness with the rest of the hug. We do have a space on the Community channel, but our favorite space and everyone else's is the Slack channel. We also have a LinkedIn space and group for everyone here, but definitely go ahead and join the Slack sparketeer Slack. Lots of amazing admins in there. I'm not going to talk about the bootcamp because it is full, but make sure you keep an eye on that so that if there's any new launches for that that you're joining it. HubSpot's week of rest, I already mentioned, is June 30th through July 4th, so we won't be having a hug that week, but you can check out all of the other hugs holidays that are there. The platform that we use is Bevy. It sometimes does its own thing. We know we love it. Just turn it off, turn it back on again. Check out all of these troubleshooting tips and it would probably start working again. If you don't see a speaker talking or if you can't see their screen or something's wrong, let us know in the chat and we'll get it resolved for you. If that doesn't work, we do tons of these once a week, so if you have any ideas for an upcoming hug, make sure you are submitting those. If you don't, you know what happens. I'm the one that gets to pick the topics. Kyle will pick those that he likes and we will move on and make sure you're connecting with both Kyle and myself on LinkedIn and I will bring this over to the team. I love it. So Paul, why don't you go ahead and take it away for us.

Amazing. Thank you so much. Really appreciate that and we are honored to be here. I am sharing my screen right now, so I'm going to kick us off. The topic today is how to scale support with Service Hub. Quick note here, things are subject to change and with any roadmap items, we reserve the right to adjust with new information. So just introducing the team you've got today. My name is Paul Weston. I'm GM and VP of Product for Service Hub. I've been at HubSpot for about eight years now, making my way through Marketing Hub, some of our data platform teams and operations Hub. And I've been on Service Hub now for about the past two or three years. Joining me today, we've got Amber, Smrithi, Jess and Mark and we're going to talk through a customer agent, KB Agent, the Help Desk, workspace, and a bunch of really cool insights in reporting features. And we'll have plenty of time for questions towards the end. So I'm going to kick us off but I really want to get out of the way. We can spend the most time actually digging into the product and answering the questions that you, that you all have got so I'm going to set the stage here. So HubSpot's mission we always like to start with, our mission is to help millions of organizations grow better. And Service Hub's job underneath that mission is to help companies scale, support and drive retention. So in the old way of doing things you kind of have like your point solutions. There's a million Help Desk tools out in the market that help you kind of scale support. But the unique aspect that Service Hub brings to the table is the ability to scale support across those humans, AI, agents and self service tools. But also now that ability to drive retention through customer success and feedback. Management tools of course all underpinned by that HubSpot customer platform, that single source of truth. We think that that's a really unique offering out in the market and how this fits in. In terms of HubSpot, you think about the flywheel. So going from Marketing Hub attracting visitors and creating leads, Sales Hub helping you create deals and close deals from those leads. What happens after that deal closes though? That's where Service Hub really comes in. You onboard your customers, engage them, retain them, grow them. CSMs kind of pick up the baton here and their job is to retain customers. And underpinning that all is are all of these awesome support and self service tools that help you resolve issues efficiently. And that's the topic at hand today. I wanted to call out Spring Spotlight about a month ago. I'm sure you all have been very plugged into this, but Service Hub had a major glow up as a part of that Spring Spotlight 74 feature updates. Just making a ton of progress here and I'm really excited for the product team to do some demos here and show you all what's going on. So some themes to watch out for as the PMs step up. First of all, surprise. Lots of good AI stuff to demo and to talk about. Secondly, Surface Hub is all grown up. We like to say that scaled, scaled up service teams can run their support through HubSpot. In fact, HubSpot's own support team runs on our Help Desk tools and our Service Hub tools. And lastly, just driving real outcomes. So using Service Hub, we're not just infusing AI for the sake of AI, we're trying to actually drive real measurable customer outcomes. And so things like faster reply times and just more ticket resolution have been proven out in the data through all these features that you're seeing. So without further ado, I'm going to get out of the way, I'm going to pass it over to Jess she's going to kick us off here talking about customer agent. So Jess, you want to take it away?

Yes.

Give me a second.

Right.

All right, can everyone see my screen? Okay.

Looks good.

Awesome. Well, thank you so much, Paul. I'm so excited to be here. As Paul shared, I am a product manager on the customer agent team and I'm really excited to be able to dive in with you today to talk about how customer agent and then my colleague's going to talk about KB agent, but we can, we're going to talk about how they can add value to your business fast by helping customers help themselves. So for starters, if you haven't heard, what is a customer agent? It's an AI agent. It's going to be an extension of your team.

Right.

Resolving queries 24. 7 and scaling support across all of your communication channels that you have like email, live chat, WhatsApp. And customers are going to be able to answer their own questions via customer agent. And as a result of that, it will help you reduce your support workload, help generate revenue, and even allow you to provide around the clock support. We see customers with resolution rates of, of an average of 50%, but some are seeing up to 90% and seeing a ton of value. I'm just going to read one of the quotes here in the middle. It says we have lowered the number of tickets handled by Customer support by 77%. Improving conversion rate through 24 hour support. BRI's customer agent is a key tool for generating revenue and increasing customer satisfaction. Like I said, businesses of all sizes are seeing value quickly from the customer agent. And we're so excited and think that you'll, we'll be able to see value too if you're not using it already. So with that, let's dive in. I'm going to do a quick demo of getting customer agents set up and actually show you the value that it can bring. All right, so I'm going to go ahead here and create my agent. I'm going to name it. You're going to give it a name. It'll let me type. Okay, a role, customer support. And you can choose from a list of different personalities. Actually, tool tip is if after you've gone through the onboarding, you can actually even connect your brand voice if you have one that's already existing. So there you go. You're going to use friendly next, you're going to give the content that Luma needs to be successful. This is, let's say I already had a bunch of HubSpot content. You can select the content here. It'll show like your knowledge base, articles, your blogs, website pages. You can sync that. Otherwise you can add files from the file manager or you can add public URLs. I am going to just let's see. My keyboard will not let me type right now of course I'm just going to go ahead and say let's say I created this agent. This is what you're going to show. This happened now again just in a few minutes. Really quick. The customer agent's been created and here at the bottom you can actually go ahead and preview and test. Um, so I'm going to go ahead and click on. I think it's already synced on like my hubs content. And here what the you're doing is like building trust, right? You want to trust that your agent's going to be able to deliver responses that reflect your brand and actually is accurate. And so here we see the agent doing exactly that. It used source source that it was trained on and a really thorough response. You can see it links up out to different. Once you now then have like tested it out and gotten a flavor for how it's going to respond to your customers, the next step is actually to assign it. And on the top right here you can kind of see all the different things that you may need to do to get everything set up and running. So I'm going to go ahead and assign it to a channel. Here we have five different channels that you can connect customer agent to. I don't have in this portal, WhatsApp, Facebook messenger or Calling, but those are available to you. But I'm going to go ahead and use email and I have this email right here that I will connect it to. This is a really cool step. We recognize that not every team has a live chat support team. If that's the case, you can actually keep your customer agent assigned to the conversation. In case of live chat, let's say I selected that channel. You can actually keep the customer agent assigned so that it'll create the ticket for you and you can follow up with the customer later on in the support channels that you already have. But in this case I'm going to because I'm connecting email, I'm going to choose handoff to a human agent and I can select the specific user in teams that I would like. I'm going to go with Zyra and then if for some reason Zyra is not available, then we'll just choose to unassign the customer agent. And what that means it's going to move to your unassigned view so that your team can, you know, cherry pick and pick up those conversations as needed. All right, so now the conversation has been assigned a couple like tips that you may want to do. Let's say there's certain keywords or like conversations that you actually don't want going to your support team, or you don't want going to the customer agent and directly to your support team. So we've heard conversations about pricing, billing, canceling subscriptions. Those are more high touch conversations that you may want your humans to get involved in. And so in those cases, if those keywords are mentioned, it's actually going to hand off to the human. So before I actually demo what that looks like or give you a snapshot, I want to talk about actions. So everything I did just up until now is going to give you that average 50% resolution. But we've heard from our customers that 50% of their questions are actually questions that require like personalized information. So questions like, what is my order status? Or hey, I need a password reset. So we've built actions for, for you to be able to connect with your third party systems so that you can get that information back to your customers. So I'm going to click into this action right here, password reset. This is going to take you 10 minutes. It's really quick. You're going to go ahead and give it a name. So I've named it password reset. Give it a description of what this action actually is. So here, customer agent can't log in, send them password reset email. You're going to give it some trigger phrases. So these are the keywords or the phrases that the agent will look for to actually trigger this action. So phrases all around, like not being able to log in and then you're going to give it some inputs like what are the things that you're looking for? So like an email address and then there's some API information that you need to include. And then the last thing you'll give is instructions. So, so what are some things that you want the agent to do during while it's carrying out this action? So in this case I want to confirm that the password reset email was sent. The other one right here is order status. So real quickly, again, where's my order? Looking for order input. The inputs will be order number, email address, API information, and then response instructions. So with this again, you're going to get up to 90% resolution. We have customers seeing that. Um, so really, really awesome way to deliver even more Value to your customers and for your business. Um, so with that, let me just give you a quick. I sent this email just for the sake of time, just 54 minutes ago. But here you can see my initial email was, hey, my plant order hasn't arrived. Can you give me an update? And then we do some back and forth, you know. Sure. Would you like me to check the status and provide the time of arrival for your order number? Yes, please. And then it goes ahead and sends the status of my order. So all over email. Just within a few minutes, I was able to get the care, the help that I need and the personal answers that I was looking for. But we all know at the end of the day that content is not easy to come by or easy to create. It's a very time consuming process. That's why we've actually created something called knowledge gaps. This is where your customer agent is going to actually cluster conversations that that went unanswered. So in all of these situations, the agent actually responded by saying, I don't know. And that's the reality. You're not going to have content for everything. But we want to make it easy for you to be able to give your agent more knowledge. And so here we've clustered these conversations. See language requests. Can you please answer in English? In this case, there's only one conversation that's tied to this topic, but you may see multip and it's an easy way for you to then go in and create an answer. Short answers are going to be prioritized over your other content. Short answers are great to use when you have like, let's say you know, your website is down and you want customer agent to respond. That's a really for answering for like a really quick short period of time. Like that's a great instance where you want to use short answers. Otherwise you can also create knowledge base articles to be able to make sure that this topic goes answered in the future. And so yeah, with that, I'm actually going to hand it over to Amber, who's going to talk a little bit more about KB Agent and how customer agent and KB Agent work together to be the dream team.

All right, thanks, Jess. I will share my screen as soon as your infinity screen has left. All right, here we go. All right, thanks again, Jess. Hi everyone. I'm Amber, the product manager overseeing our knowledge base tools. So Jess has just shown you some of the impact that the customer agent can have. Super exciting stuff. I'm thrilled to be here today. Thanks so much for having me to be able to show you the next step and how you can continue to build on your AI agent skillset. With our new Knowledge Base Agent, which is currently in private beta, I first want to highlight the problem that prompted us to work on the Knowledge Base Agent. I love context. I imagine everybody else does too. I just want to give you a little bit of that today. As Jess mentioned a second ago, and as we all know, one of the most scalable ways that you can support your customers with self service is through your knowledge base. There's a bit of a kind of twofold impact there. When you have a really strong knowledge base, your users can of course reference that information on their own, reducing the burden on your support teams. However, if you've already implemented or after Jess's incredible demo, you are now compelled to implement the customer agent. And you already know that your agent is only as strong as the content that you're feeding it. So we want to make it as easy as possible for you to create new content. So taking a step back through your feedback over the years we've identified areas of opportunity to make it really easy for you to create and deliver new self service content through your knowledge base that presents itself in a really strong way to your audience. And we'll be continuing to work on how to make that editing experience consistent with your website and blog. That focus on developing tools that will enable you to customize your content and effectively communicate that information I think is a really great place to start. However, what about creating the content itself? I think that there's an opportunity here to speed up some of these knowledge base related tasks, particularly writing the content itself. I know someone in the chat was asking about kind of smaller growing teams. I think that this is a great opportunity for some of those smaller teams to get some support in creating the content. That can be a really difficult thing to do without dedicated resources. It's not necessarily your support and services core job. So there's two facets to this problem as you can see highlighted here. First off, creating that new knowledge base requires a lot of effort that your teams might not have time for. But then after that there's also this maintenance factor that continues to take up time and resources. But if we're not making that effort, you'll start to see a negative impact to the quality as well as the freshness of that content that can impact the customer agent's ability to provide strong answers. And then we see that increase in in the support burden. So enter Knowledge Base Agent. I mentioned that we have two problems to Solve. That's also how we plan on approaching the Knowledge Base agent itself. Today we're really focused on the first part of this problem, making it really easy to create new knowledge base articles just to quickly highlight, you know, what would happen if the customer agent is unable to find the answer to a question due to that lack of resources that I mentioned earlier. Those conversations will likely end up getting rerouted to your support teams, thus increasing that burden. And they've likely provided that answer before in a previous support ticket. So the Knowledge Base agent will use all of this information. Jess mentioned a minute ago that the customer agent will be able to produce knowledge gaps to identify areas of improvement in your knowledge resources. So the knowledge base agent will be able to utilize those knowledge gaps and then compare those inquiries to your portal's tickets and then extract answers that your live agents have previously provided for those questions. So let's take a look together at the Knowledge Base agent. You'll be able to navigate into your knowledge Base dashboard and view AI generated article topics. Here you'll be able to view the sources for that article content. So as I mentioned a minute ago, the agent will be using both knowledge gaps as well as tickets. The knowledge gaps are informing the inquiries, as I mentioned, that are coming in or you know, the topics that those articles will have. And then the tickets will be informing the answers to those unknown questions, thus kind of creating that new content. So if we open up one of these AI generated article topics. Oh, I've just realized that I was not sharing. Okay, well let me just exit and show you that again. Okay, so we came into here, we came into our Knowledge base article dashboard and then we navigate over to the Knowledge Base agent and you'll see those AI generated article topics. We'll view the sources. So as I mentioned again, here are the knowledge gaps and here are the tickets. So you'd be able to click into each of these just to view more. And we'll talk about the roadmap in a minute here. But we do want to kind of tighten up this feature feedback loop as well so that you'd be able to provide feedback on these sources in the future. And then if we click into one of these article topics, you'll be able to see that AI generated article content right within the article editor, which is going to save your reps time and effort in manually creating this content. And this content was extracted right from a support ticket that existed in my portal. So super exciting stuff. Let me go back and make sure I'm sharing my screen. Screen Again and I just want to highlight what we've just talked about. So I did see a couple of questions about the status of where we're at with some of these features. Customer Agent is currently live. However, some of those newer features that Jess mentioned a minute ago, including the email channel as well as actions, those are in public beta. Knowledge Base Agent is in private beta as I mentioned. So you should be able to see all of that in your product updates. But feel free to give a shout if you're not seeing that. And then one other really exciting thing I want to mention while we're here, while we have you, is we announced last week that the Customer Agent will be available in all hubs, not just service hubs, starting in June. We know that there's really meaningful use cases kind of across your business, across marketing and sales as well. So that will all be available through a new pricing concept called credits. So more details on that to come. We won't go too deep on that today. And then lastly, just to highlight what is on the roadmap for Knowledge Base Agent, I mentioned earlier that we want to make it really easy for you to spin up a new knowledge base. So as it exists today, Knowledge Base Agent, as we kind of saw in that quick demo that I mentioned, the Knowledge base should exist for you to to activate Knowledge Base Agents. So we want to make that process from start to finish really easy so that you don't have to kind of go through that process at all and the Knowledge Base Agent will be able to create it for you. I also mentioned that the second facet to our problem is maintaining existing articles. So let's say that you have articles that have been around for several years and need some updating. We want to work on having the Knowledge Base Agent scan through your ticket data or even external data through that you've uploaded similar to the Customer Agent and be able to provide enhancements to your existing articles to ensure that that maintenance piece is taken care of as well. And then insights on some of your top created issues so that you can continue to track where the most traffic is coming in for your support team and how to potentially enhance your self service resources based off that. So Jess, would you like to chat about the Customer Agent roadmap real quick before we move on?

Yeah, thanks Amber. So real quickly, Customer Agent, we have a lot going on. We're really excited about the next couple months Right now Customer Agent is really good at answering support queries. It even will answer pricing questions if you want it to, or answer questions about just your products that you have in general, but we really want to be able to scale what it can do for sales use cases. So think about lead qualification and booking meetings really being able to solve for the entire front office. So we're really excited about what we're going to do there. We've also heard from a lot of customers the need for multiple agents. If you have multiple brands or maybe you have your different audiences that you want to cater to, that will be coming later this year as well. And then a concept called guidelines. It might be called something else, but guidelines essentially are a way for you to be able to give your agent more tailored instructions through natural language so that you can really have it reflect your brand or make sure that it's producing like accurate and personalized content for your customers. So really excited about where we're headed with that. I'm going to pass it over to Smurdy who's going to talk about the rep experience.

Awesome. Can everyone see my screen?

Looks good.

Great. So hi everyone, I'm Smriti. I'm one of the product managers on Help Desk and today I am super excited to share about some of the powerful new features that can make it up to 65% faster for your human reps to reply to tickets. I'll be highlighting updates in three specific areas. First is enhanced organization for large teams. Second, some of our latest AI powered features and third, significant productivity and efficiency boosts for agents. So with that let's jump into Help Desk and do some demos. So let's start with organization for large teams. On a high volume support team, a well organized workspace is super critical. It helps ensure the right people handle the right tickets quickly and keeps distractions to a minimum. To promote a focused and well organized workspace, we've released two new features, Views Customization and Spaces. So starting with Views customizations, this is available in private beta in Help Desk settings and you can see all the available views in your portal at a glance in including private views. So this gives you a sense of view usage for your portal. You can also set a default homepage for the portal so everyone starts their day in the right place. So in this case I am going to say that I want all my reps to start their day in the assigned to me. So I'm going to make the assigned to me view the default homepage and also I don't want my reps to see all open so I'm going to go to the all open view. With this new Views customization experience, you can manage settings for and sharings for both default views and custom views. So back to all open, I'm going to edit the view and set it to no one which is going to hide it from the portal. So that's views customization. Moving on to Spaces. So with Spaces we have a brand new way to organize your team's work. They offer dedicated working areas in Help Desk and they reflect tickets for the team or or teams of the space so agents can focus on their own issues without getting bogged down by other team issues. Spaces are really easy to create and manage, so to create you just hover over this create cta, you can edit spaces and delete spaces quite quickly. And then over in Help Desk, once you've create created a space, your agents will see spaces in this space switcher. They can navigate across spaces here. One thing to call out is spaces also come with custom views which are contained within the space and are automatically filtered down to the tickets of the teams of the space. So as you see here, I have I'm in my Tier 1 space, so I have my Tier 1 custom views. If I jump over to Tier 2 space, I will see Tier 2 custom views and so forth. So that's all on team organization. Overall, we highly recommend using Views, customization and Spaces together as they effectively allow you to both organize and customize the entire workspace for specific users and teams. Now let's switch gears and jump over to AI. We have some really exciting AI features in Help Desk that can really accelerate support rep efficiency by bringing information to their fingertips and speeding up certain actions. So starting off with Ticket Names first, we have this new AI generated ticket Names feature here. If your ticket has an associated conversation, Help Desk can use AI to generate a ticket name. So simply go to the top of the ticket and click this icon here and Copilot will generate a new ticket name based on the context of the conversation as well as ticket properties. To sort of run with the ticket name, just hit Insert. If you'd like to see new recommendations for ticket names, hit regenerate. Next, we have AI generated ticket summaries. So say a ticket is handed off to a new rep without a lot of context. Instead of spending time having to click through that ticket to get context, agents can now use Copilot to generate ticket summaries. So just go over here to the sidebar and that and you have a ticket summary. Next, be sure to check out the new AI Knowledge Sources sidebar card. This is also AI powered and agents can use keyword search here. To find relevant knowledge based articles to help resolve tickets. And even better, if configured the system will provide automated suggestions. So in this case I'm going to search for biking events. I'm in a biking portal and here you see that the system has served up a relevant knowledge base article. And last but not least, AI generated reply recommendations. So you can use these to help generate responses to customer requests in a single single click by clicking this button here. These save your team your team time, allowing them to focus on delivering high quality customer experience and improving overall productivity. So that's it for AI. Let's switch gears to the newest non AI productivity tools that can help boost agent efficiency. So first up is the board layout. Many of you love the visual Kanban style from our CRM and now we've brought it directly into the workspace. The board offers an intuitive way to visualize how tickets are moving across pipelines and quickly surfaces tickets that need action leading to faster response times. If you're a pipeline driven team, we highly recommend checking out the board view. Next is the task sidebar card tasks are really useful in scenarios where there is a multi step process to resolving a request. So in this example I'm going to create task to update the product team when the ticket is resolved and I'm going to set a reminder for tomorrow and create my task and the task will generate up here at the top. Last but certainly not least because I'm personally really excited for this feature is Send Options. Send Options combines message sending and status updates into one seamless action by enabling agents to do these actions in a single click. So you just type in your response here, go down here to the bottom right and select from this menu of Send options. Okay, so as you've seen from improved team organization with spaces and views customization to productivity gains like AI generated ticket summaries, the AI Knowledge sources sidebar card and Send options, the latest Help Desk updates are all designed to really empower your support teams to speed up replies and deliver exceptional. Many of these features are still in beta, so highly encourage you to check out the product update page to learn more and request access. I will now pass it over to Mark to talk about tools that can help managers improve agent productivity.

Perfect. Thank you so much. All right, let me start by sharing my screen. Can everybody see my screen looks good. Perfect. Thank you so much. Great. So hi everybody, my name is Mark, I'm the PM of Full Service Hub Reporting and I want to talk to you about how we are enabling managers to make data driven decisions in a timely manner. So that's kind of the goal that our team is focusing on. And we always focus on like three main jobs to be done of leaders and managers of support teams. We want to enable them to monitor operational health, coach their teams and plan ahead. So these are the three topics that I'm going to talk about today. How do we do that? Essentially the framing we always using is we're going to enable managers and leaders to access more data points and we will enable them for easier access, easier to digest data visualizations. So I'm going to go back and forth about this concept in the demo in a minute there. In that demo I'm going to give you a quick sneak peek just to give you a high level overview of what we're going to cover. So I'm going to cover three different areas here. I'm going to talk about automation that managers want because they want to have timely decisions being made. We want to talk about monitoring operational health, how we enable that in the help desk and then also of course optimize self service and AI. So without further ado.