Day 2 Keynote: Retention-as-a-Service Realized

Please welcome to the stage, Gainsight Chief Revenue Officer, Marilee Bear. [Music] Hello, Pulse community. Welcome back to day two. It felt like this expedition really came alive yesterday, and I think it's safe to say this community knows how to explore and celebrate.

Congrats to everybody who clearly made it out of the hidden ruins that our events team created at the venue last night. Now, as I think back to yesterday, what stood out to me were the conversations people were having everywhere. Everywhere I turned, members of this community were asking the same version of the same question. How do we keep growing customer value when the pressure on retention keeps rising?

Yesterday, Chuck introduced retention as a service and the AI-powered capabilities that go along with it. Pre-built agents inside the platform, agentic capabilities that let you build your own vision inside Gainsight products, extending the value of Gainsight out to LLMs and tools that your team is already using. And lastly, Atlas AI native services, where you can hire Gainsight to deliver the renewals outcome on your long-tail customer base. Now, I'm not up here to put you on the platform.

That'll come later. I am here as a CRO who owns the retention number and leads CS, just like many of you. And I can tell you that the pressure to retain and to deliver and adopt AI, it's very real. I know you all feel it too.

And the data backs it up. Bain has been writing about GRR pressure across SaaS. Headcount is shrinking, procurement is pushing back, and expansion requires clearer ROI proof now more than ever. Mackenzie has also shown how deeply retention connects to company value.

A few months ago, it released a study of 100 B2B SaaS companies. Those with best-in-class retention trade at 24 times company revenue. Those with lower NRR trade at only five times revenue. The difference is retention, and that's exactly what's at stake.

Inside Gainsight, we are responding like many of you. We've started using those three agentic paths ourselves. We're using what's already available in the platform. We're building and extending where it makes sense and we need something more specific.

And we're evaluating how AI native services can help reshape our approach to renewals. So today, you are going to see what that AI strategy with Gainsight looks like in real workflows. You'll hear from post-sales leaders at Gainsight who are closest to this work every day. They're going to show you how it's taken shape within the Gainsight platform as real workflows, real capabilities, and real outcomes that help your teams respond to the pressure and put retention back in reach.

So with that, I'm going to hand it over now to Gainsight's Chief Customer Officer, Brent Kremke, so you can see what's possible in practice. [Music] Good morning. As Merily said, yesterday we talked a lot about the challenges that we're all facing as customer success leaders in the room. And I peeked to a lot of you every single day and I know we're all getting the same questions from our board, our colleagues, our C-suite.

It's questions like, what are we doing with the insights we're getting? Are we getting enough insights? Of course, AI and agents. We're getting these questions over all these things we're talking about actually having an impact on GRR, NRR, and efficiency.

The other thing that we're all facing is the answer to these questions not very long ago are now very irrelevant and are creating more and more challenges. We know that this map that we talked about yesterday is more critical than ever, and we've got to get the answers not just for our customers, but for ourselves and the leaders that we work with across our organizations. Yesterday you heard us talk about how we're going to bring together both building and buying into a single path. And so today I'm incredibly excited to walk you through exactly what this looks like.

So there's three main things I'm going to talk about. One, Merily mentioned, pre-build agents all out of the box. Second is how are we building on top of Gainsight Intelligence with tools that you're already using today like Clod, Salesforce, other LLMs of choice. And then third, how are you building now directly inside of Gainsight with Gainsight Agent Studio?

Today, again, we're going to get hands on with all three of these. So where I want to start is within your customer success teams today what they can use directly out of the box. So for this demo, as we do, we're going to use Sally, our CSM. She uses Gainsight every day.

For all of you in the room that have been using it every day, don't worry. The UI isn't changing. The critical components that you've built out are still going to be there. But what's no longer on that is what it's going to take to make it work.

Instead, it's working for you because it's now powered by pre-build agents that are already on, already running, and already surfacing exactly what matters for all of you. This is what Sally, our CSM, now sees when she logs into Gainsight's view directly in her account page. AI-generated sentiment, stakeholder relationship strength, and AI-generated milestones. All out of the box, integrated AI signals pulled directly from staircase AI.

It's giving Sally both the relationship and account level insights she used to have to hunt for. Now, let's go one level deeper. And one of the most common pain points I know we all continue to hear about, which is that dreaded handoff from the sales team into customer success. The reality is, Sally used to spend days trying to go through sales tools, notes, begging, pleading, to really put together what exactly did the customer buy, what are the outcomes they're looking to drive, who the stakeholders are, and where the landmines are that she may need to anticipate.

Now, all these insights, the goals, the outcomes that the customer is trying to drive, they're already there. Now, Sally can have a meaningful conversation on day one instead of month two. This is just the beginning of what pre-built agents unlock for CSMs like Sally. AI agents like that handoff analyst, the risk analyst, and the expansion agents are already live today for all Gainsight customers, all running the moment you and your team logs in.

Once these goals are on the table, the success plan will now start to build itself around them. AI-generated success plans pull from the account insights and similar success plans that are tailored to each one of Sally's customers. They're actually built to work. So now, Sally has the CSM.

All she has to own is the delivery. All of this is the first way that Gainsight is now delivering on the map that you heard about yesterday. So now, let's talk about when Sally's teammates need to really work behind the platform, because here's the reality. Sally lives in Gainsight, but there's a lot of people that she works with like Jake, her account executive.

He doesn't. He lives and works in Salesforce, and we all know how much we have to beg them to check certain tabs, go into certain areas. And the reality is we've stopped asking because we truly believe at Gainsight we should work wherever you and your team do. So now, what you're looking at is Salesforce.

Accounts, Gainsight's embedded directly within that Salesforce account page. Now, a lot of you are probably thinking, "Okay, Brent, this isn't new." A lot of you already have everything set up directly like this in your own platform. Jake can already see the same health scores, same expansion signals that Gainsight creates, all directly within Salesforce. What's new, though, is now what Jake can do with this.

With Agentforce MCP Connector directly connected to Gainsight, he can act on all of his deeply integrated customer data exactly where he lives, which is in Salesforce. So an expansion opportunity that an expansion analyst at a Gainsight has built or identified directly turns this into a Salesforce opportunity almost instantly. Context is already filled out. All the fields that Jake hates populating and probably isn't going to do, it's already done.

Same intelligence your CS team runs on, zero workflow change for the team that you're dependent on across your organization and where they live, which is in Salesforce. Now, it's not just the CRM that can now build on the deep context and the expertise at which Gainsight is delivering. We're bringing that same underlying infrastructure to the AI tools as well. I know we've all heard about MCP over the last few days, but here's why it matters.

Before MCP, your teams were already using Gemini, Claude, chat GPT, but the reality was, one, they were getting generic answers. And those answers were staying within their login, their chat windows, etc. Now, with MCP already released in live for both Gainsight customer success as well as staircase, when Sally goes in and asks about a customer in her LLM, she gets an answer that's pulled directly from Gainsight with real data. Real context, so when Sally is asking a question like, "What top risk accounts should I be aware of across my book?" She gets the right answers where she's already working in Claude.

As a CS leader, this has changed how I operate my day. My team is annoyed because it's all I talk about. But where they're not annoyed is, I no longer have to ask them for prep. Every morning I can go in, I get a debrief, I understand exactly what's going on.

It gives me the context and the confidence every single morning to know exactly what's going on. Now what we're going to do is, you're going to see how Claude can also start to give Sally the answers to the risk and key callouts exactly where she needs to have it. The way we like to think about it is, Claude is bringing this processing power. Gainsight is what is answering the questions underneath.

Now to take this a step further, it doesn't stop at the data. Sally not only is asking questions and getting those answers, now she can take action. Creating calls to action on her highest risk accounts, all from Claude without having to go back into Gainsight. So now to shift from MCP, now I want to go a little bit deeper and you all heard us talk yesterday about CLI.

For any of you that don't know what this is referencing, it's command line interface. And I like to think of this as, for your technical admins or your agents, it's a direct line directly into Gainsight with no additional UI required. So the examples that we just talked about in regards to MCP, they're great for everyone, including CSMs like Sally. CLI is both for agents themselves, for your more technical users on your team, for your admins who really need to build exactly what they need and they need to build it fast.

We know how important speed is right now in this new environment. So here's a real example for a CS admin like Olivia. Most admins like Olivia, they spend their time setting up and managing digital touchpoints and for any of you that are familiar, your journey orchestrator within the platform. If you aren't familiar, it creates custom journeys.

So this could be everything from calls to action, automated emails, in-app engagements and so much more. It's incredibly powerful, but the reality is it does take time to build and we've got complexity, different segmentation models, different engagement models. That complexity can start to build. With CLI, Olivia just describes her situation and the program builds itself in minutes and not days.

Additionally then, the different variations for her segmentation model, her engagement model, it's built just as easy. So you just saw Claude doing its thing to build this out line by line. Claude is doing all the work for you. You get a cut straight to the output, which is the program already built for you directly into the platform.

And you can see that the Google Cloud is the inside of Gainsight. Everything you've seen so far, it works in Claude with chat GPT, with co-pilot, whatever LLM is your choice within your business. Most of you in this room already have access to these incredibly powerful AI tools. What you just saw shows what's possible when you connect those tools to the broader tech stack across your business and you start to build on top of Gainsight's customer context.

But there's one bigger question and that is what happens when you want to build what you want to build isn't just a generic app. We all are familiar again, things like Claude co-work, Claude code, replet, lovable. These are great to build anything. But a CSM or a customer success admin like Olivia, they don't need to build anything.

They need something that already understands success plans, health scores, calls to action. They need something that's built on customer success and not just AI. And so that's the question that Gainsight agent studio powered by Claude is answering. Gainsight agent studio is where your teams can build customer success and gain site workflows.

The ones specific to their account, their data, their day. And because it lives in Gainsight, it already knows everything it needs to the right APIs are already prewired permissions governance security models. It's all already inherited. So here's what this looks like in practice night before busy call CSM Sally ask agent studio to prep her.

It reads everything. It's got her calendar account health historical context from timeline calls to action. It starts to give then Sally a prioritized understanding and overview of exactly what was asked and it even creates a call to action for her based off of that summary. But then it ask would you like me to do this for you automatically every night or every morning one?

Yes. And Sally now is built a skill that can run automatically right there in her conversation. No admin set up no separate UI. It runs every single night now Sally's briefs are waiting for her every morning.

But then the same is true after a call as we know. No call goes as expected. So Sally has a call with Abbott. And she's then not sure where do I start based off of the call that I was just on.

So she tells agent studio it reviews the call surfaces what matters walks her through it one step at a time. Now it starts to mark the onboarding milestone complete flag and flags and adoption risk from the reorg that the customer mentioned to her. It also logs the verified outcome that they short that they shared and then it routes all of the follow up comments directly back to the right place and so now what it's doing is it's asking Sally to approve everything that it created. So now she can choose to make this as well another repeatable skill that runs after every single one of her customer calls.

So agent studio isn't just like taking tasks off of Sally's plate. It can also she can also start to ask it to build entirely new dashboards UI elements that fit her needs. So you can see things such as key insights reports with their daily priorities anything at risk everything that might have slipped through the cracks now for the CSM it is one click away. The way that I like to think about this is we built agent studio on 15 years of gain sites institutional knowledge.

It's every pattern that we've seen across thousands of customer success teams all baked in from the start. But we also know that no platform can anticipate every single workflow every customer situation all the institutional knowledge that's in this room all the different ways that you all operate. So with agent studio we've prewired everything we know. And we're handing the keys to you all to build the rest.

So these are just a few ideas of where you can start or hoping for is that you walk out of pulse this week with a few hundred more. Reality is we know the customer success category is shifting what you saw today is 3 ways of how gain site is truly delivering on the shift right now. It's pre built agents it's ready the moment you walk out of this room it's building on the intelligence of gain site with tools like salesforce or whatever AI tool your your organization and using and finally. It's now a place to build exactly what your team needs right inside a gain site safely secured securely all without leaving your platform.

With all of this your team is now more efficient more predictable healthier in your grr stronger than ever. This is not just adding AI adding AI features this is truly a different architecture and it's here now. The last thing is and what's even better is this is not just about gain site customer success and staircase to show you where these paths are also leading us with skill jar with community and our product experience. I'm incredibly excited to welcome to the stage Samantha Murray.

[MUSIC] Yesterday you heard Chuck and our leaders talk about the choice to be a critic or to be an explorer who charts their own path and goes on to build great things. Now all great explorers have one thing in common they don't set out empty handed they come prepared with a toolkit built for the journey ahead. So let's talk about the toolkit that we've built for digital customer success teams. There are two paths we talked about yesterday building and buying.

Let's talk about building first. Meet Emma she's a digital marketing manager at Abbott. She runs email campaigns for Abbott and email monkey is her tool of choice and this is the email monkey customer hub where Emma goes for everything and anything email monkey. It has Q&A tutorials courses release notes events and so much more.

This hub has one single login one customer profile one seamless navigation and a unified search everything is personalized to Emma. Now this unified experience for customers can seamlessly move between educational resources and community conversations is live for all of our skill jar and CC customers today. Everything you see here comes out of the box from gain site. But what if you need a custom widget that doesn't exist in our widget library.

You used to need a developer and that developer had to use an HTML editor with no version control no AI tools and no collaboration to deliver that custom experience. That's all changed with the launch of developer studios available in customer communities today and coming to skill jar this fall. Developer studio lets anyone vibe code the community of their dreams any widget you can imagine you can build it and ship it to your community. And if your community team does happen to have developer support they can finally build deploy and version directly in.

Sorry and deploy custom widgets directly in CC using GitHub. So now let me introduce you to Diego he's head of digital customer success at email monkey. Diego oversees the digital experience for 10,000 email monkey customers academy community in app experiences newsletters advocacy programs literally all of it. An email monkey is constantly running all sorts of events webinars meetups dinners and different teams across the organization run these events marketing customer success sales and even product which means that Diego gets messages like these every single week.

So after this latest message from Ravi Diego decides he's going to solve this problem he's envisioning a map right inside his community where customers can find events near them. Diego fires up quad he's already loaded it connected it to a GitHub template preloaded from gain site preloaded with everything that Claude needs to know about building CC widgets. He asks for an event map widget Claude builds it and pushes it to the email monkey community back end so let's check out what Claude's built and watch Diego drop it into the email monkey community. Diego finds his custom widget right alongside gain sites out of the box widgets there it is events near you under my widgets.

He publishes it. And just like that his vision is live in the email monkey community from idea to shipped in a single afternoon. From here his team can configure it edit it and drop it anywhere else in the community hub that they may want it. Now we've already seen customers building some really cool widgets from event countdowns advocacy scorecards and even template galleries.

If we know this community well we know that you're going to design things that we haven't even considered yet we can't wait to see what you all built. But as cool as this is Diego's day job isn't building snazzy widgets he has 10,000 customers to take care of. Most days he lives in a sea of browser tabs but these days he's been living inside of Claude. So today we are excited to announce MCP servers for the entire digital customer hub skill jar customer communities and product experience all in open beta.

As you heard earlier model context protocol is a way to connect your favorite AI tools to your gain site products. The CS MCP that Brent shared earlier brings account level intelligence into your AI tools. Our digital customer hub MCPs bring user level data and intelligence into your AI tools so that you can build incredible digital experiences that actually move the needle on customer health. Skill jar knows every lesson and every learner community holds years of conversations and feedback.

PX sees every single click that your customers take inside your product pull all of that together into a conversation with Claude or chat GPT and you've got the entire picture of your customers in one place think of how powerful that is. Now before I show you what this actually looks like you're going to want to bookmark this page using the QR code on the screen. It's your inspiration library where you can get ideas for all of the different ways that you can use our MCPs. Today I'm going to show you a few ways that MCP is changing Diego's day.

Email monkeys chief product officer Alex just messaged Diego. He shares that email monkey is launching a new feature AI email templates in only four days way faster than plant. This probably sounds pretty familiar to you guys right. He asked Diego how they can get every customer adopting this new feature in 60 days or less.

That means building training identifying the right customers and enrolling them normally that's a three week sprint today with Claude and skill jar connected it's just a few prompts. Diego asks Claude to build a course plan and shares some source material that other teams have already created like product docs release notes and a video tutorial. Claude gets to work and drafts an outline for Diego to review including level appropriate interactive lessons quizzes and more. Diego approves it Claude builds the entire course and after reviewing the content Diego asks Claude to publish it to the email monkey Academy.

And there it is the how to use a email templates course live in the email monkey Academy. Now imagine the two weeks have gone by the email templates feature is live the course is published and Diego's team has activated in-app engagements and a whole stack of adoption campaigns. So Diego really wants to know is anyone actually using my new feature are all of my team's efforts working. With Claude and PX connected he can get an expert level answer to his questions in seconds no more digging through dashboards.

He asks Claude how many people have adopted a email templates in the last two weeks. Instantly Claude pulls the entire picture together from PX. 700 accounts have fully adopted the feature 2100 have tried but didn't make it through. He discovers that the main point of friction is setting up a brand voice.

But what really stands out here and is exciting is that 70 percent of accounts that have adopted the new email templates feature had someone complete the email templates course confirmation that education is moving the needle. Diego thinks he has something concrete to bring to the product team now but he really wants the full picture so he turns to the community and to Claude to find out what his customers are actually saying. Diego asks what's the sentiment on a email templates in the community give me top themes ideas product gaps and build me a report. Claude synthesizes the information from real community conversations and provides for key themes a product gap for a flagging and even drafted answers to customers top questions all packaged into a report that his product team can act on in the morning.

Now you might be wondering how Diego became such an expert vibe coder and such a deep MCP and Claude user. He learned all of this in anthropic Academy they have a suite of free courses available rich with interactive exercises videos and hands-on learning which of course runs on skill jar but who's counting. So think about what we've seen Diego do so far he shipped a custom widget he launched an adoption course and he pulled live insights from PX and from the community for his product team. A year ago that would have taken at least a month if not more today it's one person and one conversation with Claude.

But this didn't happen overnight for Diego he's an explorer he's not waiting for permission or the perfect playbook he's picking up new tools forging his own path and reinventing what his role looks like along the way. And that can be all of you too. But not every team wants to build some teams just want agents that work right out of the box in the gainsite tools that they use every day so let's take a look at some of our recently released pre-built agents. For Diego like any community manager moderation is part of the job keeping the community healthy and thriving never stops with our moderation a agent available today.

In gainsite CC Diego can focus on engaging his community members while the agent upholds his code of conduct obvious bad actors are removed automatically anything that needs a judgment call gets routed to his admin console for a quick human review his moderation time gets cut in half. We also recently introduced skill jars AI tutor now in open beta think of this agent as a study buddy for all of your learners inside your Academy. When someone gets stuck mid course AI tutor is right there in the lesson available to answer questions grounded in your course content and your knowledge base. The learner doesn't have to leave the course they can just keep learning and progressing.

Let's take a look at this so Emma is in Diego's a I email templates course and wants some best practices on how to craft welcome emails and the AI tutor is right there for her. So AI tutor helps learners while they're inside a course but what about when your customers get stuck inside your product in the middle of their flow of work. Most of you already have a really great support agent whether that's for thought fin or decagon trained on your knowledge base and handling thousands of conversations a week. And many of you would love to bring that support experience everywhere your customers already are so that's exactly what we've built.

We are excited to announce a forethought integration with CC skill jar and PX now this is currently in a closed beta but if you'd like to join our wait list you can sign up using the QR code on the screen. Your support agents have already been trained on your knowledge your learning and your community content so it only makes sense to bring that experience into your product into your Academy and into your community to. By integrating for thought directly into your digital customer Hub search experience your customers get a richer more conversational way to find answers to their questions and get on blocks in a moment of need. Yesterday you heard Colin from zendesk talk about how they're bringing a I powered conversational support directly into the zendesk community so if you'd like to check out this experience for yourself you can go to the zendesk community or to our own gain site community to see this for thought support agent live in action.

Now let's see what this looks like inside the email monkey up. Emma is inside the email monkey app and she's trying to use the new ai email templates feature. She needs help setting up a brand voice which we learned earlier is a common place that customers tend to get stuck. Previously she would have had to search for answers in the help center outside of her flow of work.

Today she simply opens the in a pub right inside the email monkey product and types her question my company doesn't have brand guidelines how do I set up a brand voice. For thought powers the digital customer Hub search experience and answers right inside the product providing her multiple options to work through her problem. The best part about this is that Diego didn't have to build any of it he simply flipped on the fourth on integration and it's now working everywhere his customers are. So while Diego is a builder at heart the pre-built agents from gain site really allow him to hit the ground running.

Here at gain site were of course using these agents ourselves and seeing some pretty incredible results. The moderation ai agent has been really helpful in reducing spam and our for thought support agent has already seen an 83% ticket deflection rate. Every one of you has a Diego on your team maybe it's you someone who's been asked to drive adoption for a late breaking feature launch someone answering the same question over and over across 10,000 customers. Someone who needs answers not another complicated dashboard they've been figuring it out on their own stitching things together for years doing more with less.

These product innovations that were delivering allow you to focus on what you do best engaging your customers and building unforgettable experiences that drive retention and growth. Now go build something great. Please welcome to the stage merrily bear and Josh Schechter. All right it is now time for one of my favorite moments of pulse the annual game changer awards every great adventure has its own trail blazers those that chart the path forward and inspire everyone to follow.

And this morning we're recognizing eight customers who have done just that with gain site. First up CS transformation of the year a year ago pay scale CS team was reactive they were chasing renewals they're putting out fires but they decided that wasn't good enough so they rebuilt the entire platform and function from the ground up resulting in a predictive revenue focused growth engine. The outcome of five point increase in retention in just one year please welcome to the stage Michael midday VP of customer success and renewals and Lindsay oh VP of customer strategy at pay scale. Next up we have our retention champion CCC intelligent solutions CCC asked a simple question what if we stop measuring our CSM on how busy they are and started measuring whether customers are really feeling the value.

So they built it a storytelling framework engagement scoring across the full go to market and a business continuity plan and process that keeps the right conversations happening at the right time that's a team that turned retention into discipline. Please welcome Tim would head of digital customer success at CCC. All right we now have our adoption accelerator award and that award goes to send oh so here's the number that stopped me accounts with a certified send oh so admin spend one hundred and twenty seven percent more than those without. So send oh so when all in they built seven strategic use cases within gain site CE certification became the fastest path to value and the accounts that engage with send oh so a ninety eight percent net dollar retention rate.

So please welcome Danielle Evans director of go to market and customer enablement at send oh so. All right the digital architect award goes to cloud era cloud era didn't just use gain site they built on top of it they launched team agent and AI assistant purpose built for their CSM six months after the launch CSM covered accounts grew 30 percent and in our increase so one hundred and thirteen percent. The window of worldwide intensive industry and the (upbeat music) The Intelligent Innovator Award goes to PartSource. This team built something most CS orgs are still just dreaming about.

An AI powered intelligence system inside GainSight that unifies spend, operational results, and feature adoption into a single health view. It detects risk automatically, it diagnoses the root cause, and then recommends the next move. The first year on GainSight, their retention hit 99.6%. Please welcome Alana Weinbaum, Head of Client Experience Operations at PartSource.

(audience applauding) I'm A star A neighbor now The Rising Star Award goes to Brad Sova at DailyPay. In year one with GainSight, they did a full roll out of forecasting, NPS, staircase across CS, and they rebuilt their health scoring model. By the end of the year, churn improved seven percentage points and they hit 100% of their upsell target. Congratulations to Brad, Senior Director of Customer Success Ops at DailyPay.

I'm gonna fly now I'm gonna fly now I'm gonna fly now And finally we have the Summit. The Hyperscaler Award goes to Anthropic, the AI safety and research company behind Quad. With Skilljar, they built certification level onboarding at massive scale. They went from 30,000 monthly active learners last year to over 1 million today.

That's a 30x increase. Please join me in congratulating Zoe Ludwig and the Anthropic Education Team. Zoe can't be with us today, but we will proudly accept this offer on her behalf. (audience applauding) So a huge congratulations to our 2026 game changers.

We thank you for everything that you do for this community. One more big round of applause please. (audience applauding) Okay, and the recognition continues. The game changers we just celebrated represent the boldest customer transformations of the year, but they're part of something bigger.

Today we recognize our platinum gain SARS. These customers have earned the program's highest honor and a few of them are actually going double and triple platinum. Please welcome Erin White, Dane Johnson, Danny Pancrends, Keith Mattis, Bradley Basha, and Heather Hanson. (audience applauding) (upbeat music) All year long, all year long, our gain SARS customer recognition program shines a light on the customers who continually share their expertise and act as champions of gain sight.

Whether they're sharing best practices, lending their voice on webinars and reference calls, or guiding peers across our community, these stars have set the standard for what customer advocacy looks like at its absolute best. If you wanna join the adventure, just scan the QR code on the screen to join the gain SARS program, and we'd love to celebrate you. Next, please give one more round of applause to our 2026 platinum gain SARS. (audience applauding) I've got that lightning inside me Son of a gun Good job guys.

I'm like a titan that's rising Oh just you watch I'm stepping into fate There is no time to waste I've got that lightning inside me This is how legends are made "Unchurned" is the podcast where we go deep with the builders, the operators, the executives, who are reimagining retention in the age of AI. Today we dropped our 187th episode featuring Fred Reicheld, creator of the NPS. New episodes drop every Wednesday, and I would personally love for you all to follow the show on wherever you listen to your podcasts. And if you want the deeper stuff, the stories that don't fit in a podcast, we have a sub-stack.

You can subscribe right over there. Now on to today's guest. He spent a decade at Google and built search, maps, YouTube. He literally built the infrastructure that billions of people use every day.

Then he co-founded Rubrik, a data privacy company and security company that went public in 2014, excuse me, 2024. And while Rubrik was blitzscaling, he looked around at his own enterprise and saw a problem. So a quick show of hands here. Who's had that moment of work where you know the answer to a customer's question, you just don't know which tool to find that answer in?

Anybody. So our guest built Glean to fix that. And now only five years later, Glean has surpassed 200 million in ARR and Gainsight is proud to call them a product partner. Please help me in welcoming to the stage, founder and CEO of Glean, Arvind Jain.

(upbeat music) I'm falling free I keep moving down Arvind, thank you so much for being here. We're so excited to have you on stage and have you on "Unchurned." I wanna kick things off here. I wanna rewind the clock all the way back. I wanna know, Evi, you're a technologist to the absolute highest degree, you're a distinguished engineer, former distinguished engineer of Google.

Before that, before everything, tell us about when Arvind fell in love with computers. Well, first of all, it's so nice to be here. Thank you so much for inviting me. And my first interaction with computers actually happened a long time back.

Quite embarrassing, actually 40 years. I think it was 1985, 86. It was my elder brother who actually got a PC at home and he was doing some work on it, like I had no business doing work at that time. But we had these video games and he had five of those at the time.

One of them was Pac-Man. And that was my first introduction to computer. Felt it was a video gaming thing and I fell in love. And since then, of course, I've been playing with computers in my work, in my day-to-day work.

Who was better at Pac-Man? You or your brother? I used to beat him, yes. Good.

Yes. Good. For me, it was Commodore 64. Do you remember the old Commodore computer?

That's right, yeah. And it was the Rampage game. Were you serious of applause there, right? You got the bass that are climbing and somehow they're consuming, they're eating the bricks on the buildings.

I don't know how that worked. I think our PC was 286. What's that? PC 286.

Oh, okay, yeah. Do you have a favorite video game now? Are you still a gamer? No, I don't play games now.

You don't have quite the time on your hands these days. Yeah, unfortunately. Okay, so we flash forward from the 80s. Now we're in 2003.

And you remember where you were in 2003 where you just adjoined. This smallish company called Google. At the time, Google was a $20 billion approximately market cap company. It's still substantial.

Google's now 4.5 trillion. So that's a mere 225X of what they were when you joined. So you're still in the early days of Google. And like I'd mentioned, I mean, you built the infrastructure for so much of this.

You were a distinguished engineer. You rose the ranks there. And so I guess, you know, then you left Google, you had accomplished so much there. You started Rubrik in 2014.

Rubrik went public in 2024. You guys grew at a tremendous clip. And as you were what I would call, or what Reed Hoffman would call blitzscaling at Rubrik, you noticed something. You noticed a challenge.

Tell us about that. Yeah, so this was 2018. We were about five years old as a company. And we had really good success.

The company grew rapidly. We were one of the fastest companies to grow to $100 million, $500 million in revenue. And the company grew to 2,000 people from about 600 people the year before. So we're like, you know, like 3X growth in one year, something like that, you know.

And when that happened, suddenly it felt like everything stopped for us. We couldn't produce code anymore. We couldn't ship our product on time. We used to ship at the time, every three months, a new product release would go out.

And now we're sitting with the release that was 15 months and we still could not launch, we could not stabilize. Similarly, our sales teams were struggling. Like, you know, they used to sell a good amount before and now they're not able to sell as much. And so that was actually like a big sort of scary moment for us at the company.

And we tried to learn like, you know, what was coming in the way of productivity and why we couldn't be as efficient as we used to before. Some of it, you know, I discounted that like, you know, looking as companies grow, like, you know, it just, things become slow. But then we did a survey and asked people, what could the company do better? And the number one complaint that we heard from everybody was I cannot find anything in this company.

I don't know where to go and look for information. I also don't know who to go and ask for help and I need help. So this was a big enablement issue. Number one problem, it actually, we scored words on that question, compared to even compensation question, which is typically where you score the lowest always in pulse surveys.

And it was not a surprise because we were built as a modern company in the SaaS era. We had 300 different SaaS applications and our information and data was everywhere. So, but me as a search engineer, that's how like I've spent most of my career in. When that problem came to me, I said, well, like, you know, let's just go buy a search product and we'll put it in front of all of these systems so that it'll become easy for people to find, not just information, but also like, you know, find experts on different topics who can help them.

And we realized that there's nothing to buy. There's no product. And that's sort of what led to the creation of Glean. We wanted to build a Google for you and your work life.

So here's a problem. There's no solution out there. Did you know that there was a willingness to pay that you had a market for this at the time? Oh, I think everybody told me there's no market for it.

So, in fact, like, you know, it's interesting. When I decided to go fundraise, there were investors who wanted to invest in me because I was already, I've done one successful startup and that's the pattern matching that VCs do. So like I was the right person to go and invest in again. But when they would hear the idea that I was working on, like, you know, their shoulders would droop, you know, their excitement would go away because Enterprise Search was the most boring, the most unsuccessful domain.

Like, you know, there's no company that really succeeded in that space. So we got a lot of advice not to work on it. But I felt like, you know, this is a problem that everybody who I know faces every day. And so it doesn't matter to me what the time is, what all the fancy finance terms are.

I know that, like, you know, when everybody has a problem and you solve it, it's got to be valuable. Don't listen to the pundits and the critics and is the moral of that story. And riches and the niches and the most unglamorous stuff is where, you know, the fortunes can be made. Okay, so you've been at Glean, you know, you've been building Glean for a little over five years now.

How does Glean solve this problem that was the initial tipping point for you? Yeah, it's actually seven years. So we are now more than seven years old. We started in early 2019, and that makes us the world's first enterprise generative AI company.

Not that we knew, actually the term generative identity didn't exist, but transformers and language models were there at the time, and in fact, this technology was initially built to make search better. Origins of transformers, this AI revolution that we are in, all of this happened at Google. In like, you know, during 2012, 2013, you know, that time period, and the goal was to build these language models that could help understand world's information better, more conceptually, so that we could go beyond keyword search and make Google smarter. So that's when this technology was built there.

So when we said, you know, we're gonna build a Google for your work life, it made all the sense for us to actually use transformers, Glean language models on our customers' corpus, you know, build dedicated models for them that could help understand your information, your data, in a much more sort of advanced manner, and then build a really good search product. So we were actually able to launch semantic search, vector search, before even Google did, because while Google invented a lot of these things, you know, it was very hard to actually bring those technologies in a product that's already being used by billions of people. You can't experiment that easily. But for us, when we had zero users, we got to actually play with it, and it turned out to be a really good decision.

So yeah, so we started seven years back. The company since evolved. You know, what started as a search, a really good search product now, is a horizontal AI platform that really powers all AI, all agents inside your enterprise. Horizontal AI platform.

So you've always been one step ahead. And we're very proud to say that Gainsight integrates into Glean. You can pass your CS data into Glean, and Glean can operate its data intelligence and its agentic flows on top of your data structures that come in from our CS platform. So I did need to throw that out there, because it's something that we're really excited about, because we want to let our customers meet them where they are.

As I know, Glean is, you know, that's your philosophy as well, which we'll talk about in a moment. Okay, so at Gainsight, we have this concept of Gong. Before I joined Gainsight, I came from the call intelligence world, and so when I heard Gong, I was like, oh, gong.io, right? No, no, Gong is Gainsight on Gainsight.

It's how do we drink our own champagne, dog food, whatever expression you want to use. So what I'd like to know is what is the Gong, the Glean on Glean, at Glean, specific, you know, it could be across any function, but I think this group would love to know on the post-sale side. Yeah. Yeah, so we use Glean extensively in our post-sales organization.

So first, like, you know, just so that you understand what Glean is. So you bring Glean inside your company, you connect it with all of your enterprise systems. So you are connecting it with your emails, with your Slack, where conversations happen. You connect it with, like you said, Glean, with Gainsight, with Salesforce, in fact, with, you know, Gong, if you're using call recordings.

You're basically connected with all the enterprise systems. And now we actually go and understand, like, you know, how your business works. We understand your people, we understand your customers, your partners. We actually see what work is happening.

And some of the things, you know, that now we're able to do is, for any given customer, we're able to actually create a very detailed 360 dashboard. And of course, like, you know, it's not being created just with Glean, like, you know, the actual work is happening in all individual systems, like Gainsight, or Salesforce, and all of those. But Glean actually now brings all of that data in one place, and allows us to actually do a few different things. So one, our post-sales team actually use, like, you know, whenever they're gonna have a meeting with a customer, they will actually use Glean to actually get full context on that account.

And in fact, get, like, you know, what are the key action items, you know, that they should be undertaking, you know, for this call that's gonna be happening now. Whenever customers have questions and queries, you know, we typically route them, you know, to Glean directly, or if our, you know, customer success team is actually trying to handle people's questions, again, like, you know, the default place to go and get answers to questions, get next suggested steps for any ticket that people file, all of that happens in Glean. So again, it's a little philosophical, but I heard here that you have an expression of the company internally, you say, "Do it in Glean first," is that correct? Is that again?

Do it in Glean first? Yeah. There's that internal expression. And I think we're all trying to figure out, like, where do we do our work?

What's your take on that? Like, where should employees be doing their work right now? Yeah, that's a really good question. I think employees should do work where they like to do work.

And I don't think, you know, they have to, I think that when you think about AI, the best AI is the one that actually comes to you in your workflows and not force you to go to some other system, you know, where you can actually use AI. Because I think if it is that, then most people will actually not find, you know, that AI and actually not use it as often. So, like, even with Glean, like, you know, we follow this strategy called Glean Everywhere, where we basically bring our capabilities into all the surfaces where people are already spending time. So if you're, if you're even spending time in Salesforce or Slack or an Outlook or Gmail, like, you know, it doesn't matter.

Like, you know, we want to make sure that we can bring our product capabilities right in those product surfaces. So, and so that's important. Like, I think, I also feel like a lot of AI, ultimately, like today, like, you know, when you think about AI, a lot of AI has been experienced in standalone AI tools. Like, you know, for example, you go to chat GPT or you go to Cloud or you go to, in fact, Glean.

And these are all like standalone AI tools. You know, they've worked like a personal coworker, your companion, and you can do a lot of work, you know, within those interfaces. And that'll continue, but I think the majority of AI in the long run is going to actually shift back to the actual systems, you know, where you run your workflows. Yeah, yeah, and along those lines, you've been pretty outspoken that the AI race, the enterprise AI race, I should say, it won't be one with the best model.

It'll be one with the best integrated workflows, and specifically also with the best context. So, I'd love to understand, I mean, first of all, we're all learning here together. All of us have some sense of imposter syndrome when it comes to understanding AI. Can you explain what context is?

And then, yeah, why is it so important, specifically when you think about these enterprise tools? Yeah, so, when you think about these AI models, you know, they are incredibly powerful. They have amazing reasoning capabilities. They can think like humans, but they still don't know anything about your business.

Internally, they don't know who are, who are all the different people in your company, who are experts on different topics, what products do you build, what are the bugs that are being solved today, like none of that context is there in the model. So, model is a piece of technology that you've got to use, and in fact, as an enterprise, you should make sure that you're using, you have access to all the different models that are being built in the industry today. Today's industry is dominated by closed-source models like GPT, Cloud, Gemini, but in the next 18 months, you're gonna see a lot of open-source models too. So, like as an enterprise, you need to make sure that you have access to all of those models.

But now, to have those models do something useful for you in your business, you know, you need to sort of bring that understanding of your business, your context, to these models so that you can actually solve meaningful tasks, you know, like that actually matter within your company. So, context as such is simply, number one, it's assembling all the data, knowledge, information that you have in your company in one place so that you know what information is there and what work happens in what systems. Number two, you've got to actually go a little bit deeper and start to sort of like assemble and organize information in a little bit, in a better way. For example, take a project.

You know, for a given project, there is a lot of information. Some of that could be in Jira where you're tracking, like all the different tasks that you're working on. There may be documents in Google Drive where you talk about how you design the system. There are conversations that are happening in Slack.

And there's a lot of information on this project which is scattered across many different systems. So, when you think about context, context is on that project, is basically bringing all of that information and contextualizing it, understanding how all of that information relates together for this logical concept which is this project. So, that's what context is. It's just a little bit more derived data artifacts on top of your raw data that's in multiple different systems.

And why is context important? I mean, as we said, ultimately, AI gives us this promise that it can do some of the work that our teams today, our humans are doing. Like, you know, we think some of that work can be automated with AI. And the only way it can do that, the only way it can do a good job with it is if it actually is as knowledgeable as the experts in your company are.

And the experts in your company, like what makes them better, it's the context. It's the understanding of how your business works. It's that knowledge of how to actually get a particular task done. Those are things that actually makes people in your company so much more valuable.

So, building a context graph is essentially taking all of that human intelligence, all of that knowledge, aggregating it, and then making it available to AI so that AI can start at the same place where your experts are today. Is there any cool piece of context that maybe is a sleeper context that people wouldn't think about, but some cool data that you've pulled in, that you've seen in use cases? It's like, wow, this really helps to elevate the intelligence that you can do. Yeah, like, so for example, so at Glean, like, you know, one thing that we will do is when we get deployed in your environment, we'll start to understand all the different sort of concepts and subjects and subject matters that are actually important in your company.

We'll understand your projects, your teams, your lingo, like, you know, how you like to speak in the company, and then we assemble, sort of, write information around each one of those, each one of those sort of concepts. So, there are a lot of interesting applications, you know, that people do, like, you know, which has surprised me. One of the first agents, for example, that people build with Glean were performance review agents. Nobody wanted to write self-assessments or manager assessments.

And, you know, one thing that we do in Glean is, you know, we build a people activity graph, which means that for every individual, we actually keep track of a chronological, chronologically we keep track of all the work that they've done for the last six months, last one year. And so that actually turned out to be really useful context for people to now go and build these HR, like, self-assessment and manager assessment agents. Yeah. So, that's a very hot use case now.

That's a pretty, yeah, I could see that being a hot use case. Okay, so let's flash forward now to 18 months from today. What does the future of AI in the enterprise look like? And I think there's lots of folks here and everywhere that are wondering what happens to the humans in the room.

Love your take on that. What happens to the jobs? Yeah, well first, the AI today, if you think about it, is very reactive. There's no AI, you know, which actually comes to you.

Like, you know, if you have some tasks that you want to do with AI, you have to go into chat GPT or Gemini or Cloud or Gleen, and then you will go and ask question, you will converse. But the initiation always happens, you know, from a human today. And unfortunately, when that's the case, we notice that very few people actually use AI very actively. I mean, everybody's using AI now.

Like, I would say like 80% of employees are using AI in some limited shape or form, like mostly to ask questions, get answers. But there's maybe like five to 10% of the people who are truly reinvented, like, you know, how they work. They've fundamentally changed, like, you know, how their day-to-day work looks like. And that's because, like, it's hard for us to actually find time to figure out what AI can be for us.

And so that's one of the things, you know, which I think is gonna change in the next 18 months, is you'll start to see that AI is gonna become a lot more proactive, and it's gonna start to come to you. Imagine, you know, AI that actually knows you and your work life, you know, at the deepest level. You know, it knows about your preferences, how you like to read, how you like to write. It knows, you know, all the OKRs that you have to actually deliver in this quarter, the weekly tasks that need to get done, or all the meetings that are gonna have today.

So with all of that deep understanding of, like, what work you need to get done, if AI can come to you and say that, hey, here's five things that I can do for you today, before the person actually goes and asks for them. This is one thing that you're gonna see, you know, big shift that's gonna happen in the next 18 months. And that is what is gonna actually lift everybody and make all of them equal AI experts. Today, you know, we're seeing a power law, but when AI becomes proactive, then everybody will get to learn, like, what to do with AI.

So that's one big trend. I also think that in the next 18 months, open source will become a big factor. And a lot of, like, AI workloads in the enterprise are gonna be done by open source models, which will be an order of magnitude faster and cheaper. This is very important because right now, AI as a technology is very, very expensive.

And every company is struggling with how much money they're burning for it, like in order of budgets you chose, which is established for AI, you're actually burning through them in less than one month. And so that's a big thing that's gonna happen. Its cost is gonna come down. And then finally, what happens to people?

And so I feel like a lot of people talk about that AI can replace almost all the jobs, but I don't think it's anywhere close. I don't see a lot of jobs getting displaced. Of course, I think all of us will probably be doing a significant part of our work with AI. It doesn't mean that there are fewer of us needed because remember there's competition, and they also have AI just like we have.

And they're gonna actually use AI to build more features, more capabilities, and so we also have to actually deliver more. So unfortunately, like all AI does in some sense is, pushes us to actually do more with the same amount of time, but of course with a better tool that is in front of us. And it shifts the focus a little bit as well. Yeah, of where things are.

Okay, last question for you. There's another sub-stack blogger that I follow. His name's Scott Barker, former VC. And he published a blog article around this AI age of acceleration, where there's a lot of overwhelm.

People are so excited about AI, but there's also this feeling of, they're daunted and they're overwhelmed. Do you ever feel that way yourself? I know you guys have done things like move to monthly planning from quarterly planning. You're picking up your pace.

Do you either as an executive of an AI company ever feel that way around? Or just as a person, as an individual like the rest of us? And if so, what do you do to ground yourself? Well, I think the number one thing that I hear from customers, I do maybe three or four meetings with customers every day.

And the one thing that I routinely hear from every single person is that, we are very behind with AI. And everybody feels that. Like, you see in media people talking great things, companies saying that we completely reinvented ourselves with AI. There's a lot of pressure that gets put on each one of us.

And I think that's the number one thing that I will remind everybody here that nobody's behind. And I think AI is a pretty good technology. It's gonna come to you and it's gonna carry you with you anyway. So all you have to do is make time for it.

Just force yourself to make time to actually do some exploration with AI tools. But there is nothing that the trainer's already left and you're fundamentally behind. Now for us internally, we feel the same way. We don't actually feel stability as an example.

Like in this company, we are at scale company now. We are over a thousand employees. And five years back, with our revenue scale, we would have gone public today. Today we don't feel that way because AI actually makes us feel...

There's this feeling of that something's gonna change tomorrow. And whatever you build as your product may become obsolete in an instant. And so that's the feeling that we have, that our engineers also feel, like when they're building stuff, they constantly feel that, "Hey look, "it's gonna take me two weeks to go and build this thing. "And in two weeks, will it even be relevant anymore?" So yes, even in an AI company, we do feel that way.

And I think what you do with it is, ultimately, I guess you just have to get used to it. You just have to acknowledge that we are beneficiaries of really good technology and good surprises that come our way every week, every month. And we just have to be ready for it. So one thing that for example we do now is, because our engineers are having anxiety, in fact some of them actually even needed that support, it was really stressful for people.

The one thing that we've done is we tell people that, "Look, if you build something, that's great. "We reward you for that. "But if you actually also throw something, "you build something maybe a month back "and now you actually retire that "because there's something new, a new way of doing it, "we actually reward it even more." And that's one way for us to actually make sure that we keep innovating. So yeah, it's a feeling that all of us have and I think we just have to get used to it.

Adapt to this new world. We have to move fast. We have to be willing to throw technology that we build. We have to be willing to change ourselves and our habits and whatever way we're doing things.

Yesterday, we got to be ready to start doing it differently tomorrow. I love it. Well, keep growing that rocket ship, Arvind. We really appreciate it.

I have the last request for you here. We're starting the tradition today, you and I. I've got your hat. You've got the unturned hat.

I would love for us if we could bring up the house lights. That'd be awesome. Yeah, that's nice. But I would love to do selfishly a selfie with Arvind and our best friends behind us here.

Thank you so much. Awesome, well thank you so much. Arvind Jain, founder and CEO of Glean. Thank you so much for being with us today.

Thank you. Thank you very much. (audience applauding) Impossible. Nobody can jump this.

(dramatic music) It's a leap of faith. (dramatic music) It's Sarwe versus the clock. He's gonna do it. We are gonna see a legal sub two hour marathon to Sebastian Sarwe.

It's the world record by nearly a minute. (dramatic music) Please welcome back to the stage Gainsite Chief Executive Officer Chuck Gannapathy. (dramatic music) (audience applauding) Over the last two days, we've covered a lot of ground together. Exploring how we can use agentic AI to drive higher retention.

Agentic tools that let you build, that are built for you. And AI native services that will deliver renewals in the long tail. Retention as a service. And through it all, there's been one question that's been sitting beneath the surface.

What happens to the human when AI can do so much more? There are a lot of voices trying to answer that question. Some optimistic, some cautious, and some flat out doom. I understand why, because for many leaders right now, everything feels impossible.

Impossible to protect GRR. Impossible to get more budget or more headcount. Impossible to keep up with the pace of change in AI. And as explorers, we must be willing to commit to live in that tension.

You know, the impossible has always existed right up until the breakthrough. My wife Uma, my wife Uma, just completed a big personal journey by finishing all six marathon majors. Boston, New York, London, Chicago, Berlin. And back in March, she ran Tokyo, one of our favorite cities to visit as a family.

And by the way, to our friends, visiting us at Pulse all the way from Japan, thank you very much. (audience applauding) While Uma's achievement was big news in our family, the bigger news in the running world was something that happened at the London Marathon just a month ago. The top three runners all broke the previous world record of two hours and 35 seconds, all three. That's an incredible feat in and of itself.

But what's even more incredible is the top two runners ran the marathon in under two hours. And experts had thought crossing the two hour barrier was impossible, physically impossible for humans to break. Two of them broke it in the same race. I guess nothing's impossible.

You know, it reminded me of the story of Roger Bannister, who in 1954, also in the UK, broke the four minute mile barrier, which was also considered impossible until it wasn't. You know, the London record breakers, they were wearing this $500 Adidas Adi Zero shoes. By the way, if you have $500 sitting around, don't bother, they're completely sold out. But hearing about those shoes reminded me of a slogan that Adidas had launched back in 2004, and then they brought back in 2021 that I absolutely adore.

Three words originally spoken by the greatest boxer of all time, maybe the greatest athlete of all time, Muhammad Ali. Impossible is nothing. In working to overcome what feels impossible right now to all of us, we mustn't forget what the work actually is. What the work actually is.

Our work, your work, is not just about the workflow. The work is judgment. It's context, it's relationships. And above all else, it's care.

I recently read an essay by Alex Emis, who's a behavioral economist at the University of Chicago. It's called "What Will Be Scares?" The question he raises is stop provoking. In a world where AI makes everything easier, faster, and cheaper to produce, what becomes scares? Scarcity doesn't disappear, it just relocates.

And as AI makes more things easier to produce, the commodity gets cheaper. But the human connection becomes more valuable. The person who knows you. The team that remembers your context.

The experts who make you, make sense of the complexity. The community that helps you feel less alone. These things become more valuable in a world of AI, not less valuable. And that's why I believe human connection will always be at the center of all relationships.

Not displaced by AI, but thriving alongside it. AI doesn't make customer success less human. It gives us a chance to make the human work more visible, more focused, more strategic, and more impactful. Look, the technical work still matters.

But in this next era, what we used to call soft skills may actually be the hardest thing to replicate. The future will not belong to the spectators or the critics. As we adventure through our final day of Pulse, let's commit to being explorers. Let's keep moving towards possibility, even though the path isn't fully clear.

Let's learn, adapt, experiment, and build what's next. To treat every customer like your best customer. To turn people, your people, into super humans. To use agents in ways we could have never imagined.

Impossible is nothing. (audience applauds) (upbeat music)