Scaling with Soul: Balancing Automation and Human-First Leadership in Digital Customer Support and Success

[Speaker 1]
All right, thank you so much. Hey, everyone, it's great to be here. Again, my name is Kayvon, and we are here to chat with you a little bit about how we've scaled with Sol, right?

Like, I think AI is the big term right now, but in our progress through, in our journey, I would say we've taken a very sort of intentional approach through different pillars we're gonna walk you through, and AI is now the next iteration of that. So, you know, high level, we'll talk a little bit about our vision, our background, then we're gonna dive into the pillars, and we're gonna make sure we leave some time at the end for questions. There are a lot of, I don't wanna say a lot, but there's a handful of polls, so if you can have the app out, we're gonna have some engaging parts of this session as well.

And so, I always wanna start with, you know, a question. So, I'd love to hear from all of you where you're coming from, where you based that off that you came all the way to London to join for this conference. I'm sure some of you are local, some of you had a longer journey, and while those results are coming in, I'll tell you a little bit about myself.

So, I began my journey in tech in 2007 when I joined Google. I moved to California, spent four years there, and then actually moved not too far away from here to Dublin, Ireland, where we had our European headquarters. And a lot of my work there was really focused on scaling support for Google's consumer products, things like Android, Chrome, Gmail, YouTube.

And in 2015, I relocated back to the U.S. to Salt Lake City, Lucid's headquarters. And more recently, I moved to Raleigh, North Carolina, which is our newest but fastest growing office. So, quick background on me.

Do we have the results pulled up? Perfect, all right. So, London and Dublin are the big winners.

Good to see Italy in third, because Azura hails originally from Italy. And then Europe is very well represented here. We also have D.C., very, very cool. So again, started my journey in Google, just some of my favorite photos from the early days. And then more recently at Lucid, first in Utah, more recently in North Carolina. Currently at Lucid, I'm the VP of Customer Ops, which is all of support, and then the scaled parts of sales as well as customer success.

Sorry, the scaled parts of customer success. So, I will get into some more of those details in a bit. And Azura, tell us about yourself.

[Speaker 2]
Yeah, very nice to meet you all. Thank you for coming, and benvenuti for the Italians. My name is Azura.

I was born and raised in Italy, and then I moved a little bit around Europe. I lived in Barcelona in Belgium, and then I moved to Amsterdam two years ago to join Lucid Software. Started in the sales team, and then decided that my goal was mainly around building relationships with customers, and I decided to join the customer success team.

And I'm currently in the team of Cave, and in the scaled customer success motion. It's quite a new motion at Lucid, which we're gonna touch on a little bit today. We try to work with our long tail segment.

So, we have a group of four people working with 20,000 accounts across the globe. So, it's a very different approach from most companies, and we're happy to talk about it a little bit more today.

[Speaker 1]
Awesome, all right. So, I wanna talk a little bit about our vision when it comes to CS, and a little bit of more of like our philosophy. And I always like to start with analogy.

When you drive up to a bank during normal business hours, you're given a decision that you have to make, right? Do you wanna walk in, wait in line, and speak with an agent? Or do you wanna go to the ATM and self-help, right?

I think that's a fundamental choice that you have. And I think, I don't know about you guys, but for me, 95% of the time, the transactions I actually want, I end up helping myself with an ATM. Now, if I wanna get a loan to buy a house, or something else that requires that higher touch service, I'll go and wait in line.

And I think when we think about Lucid software, and again, I just wanna emphasize, we are not Lucid Motors, so if you came here to talk about cars, you're in the wrong place. We're Lucid software, and I think a lot of our customers find out about us by doing a Google search. They may say, I wanna make a flow chart, right?

And then they self-discover our software. They'll use it, we have freemium models, right? So they'll use it, and then they'll subscribe for it on their own without talking to a sales rep.

When they run into an issue, their first instinct is to do a search either on Google, or on one of the AI platforms. And then if they want more tutorials and stuff, they'll go and self-help our skills, our resources, et cetera. So again, I think it's different for every model, but for us, where our customers tend to like the self-serve motion, this is largely successfully.

If given the option, they will choose to self-serve. Now, there's exceptions. If there's billing issues, right?

If they wanna strike a deal on an enterprise account, right? There are definitely the exceptions there, and we're gonna talk a little bit about that. And so, you know, that's really the heart of a lot of what we're gonna talk about today.

So I'm curious now, right? Like, are you guys familiar with our software? So again, another poll for you.

Just wanna know how many of you have heard of or used Lucidchart, and then LucidSpark. We also have AirFocus. I didn't include it on here because it's a little bit less known.

But are you familiar with both products, one product or neither? And we'll see how those results come back.

[Speaker 2]
Interesting.

[Speaker 1]
Wow, okay. Very cool. Well, the good news is, we're not gonna be doing anything in the product.

We're gonna be talking very much philosophically about our support strategy. But definitely check us out if you haven't heard about us before. I would say our biggest competitor in the charting space is Microsoft Visio.

If you've heard of that. Some of our competitors in the whiteboarding space include Mural and Miro, right? So maybe you're familiar with some of those products.

All right, now I'm gonna pass it off to Azura to talk a little bit about our four pillars.

[Speaker 2]
Absolutely. So we want to start talking about how we've been working with our customer success and support teams. As I mentioned before, we manage a lot of accounts on a global scale.

And despite our customer base consistently growing in the previous years, we've been able to keep more or less the same amount of support tickets. So how were we able to grow and at the same time not overwhelm our support teams? We use this CS flywheel model that is based on a four pillars approach.

So first and foremost, of course, we work with dynamic content. So as Kevin mentioned, most people want to find the answers immediately when they need them and they want to self-serve. So we made sure to have a strong training catalog where customers can come in and self-serve whatever they need.

For peer communities, we have created a community using insights. And what we think is that most of the times when you need an answer to something, the best person to help you is somebody that has already gone through the same struggles, that is a peer like you that is trying to solve the solutions in the same tool. And so we encourage knowledge sharing within the community.

And the third pillar is intelligent agents. So that's where AI comes in. AI is good at responding to things fast, at easy questions.

So we always make sure that we have AI as well among our pillars. But then when it comes to complex questions, escalations, frustrations, people want to talk to human beings. I think we've all been in the situation where we're going on loops with an AI bot that is not helping us in a support chat.

We want to avoid that. And so that's why we make sure that we always have our human leverage, people as well in our pillars of success. So we'll pass it on to Kevin to talk a little bit more about our first pillar that is dynamic content.

[Speaker 1]
Thank you so much, Azura. And I'm just gonna backtrack a little bit to just gloss over this slide, right? When I started in 2015, the next year we had 9 million users.

2016, Lucid had 9 million users. Today we have over 100 million, right? And so tremendous growth over the last decade.

But what you're seeing here is our support and CS volumes over that same time period. And it's actually gone down, right? And I think you guys are all in the industry and you know that this is not something to be celebrated in and of itself because we all know that we can hide the contact us form and make it very hard for users to get to an agent.

But we haven't done that. Anyone in this room, regardless of whether you have an account or not, you can just email support at lucid.co or customersuccess at lucid.co and submit a ticket. You can bypass all the pillars that we're gonna walk through right now.

And we always wanna give our users that option because then it's gonna force us to make sure that the automations we're putting in place are actually adding value, right? We really want our customers to know that if they want a really quick answer, our automations or our AI are gonna be able to provide that. But if they're adamant that they don't want that, great, right?

There's another way you can go to the community. You can reach out to our email support. And I think this is really important.

I know we talk a lot about deflection these days, right? Saving money, decreasing volumes. And I could talk about deflection here, but I want everyone to know that we emphasize resolution.

That is the holy grail for us. Deflection is something that we look at, but for us, deflection is moving something, let's say, to the community where a question is responded to once and other people can benefit in perpetuity, right? And so I'm gonna give you an example, actually, of Apple.

One of my favorite stories of Apple is when the iPad was released, a customer purchased an iPad, a gentleman, and then he sent it right back to Apple without even returning it. And then he just put a little Post-it note on it, and he said, wife said no, right? What did Apple do?

They refunded him, sent the iPad back to him, and said, Apple said yes, right? And I think this is just a good example of resolution is what we're really trying to accomplish here. We want the customers to be happy.

We're not viewing all of these support strategies as a way of deflecting costs. We are very happy with the cost of our support and CS teams relative to our revenue, but we want to do better, right? And we think that there are ways that automations and AI by providing an immediate answer is actually gonna be fundamentally better than the alternative, right?

Not for every case, but that's something we're gonna talk about. All right, so the first pillar, dynamic content, right? For us, this is the help center.

We think that a very robust user manual that represents 90% of the questions that customers have is the foundation for, again, the self-help model, right? Customers who are finding the product on their own and they wanna self-help on their own, they need these rich resources. And so when we think about our help center, we're viewing this as the user manual to our products, and there's a cost that goes to this, right?

Our product is in 10 different languages and every help center article is translated into 10 languages, right? So we have about 300 articles collectively representing about 350,000 page views per month, right? So it's clearly widely accessed, and it's also important because as we move from a world of SEO to AIO, right?

This is also driving how different tools like ChatGPT, Gemini, Claude are accessing the correct answers, right? In terms of when customers are asking them questions. All right, talking a little bit about what I like to call the long tail, right?

We have about 300, let's say 300 articles in the help center representing about 350,000 page views. Well, the community has about 14,000 articles, right? It's the long tailed content.

Users having very specific questions. An example might be like, oh, my Lucy chart app for Android is not working on a Samsung Galaxy 18, right? That is such a specific question that we're not gonna have a help center article and translate it into 10 languages, right?

But there's a chance that 100 million users, someone has faced that exact same problem and they can help others by providing the response. Now, community doesn't come back without its own drawbacks, right? It is generally, I don't wanna say English only because it's not always English, but whatever the user chooses to submit the question in whatever language, that's the language that is gonna be the default.

We do have AI translations in our community, but we all know AI translations are not always as good as we would like. But again, the power of the community is not just that you can really address the long tail content. You can answer a question once and other users will benefit from the answer in perpetuity.

But also, if you look at, again, this move from SEO to AIO and how these algorithms are ranking different resources, one of the big things that we've seen in the early days is that this conversational nature of communities of Reddit rank very, very highly in trusted sources for AIO, right? So community becomes such a powerful tool from that perspective. And again, just some of the bullets that I went over.

The real challenge here is when you have over 10,000 posts in the community and a SaaS product like Lucid, we have a new release every single day. Things constantly change. It's hard to keep these up to date.

And we're actually trying to use AI to solve that problem, to let us know what's out of date, but it's not easy. And then the other benefit of community is this whole notion of a community and peers helping one another, right? This is something that we've really been focused on is how can we incentivize users to help one another?

Because that is truly the, for most of these issues, a good user response is far better than what our team can provide. Example, like the one I gave earlier, right? No one on our team has a Samsung S18, right?

And so it's gonna be harder for us to provide the right answer. But I'm sure one out of our, at least one of our 100 million users ran into that same problem and can provide a very direct answer. So peer support is climbing quickly.

And again, for us, deflection is not cutting costs. What deflection is is when you post a question in the community, you're not just helping that one user with an answer, right? You're helping all the other people who are benefiting from this answer, either through Google searches that lead them to the community, or even through what I was talking about earlier with the AIO, right?

Like tools like Claw, Gemini, answering the questions. We call them zero clicks because they don't actually link to our community, but they're surfacing the answer. And this becomes critical for our success.

And this is our third and newest pillar, AI, right? It would be impossible in this day and age, I feel, to give a presentation without talking about AI. But I think you'll see we take a very different approach to AI from a lot of companies, because again, our goal is not to cut costs.

Rather, our goal is to provide the best user experience. And so what does that actually look like? And before I dive in, I always like walking people through this hypothetical scenario, right?

If we figured out AI completely, and every question that a user asked was immediately solved by AI, and the user got an immediate answer, the problem is you're not getting any new content, right? As the product evolves, the AI corpus becomes out of date. And so within weeks, your answers are gonna be wrong because AI is referencing a stale content corpus, right?

And so for us, the actual goal is the following. If a question has been asked before, and we're confident in the answer, we wanna give the answer to the user because it's immediate and they don't have to wait. However, if the question is the first time being asked and we're not confident in our answer, we wanna post it in the community because that's what's gonna keep the content fresh.

And then the next time that same question is asked, then the next user will get an immediate answer, right? So that's kind of what we're working towards. And I think at the keynote, you guys heard about the four thought integration, right?

That Zendesk is using, Gainsight is using. That's what we're starting to pilot in this space. But again, I just wanna emphasize, for us, AI isn't replacing anything.

It's just augmenting our existing strategy. And so I'm gonna take a quick step back and talk a little bit about what we are doing with AI and what we're not doing. And again, we're being very cautious and data-driven.

We're actually only allowing AI to give the answer once. If the user responds, it goes straight to an agent and we figure out like, why wasn't that response enough? Did we get it wrong?

Were we incomplete? How can we make the AI better? We're building on our strengths, right?

Like we are constantly figuring out ways of optimizing workflows. It started with the Help Center. We looked at what tickets were coming in.

What were the same questions being asked over and over that would inform our Help Center content? Then we thought about the long tail, right? The questions that are more user-specific and that's where the community came in.

Now AI, it's really the allure of AI, in my opinion, is that immediacy of answers, right? Especially when you're in the product, you don't wanna get out of the product, you want an answer right away, right? And so that's what we're doing is building on our strengths.

We're up-leveling roles on the team, right? AI is not replacing these really, really hard, charged conversations with customers. It's replacing what I like to call kind of like the low-hanging fruit, right?

Like the basic, how do I draw a straight line in Lucidchart, right? Like not challenging questions. We wanna get those handled by AI so our agents can really up-level and focus on the moments that matter, the magical moments, right?

And then also keeping up with the industry. What we found is today, AI doesn't fit a lot of the use cases that we had because we got rid of a lot of that low-hanging fruit through the help center, through the community. What remains now is actually a lot more challenging than perhaps some other teams that took a different strategy.

And so I wouldn't say we've had these crazy gains through AI, but we're ready for the moment. We know that when AI gets really, really good to tackle some of our thorny issues, we're gonna be ready. It's almost like tuition for our team members to learn AI.

What we are not doing, right? We're not getting caught up in the hype. This isn't changing everything.

We're not trying to deflect costs. It's not replacing people. I showed you the ticket volumes going down.

What's actually happened is we're spending less bandwidth on reactive responses to user questions and more time on the proactive stuff, right? We're sending, we're triggering emails under the right circumstances when we see an expansion opportunity or potential risk, and then we are proactively tackling that. And it's, our team has only grown, right?

But the day-to-day has looked a lot different. And obviously, it's all about resolution, right? We wanna support our users, and that is what this is really focused on.

So I'm gonna talk real quick about our support use case, then I'll talk a little bit about our sales and CS use cases but in support, prior to the advent of AI, we basically had a form in our help center that users would self-select their issue type, right? Did you lose documents? Did you run out of storage?

Do you want an educational account? And depending on what they selected, right, when they submitted their question, we gave them an immediate answer. And it was only purely based on what they selected, right?

But it was like, hey, we got your ticket, we're gonna respond, but in the meantime, let us know if this helps, right? And obviously, if they responded, we knew we didn't help them. And so that was a great win for us because the user got an immediate response, but if it didn't help them, they would have gotten the same response time that they otherwise would have.

Now, there are some limitations with that, right? We can't make the form so advanced that users get lost, right? So we're limited to about five different issue types, but now what we're doing is we're allowing AI to look at all the emails coming, regardless of whether they're coming from the form or just sending an email to support at Lucid, and we're looking for these categories that we think we know the answer to.

And the AI self-selects these into categories. If it's a category we're confident in, we send the response. If the user responds, that's it.

AI had a chance, it's gonna be responded to a human after that. That was a very conscious decision, and I think a lot of the companies that we work with, like Zendesk, they're against that. They're like, no, we really want AI to handle it from start to finish.

Well, guess what, we don't want that, so we're not gonna do it. Okay, and so talked a little bit about the benefits of how AI allowed us to extend our auto-reply strategy, and this walks through some of this. But the greatest benefit, I think, is with AI, you don't have to worry about the complexity of a form.

You can give it as many categories as you want, and you can constantly iterate, and that's been the huge use case for us. On the CS side and sales side, we're gonna talk a little bit more about that, but it's around signals, right? When we have customers who purchase 50 accounts, and 49 of them have been allocated, they're right near full capacity, right?

So we proactively reach out to them and say, hey, did you wanna purchase more licenses? This is how you can do it. Conversely, if they purchase 50 licenses, and they've only assigned five, and it's three months out, right, then we send a different type of email, which is like, hey, there's a lot of people at your company who could benefit from this, but you're not using all the licenses that you've purchased.

And we put them through flows that doesn't just change the, like, they don't just progress in the email flow after we send the email. We're actually measuring, okay, once they got the email, did they read it, and did they allocate those additional licenses? If not, we're not gonna move forward in the progress.

We're gonna keep hitting that same point, which is like, okay, you have unused licenses. Let's get them used up. All right, and then human leverage.

I'm gonna pass it off to Zura, but before I do, I just wanna like reiterate. I'm gonna quote myself here. Unchecked AI containment produces cold, sterile customer journeys that frustrate buyers and burnout support agents who inherit escalated resentment.

Automation without intent erodes trust. And it goes back to what I said earlier, which is use AI to provide immediate responses, but allow users to opt out when you've gotten it wrong, right? You don't wanna be one of those call centers where you call and you're just like, agent, agent, zero, zero, zero, and you're just going in loops, right?

Allow the data to tell you what customers are finding useful and what they're not finding useful. And with that, I'll pass it off to Zura.

[Speaker 2]
Yeah, so the idea here is quite simple. When it comes to very simple requests that come in, I'm thinking about a password reset, account access my account, very simple, straightforward questions where speed is the key. The AI is great because it will answer immediately.

They don't have to wait even for 10 minutes for a person to come in and respond. They immediately get their answer and we move on. However, when it comes to escalations or issues that create frustrations, that's when a human is genuinely needed.

And also when there's churn risk, like we're seeing an account is actually at risk. We don't want an AI to be in there and doing its robotic conversation. We want to have a human that talks to them that can actually understand what's the underlying issue and try to find a solution for it.

So this is kind of the comparison between a transactional AI model, which a lot of companies are nowadays trying to implement, and what we call the connected CS flywheel. So as we mentioned before, the main difference is what we're looking at. What is the primary North Star?

So with AI transactional models, what we're looking at is ticket cost reduction. That's not what we hear about at Lucid. That's not about cost.

It's about actually creating a good relationship with customers, a good customer lifetime retention, renewal leverage. That's what we're looking at. It's not only about cost at the end of the day.

But also when it comes to the customer journey, as we said, so many customers end up being in loops, fragmented walls, difficult conversation that never get resoluted. Whereas when we're looking at the connected CS flywheel, we actually have seamless handoff with good context because the person will come in to the AI, giving the problem. The AI can ask maybe some follow-up questions that are gonna be useful at the end of the day to the human, but it's the human that needs to come in and actually talk to the person.

And also when it comes to the agent role, the AI is simply triaging, answering, immediately passing it on. Whereas the people, they become advisors more than just a support ticket taker. They actually start becoming consultants to the customers.

And this, of course, also leads to better engagement and loyalty as well to the company, right? Like if I talk to a company and I'm trying to get support and I'm only met with the robots, I'm not gonna trust them. If it's a person that talks to me, understands me, empathize with me, then probably even if there's an issue, I will trust them or believe in them.

So what are some ways in which we can still implement AI without breaking that trust? First of all, it's so important to be transparent. When I'm talking to an AI, I want to know that I'm talking to an AI.

I want it to be clear. Of course, eliminating that robotic corporate jargon. I think we all have a bit AI fatigue.

We're all tired of seeing the same type of sentences. So we want to avoid that and also give them an exit pathway. If I don't wanna talk to an AI, I don't wanna be forced talking to an AI.

I wanna move on to a human. And of course, ensuring that smooth dialogue history. Transfer as well so that the customer doesn't have to repeat the same issue that they've already told the AI to the person as well.

We want to make sure there's all the context that's needed. So when we look at AI and human leverage, a lot of people ask themselves, is it actually true at the end of the day that human leverage is so important? Is it factual or is it just people talking?

Well, if we look at the data, they speak for themselves. So actually, this is a research from Gartner. A high value human intervention at critical touchpoints drives an 82% probability of contract renewal or expansion compared to contained account cohorts.

If we look at some more data, this is quite recent as well. 29% more empathetic humans, of course, compared to robots. We're not surprised about this one.

77% of people, of customers, report better outcomes when they're talking to humans instead of AI. And 55% just prefer speaking to a human being. We're humans at the end of the day.

That's what makes us strong. That's our strength is the relationships. It's helping people out.

So we're gonna look at some of the best practices and I'm gonna pass it on to Kevin to talk about this one.

[Speaker 1]
Thank you so much. And I know we're nearing the time, the end for the presentation. I wanna make sure we have some time for questions.

So I'm gonna go through these pretty quick and we can go in more detail if questions come up. But again, you want the AI to be very targeted, customized, and specific, right? And allow for an opt-out.

Like for us, the magic formula was AI gets the first app, but then anything beyond that goes to humans. And I think as we get more confidence in the AI, we can adjust that algorithm. Our sales and CS automated campaigns, they focus on completion of goals, right?

If you send an email and the user doesn't progress to what you need, don't send them the next step in the flow, right? And I think Journey Orchestrator and other tools make that very easy to do. And then for support use cases, right?

Like again, when you're being proactive, make sure you're using valuable data, right? To make informed decisions. All right, the other thing is like bring your team as part of this journey, right?

We all know that there's a lot of skepticism, but also fear of AI, the impact of jobs. When they see how their job is evolving and they're able to be more proactive rather than reactive, right? They're coming along for the journey and that makes such a huge difference.

You're getting rid of the tier one work, which is lower on the, you know, just curiosity perspective. And I think they'll appreciate that. All right, last slide.

I know that often when we talk about AI, we tend to focus on AI replacing people. But in my opinion, it's actually the opposite. These are the key roles.

Like if I were to build the CS team or support team, these are the roles that really highlight what we've talked about today. You know, from the content specialist, writing help center content, community manager, data analysts, systems admin, setting up the AI, right? These roles haven't changed in terms of their titles, but they've evolved in terms of their functions.

And so you want team members who are passionate in these areas. And I just wanna end it by saying, when you choose resolution over deflection, you build loyalty, right? Loyalty builds revenue.

We no longer have to choose between depth of care and operating at scale. AI allows us to handle both by really bringing forward the moments that matter and allowing the agents to focus there. We see this motion driving retention, which drives revenue and revenue drives our mission, which is to empower users to see and build the future.

We're focusing on providing the right solutions to the right customers at the right time on the right platform, right? The future of service is not less human. It's actually more human and humans that are emboldened by AI that allows them to do more in the operations of their day-to-day.

So with that, I would love to open it up for questions. And I know that we have an app, like in the app, you can submit questions. Hopefully you've been submitting them.

But yeah, thank you so much. Thank you.

[Speaker 3]
Okay, we have got quite a lot of questions. So I'm just gonna quickfire straight in. How do you incentivize peer support?

[Speaker 2]
Yeah, so we have actually a program that's called the Lucid Legends. It's handled by our community team. What we do is we recruit users from companies that are interested in participating in a program in which they will get some incentives such as free licenses.

They will get to talk to our developers. They will get to see the new features firsthand. They can have regular meetings with our community team, live meetings.

Again, people really are craving that human connection. So we create networking opportunities for them. They get also some nice gifts like swag and whatnot.

And in exchange, they offer to talk to other people in the community. So they actually go in and respond to questions on a daily basis.

[Speaker 1]
One last thing to add there is we have an annual hackathon. And now we invite our Lucid Legends to it. And in fact, they have an award that they can give.

They submit their own votes. And now there's a Lucid Legend award. And often what they vote on will make it into the product.

So just another perk. I love that, okay.

[Speaker 3]
Okay, so keeping with the let's crack on. Which team is responsible for creating and validating resources, customer success, or support?

[Speaker 1]
This is a great question. I would actually say it's not any one team. Our content creators live on our support team.

And we call them support content. And they do all the content for our help center as well as a lot of the product or content in the product. But everyone at the company surfaces when things are out of date.

We have a Slack channel for it called Hashtags Help Center. And everyone from CSMs, solution engineers, product managers, product marketing managers, they're all putting in requests. Hey, this is out of date, can you update this?

But if you're asking who's actually doing the update, it's the content managers within our customer operations team.

[Speaker 3]
Thank you. How can we encourage customers to use the help center instead of contacting the human support team directly?

[Speaker 2]
I think one of the keys here is definitely leveraging keywords and making sure that as soon as they go and Google the question, they immediately are redirected to the correct help center article. Also, as mentioned, we have hundreds of help center articles. They really go in depth.

They're really well organized and it's really easy for customers to access. We also think that most people prefer to just find the answer themselves. At least like even talking for myself, as the ATM comparison, we all kind of wanna do it on our own if it's possible.

So just making sure that they have the correct resources immediately when they Google it is the best way to go.

[Speaker 1]
And the one last thing I'll add is that when they ask us a question in email, we often will just link to a help center page for you, right, directly in the article. And then if you do this enough, they'll be like, if I email support, they're gonna send me the link to the help center. And guess what?

You can do that a lot faster yourself.

[Speaker 3]
Okay. Do you have many high touch customers and have you had to battle against the unwillingness to use community or engage with AI? If so, how have you dealt with those customers?

[Speaker 2]
Yeah, that's a good question. So we do have some different motions depending on the type of customer. We have the regular customer success managers that work with the biggest clients, so enterprise clients, and they have more like of a one-on-one high touch perspective and relationship with them.

We also try to regularly, every like few months we get new accounts that we try to engage with. So we alternate, every few months we get new ones that we work on so that we're always able to kind of maintain a sort of relationship with big accounts. Of course, there's a lot of people that don't want to engage with AI and we respect that.

We don't force it upon them. There's always the possibility to opt out of AI within our platforms. We really believe that people should have the choice.

And when it comes to talking to support and interacting with an AI, we made it very simple for them to overcome that and go and talk to a human. So the issue is not really there. We try to make sure that they always have the possibility to opt out and it's not a problem for us.

[Speaker 1]
We also have premium support, which we sell. And if you buy premium support, you can call us at any point with any question. But there's a cost to it, right?

And some people, like I said, AI and community and help center is not for everyone. And we offer that option for those as well.

[Speaker 3]
Okay. How big does your user base need to be for a community to be the right solution?

[Speaker 1]
This is a good question. I will say, well, two things happen for us. We started on Zendesk Community.

I don't know if any of you are familiar with Zendesk Gather. They kind of add it to their normal platform. It's not very good.

There's no leaderboard. There's no badging. The metrics are not good.

We had that from 2016 up until 2020. And it did the basics, but we had no lucid legends because there was just no incentives. I would say for us, again, every product is different and every SaaS offering is different.

For us, the sweet spot, I would say, was around 50 to 60 million users. Right now at 100 million plus users, we have about 130 lucid legends that are doing the majority of the work in the community. But again, I think if we didn't start in the early days, we wouldn't have learned everything that we did to get to where we are today.

So I would say start early, but I'd say the magic spot for us was 50 million users.

[Speaker 3]
That is an impressive recognition program. How do you moderate your community content?

[Speaker 1]
Good question. A couple ways. So all of our lucid legends are in there and they have ways of flagging the content and making it to our community managers.

Our community managers are doing this as well, but also every user has the opportunity to flag something as inappropriate or a violation of our terms of service, and that comes to our team. And we have a terms of service. If you break the terms of service, you're gonna get censored, right?

It is what it is. We try not to, right? But sometimes things get toxic and we have to protect our users.

[Speaker 3]
Okay, we might be able to squeeze two in. Let's see how fast the answers are on these. How do you distinguish support from CS?

[Speaker 2]
That's a great question. So we actually sit together. We sit next to each other.

Remi from customer success and the support team. The main difference is that the support team is mainly dealing with very technical questions. So it's more about, right, something is broken, there's a bug, there's something billing related, or those type of questions, that's where support comes in.

They also take care of filing bugs, for example, coordinating with the product teams on those, making sure that there's the proper escalations. Whereas customer success, especially our at scale program, we work more on use cases, supporting clients to understand what different type of features they can utilize in different departments, making sure they have the right enablement resources. So if a company's, for example, working on creating project planning and doing agile, we want to make sure that they have access to the correct resources in the help center, for example.

We'll put together a board with all the different help center articles, the different training catalogs, and we offer to guide them through what type of options they have to develop those use cases internally.

[Speaker 3]
Great, okay, thank you. One more question. How much impact your community, how much impact your community has had on your support ticket volume?

Did you see a measurable decrease after putting it in place?

[Speaker 1]
Yeah, a couple things. Number one, I would say it's not like a overnight spike in any way. And again, from our experience, we had Zendesk community for a long time, and that limited what we were able to do.

When we moved to Gainsight, and it's not just because I'm at a Gainsight conference, we're actually paying Gainsight more than we're paying Zendesk for just our community, because it's that good. But at a high level, when we moved from Zendesk to Gainsight, we saw a dramatic increase. You saw the ticket volumes.

It was down and to the right over time. On the opposite end, on the community side, we have seen, and I gave you roughly 120,000 page views per month to our content. Support ticket volume-wise, today we get about 2,500 tickets a week.

So you see there's this massive difference in the number of people getting support via the community versus one-to-one support. But the change doesn't happen overnight. But that is, I would say, community more so than anything else.

AI especially has been the single most important thing that we have done that allowed us to not grow our team at the same rate that our revenue is growing.

[Speaker 3]
Thank you. Thanks, Kevin.

[Speaker 1]
Thanks, everyone.

[Speaker 3]
Thanks, Azura.