Atlas in Action

Thanks for coming. Hopefully you can hear me OK. Yeah, again, I'm Francisco Lozano. I lead Go to Market.

I've been here for two years. I came through the staircase acquisition a couple of years ago. So Justin and I are lead all the customer-facing engagements for our Atlas agents, which you saw Chuck announce this morning, AI Native Service-- AI Native Services is what we're calling it. But the brand itself is known as Atlas.

So I'll get into a little bit about what we'll cover, right? So why people need agents right now, but more importantly, where people get stuck, where they struggle, right? And then we'll talk about what agents we have today from ones that are already out of the box, ones that are a little bit of set up, and then your Black Diamond agents known as Atlas. So how many-- there's not a lot of people in here, but I'm just curious-- how many have started to experiment with either a gentrification workflows or actually deployment of agents in their roles, or at least in their orgs?

Fairly good number. Good number. Yeah. Yeah, it's new, right?

We were up here a year ago announcing Atlas. And it took us a while to kind of get it configured. I think we even assumed it was going to be much, much easier than it was. But it took our engineers quite a lot of time because of the complexity and everything that it goes through.

But right now, the way that the speed of the market it's moving, which Chuck talked about this morning, is you absolutely have to really kind of take that leap of faith. At the very least, start experimenting, right? Whether that means just starting to say, hey, can I-- how do I use Claude, or how do I use ChatGPT, just to make my day easier from things like meeting prep, helping me with the QBR deck, writing and responding to emails. Very simple, kind of low-hanging flute.

And then from there, you can kind of start to embrace it a bit more and more. But this is now the baseline, right? This isn't like, oh, maybe I'll get to this now. It's really important that you guys lean in and start to do it a little bit more.

Because the expectation is the way that things are moving so fast, your customers are going to expect you guys to move equally as fast, right? People know that I can build a deck in Claude in two minutes. Literally, in two minutes, I can build a QBR deck, right? So the expectation now is that you respond as fast as the technology is moving.

And we're all being conditioned in our personal lives, right? I think we all recognize that. We're all here thinking about our customers. But we're also customers ourselves in our own right.

So it's like, raise your hand if you have curated Spotify playlists, right, where AI is giving you a little help with that. So these are the experiences that are in the market. We're all so used to it. And so now your customers have that same demand of you.

And I think that feeds nicely into the later part of our talk track. Yeah. And if you're not doing it, your customers are. It's moving incredibly, incredibly fast.

We're having to keep up with it as quickly as possible. We started at Gainsight with some Claude licenses for people. I think a couple of years ago, we've actually now deployed-- not Claude, excuse me, chat GPT. We've actually now deployed Claude almost throughout the entire organization.

Everybody at Gainsight is really mandated, whether you're in support, whether you're in sales, whether you're SDRs, whether you're executives, throughout the org, to really embrace it. Because we know the competition is nipping in our heels constantly. The two things that matter the most are, in terms of, well, how do I know this is working? And how do I know it's effective?

Is productivity. Same adage of can I do more with less? So how can I maximize the amount of time I'm spending? And trying to look at, if you think of your day-to-day activities, where are things that you spend a lot of time doing?

It's things like logging meeting notes. It's building decks. It's preparing for meetings. So that's where those things about productivity and efficiency are really the keys of success internally, as well as where you're going to win with your customers internally by deploying this.

Lastly, as Chuck talked about this morning, we're now being measured on outcomes. It's no longer how well am I using the software. How am I doing my day-to-day tasks? No, it's how well am I helping my customers retain their own business?

How well am I making them grow their business? How profitable can I make them by being more efficient in my duties? And so because of that, you're not having this proliferation of buying and buying and buying and buying. You have to be able to work smarter with the tools that you have.

And maybe some of you have in Claude or Chat GPT or the LLM of your choice done this before, where you've given the LLM an outcome in mind, and you ask it to work backwards and to fill in the gaps. So that's the mindset that we all need to have. The LLMs are really good at that. So if you haven't tried that, even if it's something simple like you're trying to book a family vacation, I'm going to give it the end point in mind and say, this is where I want to start.

Help me fill in everything in between. Think about that in terms of a workflow for our world here in customer success. And so now it's incumbent upon us as humans to be in the loop, of course, and to work closely with all of you to really understand your workflows and personalize that experience and bring it to life for your customers as well. Yeah, that's a great point.

As you're testing these, one thing that we use in our AI workshops that we start off with is pretend you're going to write a press release announcing how successful you've just been by deploying AI. What would the headline for that press release be? Or the headline you would want to see across the Wall Street Journal. That's the prompt that Justin's talking about.

What do I want to announce in terms of success? And then the actual LLMs will be able to build backwards from there. So this is the fun part. Where do we get stuck?

You've got to think of an agent almost as a new hire, as a new entry-level job, or as a really, really smart intern. The same way you would onboard a new team member, you have to give them clear lines of ownership. What's your roles? What are you doing versus what am I doing as the human?

They have to have clean data. Part of our onboarding exercise in a lot of the scoping is making sure we understand what those agents are pulling from, whether that's Snowflake, whether that's Salesforce, whether that's Gainsight, contact management. You do not want an agent reaching out if you have old stale contacts, multiple contacts, Gmail, Yahoo addresses, if anybody still uses that. But the point is, you have to make sure your data ready before you do any agent employment.

And before you all panic about the term clean data, just keep in mind, it is relative. Every meeting I'm in, I think that's a pain point that all of us experience. So it's OK. We're all in this together.

And again, it is relative. But some sense of data quality is part of these types of projects, because we do want to give the agents some sense of clarity of who the target set is, and making sure the contact name is where it needs to be. But we can work with you on that. And that comes back to Gainsight bringing 15-plus years of industry experience to work with you in this manner.

And we kind of go on with the conversation about why these workflows and these agentic projects are challenging. That's one of the reasons why. So we'll talk more about how you can lean on us for that moving forward. Yeah.

And then the last point in terms of where agents can struggle is you have to have a combination of probabilistic and deterministic. They have to have rules. You can't just turn them on and expect them to understand where to go, what the rules. So you don't want to be super, super constrictive, but they have to have guardrails around them when you're deploying them.

Humans, on the other side, also get stuck. So you've got to understand, where do I pick up the baton versus what the agent is doing? I need to know every step of my workflow so I know what is the agent going to do versus what I'm going to do, where those handoffs are in terms of the process. Again, think of if you have an employee, your agent is your direct report.

What is she or he going to do versus what am I going to do? That's how you have to think about it. And then lastly, in terms of bandwidth, you guys have to make sure you can support or there's got to be people that can provide the humans in the loop, can't say it's offering those services, but they have to actually understand where those guardrails are. So a game site is a little insight into where they operate today.

So we already have today out of the box agents, zero setup, ready to go. That's our expansion agent, which Chuck talked about this morning, tells you where the opportunities lie for growth or NRR. Based on all the conversational data that's coming from staircase, it'll tell you what to focus on, what the dollar opportunity is, if it's ready or not, if it's heating, et cetera. Risk agent, which everybody loves, right?

Again, mining all that contextual data, building a context library to understand what's going on, and then knowing what's worked and what's not in the past based on all the history on the account. It'll actually tell you not just where the risk lies and why, but what actually to do about it. So it comes with full recommendations with it. And the last one, which we actually just announced today, is a handoff agent.

So when businesses close and an opportunity is closed, there's been a lot of history across that account. From the first point of engagement with sales or an SDR all the way through all the conversations to when that Salesforce gong rings, in the past, the AE would have to fill out a big form, hand it off to professional services or onboarding or to customer success now that it's all done for you and it's all automated. So no more having to hand write and spend a bunch of time in onboarding meetings. Did I get the outcomes that the customer wanted right or not?

Did I miss anything? Nothing's missed. It's all done agentically now. Quick question.

Show of hands. So for the handoff part of your customer's journey, raise your hand if you feel like there's been friction in that part of your customer journey within your organization. Yeah, that's great. And that's our experience too.

We've noticed just in the last week or two since we introduced handoff analysts, it's quickly become my favorite agent that I talk about with a lot of our customers. It sounds like a no-brainer, but just simply having a great understanding of why the customer purchased whatever they purchased from you. It's amazing at understanding that, understanding the outcomes, the personas involved, who was the economic buyer, those types of things. Agents are really good at collecting that information and handing it to you in a really consumable format.

And that's what these agents do across these moments of your customer's journey. OK, so level two. Now we're getting in from the green runs to the blues, if you will. Little bit harder to deploy.

Take a little bit of the setup. But with a little bit of intermediate level work, these can get set up. So this is the moderation agents, which we talked about today. That's mostly for community rather than having a human going in and making sure nobody's going off the rails with their posts or whatnot.

The agent actually do it for it. And then your AI tutor, which is going to be found in our learning systems as well. Lastly is the actual agents that act autonomously. So gain sites deploying to adoption agent, which will look at a variety of signals and actually be able to take action, reach out to the customer, hey, we notice you haven't logged in for a while.

Hey, we notice you're not using this feature or whatnot. Think of it as sort of an advanced, very, very advanced journey orchestrator, if any of you have used those programs. So that'll be coming online in about a couple of weeks or so. And then what we announced this morning is our renewal agent, which Justin will actually show you live.

We have four customers live right now. There'll be a session tomorrow afternoon with our actual product managers talking about what they've learned during those deployments. So part of the renewal agent was what we announced this morning, right? Atlas AI native services.

What I talked about earlier is we found, just like we found a gain site, many customers, while they're super enthusiastic and they want to get going, they don't have either the bandwidth, the expertise, or the humans to actually be able to effectively run these renewal outcomes. Takes a lot of not just the configuration of setup, but you have to be able to manage things. There's full-time jobs. So what we want to do is be able to own that from end to end, not just the technology part of it, which we'll provide, but the full managed service.

So you let us focus on the actual outcomes of the renewal motion itself. And the goal is actually, obviously, we're not just provide better performance, but better cost effectiveness and obviously GRR lift. And the reality is, everyone, if we built this project with you, built the workflow out agentically, got the project started together, and then handed off the agent and everything that comes with it to you, you might have some questions about what comes next, right? The next chapter, the next phase.

And so we're eliminating that ambiguity and that lack of clarity for you. We're taking that responsibility and ownership on ourselves. And of course, as an end result of that, we own the outcomes also. So it's really a bespoke, almost turnkey solution for you in terms of helping you manage that all-important long tail in your renewal motion.

Now the fun part. OK. What it actually looks like. Absolutely.

So let's jump into the demo. OK. So before that loads up here, I'll give you just a quick housekeeping safe harbor statement. This is a bit of kind of a pilot that we're showing our customers.

The whole goal of this demo is to continually educate you all on what's happening under the hood. That's the best way to think about it. As Francisco said, agents are like really, really smart interns or that brand new smart employee that you just interviewed and hired, and it's day one for the agent. And so on day one, the agent needs to understand context of your business and needs to understand your products, how you speak and interact with the relationships you have with your end customer.

And it needs access to tools as well. So if you all remember back to your first day on your current job, you probably said, OK, how do I get access to my email? How do I get into Salesforce? How do I use Gainsight?

So the agent's asking the same questions because it needs to understand how it can get to the data it needs to go do that job and ultimately drive that outcome that your business and your leadership team will certainly expect from us here at Gainsight. So let's talk a little bit about what that looks like. I'm going to jump into our renewal agent demo here for everyone. And the first thing I'll do is sort of set the stage on what that day one onboarding looks like.

So as I mentioned, a lot of it comes down to context for the agent. Everything from providing it with the correct identity, understanding what kind of tone it should use, professional in nature, in this case, understanding things like contract terms, and really bringing it to life. So I bet some of you in this room have experimented with Claudier, GPT, and done some of this yourselves, like you're trying to get help doing your taxes in April or something. And you say to the interface, hey, you're a super smart professional tax attorney.

Help me out, right? It's the same idea. You want to give it the feel and the flavor for what your best colleagues would do when they interact with your customers. From a data connection standpoint, you could see some examples here.

In a renewal workflow, baseline would be Gainsight CS. Gainsight CS is a requirement of our agentic projects in this space, because the agent needs an outstanding data layer to work from. And many of you in this room probably use Gainsight CS every day. And it's a great place for the agent to start to get acquainted with who are your customers, who are the targets, what are their health scores, what are their renewal dates, and all those really important data elements that kind of bring the agent to life.

And then you'll see a few other elements on this screen. Staircase. You've heard a bit about staircase starting this morning, probably in some other sessions that maybe you've attended. And maybe you've even seen some of my teammates at the booth and done some demos.

I hope you have. If you haven't, please do that tomorrow. They're great folks and can get you up to speed quickly. Staircase is, while it's not a required component of an agentic workflow, it's pretty close.

And I say that because staircase introduces a whole layer of conversational intelligence to the agent. So imagine, now you give the agent insight to what your customers are actually saying to you through email, through recorded meeting transcripts, through support cases. Those are incredibly rich channels of unstructured data where staircase can work its magic and bubble up risk insights, expansion opportunities. We already talked a little bit about that handoff moment a few minutes ago.

So all that insight can really help further inform the agent in terms of what it's trying to achieve. A few other areas here. Now, this particular part of the onboarding of the agent is just to kind of illustrate the point. It's very precise.

These are data fields. You don't need to look at each one of these and memorize this list by any means. But this is just to show you that the agent is comprehensive. It's going to target whatever makes sense for your particular workflow your respective organization.

And so with that said, we always work backwards from the outcome in mind. What are we trying to achieve? We're trying to renew this customer, hopefully autonomously, right? And if we work backwards, what does that workflow look like?

And what do you need to give the agent to bring that to life and to make that possible? A couple of remaining items here that you might find intriguing. I could talk about this area as the backdrop that the agent needs to operate. Think support documentations, examples of contracts, examples of email outreach, anything that might be useful, quite honestly.

That would be appropriate for a renewal conversation between an agent and the end customer. And then last but not least, we've talked a little bit about maybe some moments when some of you have experimented with CLOD or GPT other tools, right? Well, in those environments, it's all about building skills, right? We always talk about CLOD because we use it at Gainsight.

Two of us probably use it a lot and are personalized as well these days. So within CLOD, building a skill is an important concept. Because if you think about the agent on day one, when you bring it to life, the agent doesn't know how to go create a call to action in Gainsight. It doesn't know how to post something to Timeline.

And there's a lot of other things it doesn't know how to do, including handling maybe a docuSign in the renewal process, or updating Salesforce, or generating a quote through CPQ. So these are all the things that are in our tool belt and tool kit for bringing the workflow to life for all of you. So hopefully that gives you some context. That's kind of where we start with the agent.

And then we bring things to life further through the playbook itself. So up here at the top, you'll notice I have one active playbook. In this case, it's called auto-renew, which seems to be appropriate for what we're trying to achieve. And for those of you who are familiar with Gainsight CS, you know the concept of playbooks.

Playbook is a series of tasks. It's very contextual. It usually shows up at the right moment connected to maybe risk or an expansion opportunity or something going on with your customer. So that's the right context to think about for this.

But this is quite a bit more comprehensive, as you'll see in a second. So if we go into our configuration screen, this allows us to tell the agent right out of the gate, what's our success criteria? What are we trying to achieve? In this case, well, we want to sign renewal, right?

Hopefully autonomously. We'll talk about the edge cases and where that's not always possible in a few minutes. We also want to update the opportunity in Salesforce. So for those of you who use, of course, Salesforce, probably many of you do, or other CRMs and other tools, those are also part of the plan, right?

We want to bring that into the project to make sure that we're basically handling all the steps as autonomously as possible. So teams don't have to fill in the missing parts. The agent can cover all of those gaps. And then down here, you'll see there is an area where we can decide how that agent is triggered.

We can trigger it based on a data signal. So perhaps a renewal date's coming up, maybe t minus 180, t minus 90, whatever the case may be. And then that will trigger an autonomous workflow on its own. You can also trigger it any way you want, manually or through some other mechanism.

And then we start to get into things like the outreach strategy. Now again, forgive the in the weeds view here. But the idea here is very comprehensive, as you can see. The agent's going to learn exactly what it needs to go do and in what steps and what sequence.

Now the first item on this list is one, Francisco hears me all the time in our customer meetings pointing this out, where it says AE coordination. I think that this is really intriguing. In fact, I was mentioning it to one of your peers in a meeting earlier this morning. And they said, yeah, this is actually really cool to see this come to life.

So what's this step all about? This is internal coordination. Imagine for one of you here in the audience, the agent reaches out to one of your teammates first, internally, before it reaches out to your customer. Simply ask the question, is this a good time to reach out to the customer?

Is this customer in a high risk situation? Or maybe if it's somebody on the sales team, are they currently engaged with another opportunity? Still might not be the right time to reach out for the renewal. Maybe next week's better, maybe in a couple of weeks.

So the agent can give you that flexibility. And of course, that's just one example. You have ultimate control over the workflow and whether or not that's part of what's included. Now this third area here, inbound strategy.

You're probably curious what happens when the customer, your customer engages with this agent. And you're probably thinking, wow, there's lots of scenarios and you would be right. So sometimes the customer is super friendly and says, yeah, terms look great, let's renew, let's move forward for another 12 months or whatever the case may be. Other times there might be a little negotiation.

I'm sure none of you have had to happen before. So that comes up. And then other times maybe the customer has a question about the product. Maybe there's something that's stumped them and they need a little bit of assistance.

So this screen is just meant to illustrate, we can account for these scenarios. And by the way, this is one of those areas that when we talk about these workflows and these projects being a bit challenging over time, this is an area that really matters. Because for our early days customers here at Gainsight, we're spending a lot of time with them watching to see what customers respond with. Because we want to account for those scenarios.

And we want the agent to be really personalized and thoughtful with its response. We don't want the agent just simply sending a canned email back to the customer. If the personalized approach can be used, if it could even include documentation or do a little negotiation on your behalf where it makes sense, well, let's do that. That makes things even more autonomous.

And then lastly, I'll show you guardrails. Probably an area that most of you will find intriguing. This is an area where we need to make sure that that brand new teammate, that super smart intern that's working for you 24-7 that we've been talking about during this whole session, also knows what it shouldn't do. So if you have a customer who has a P1, P2 support ticket, going back to the staircase example, staircase gives you visibility into sentiment.

A leading indicator, you see it right away through emails, recorded meetings, support tickets. So with staircase, you'll know right away if this is not a good time to reach out to the customer. That could be a hard constraint that you could use. And then in terms of things like negotiations and whatnot across pricing, you'll see here-- and it's probably hard on this screen-- we've given the agent, in this case, a parameter that's anything that's equal to or less than 8% discount as a request by the customer.

We're going to let that slide. We're going to give it the green light, and we're going to renew that customer. But if it goes beyond that, we're going to hand off to a person. And that's where all of us still really matter and move the needle in this scenario.

We need to be involved at the right time, that right moment. So the last part of the demo is, let's bring it to life a little bit. I'm going to show you an end-to-end workflow, walk you through what that looks like, and also share some of the finer points, which will relate back to what I just talked about-- things like hard constraints and the responses and the personalization and so on. So let's take a look at our agent cockpit.

What's interesting here, too, is this also illustrates, if I just switch over to all for a second-- notice there's a toggle at the top of the screen where I can switch gears between different agent workflows. I call this out a lot in the meetings that we have with all of you and your peers. Because as time goes by, I think it becomes even more and more critical to understand how the agents interact with each other. You've all raised hands about tinkering with agents, building things on your own.

No doubt your company's IT teams are all experimenting. So we're going to be in a world very quickly, if we're not already there, where you have multiple agents doing many different things. In our space, adoption, renewal, perhaps onboarding, there's some things we're doing in that space. So how do these agents stay out of each other's way?

And how do they also effectively coordinate with one or another? Again, another area of the project which can be a little bit challenging, which, of course, we're happy to take on as well. So let's talk about this renewal. So we're going to take a quick look here.

I'm going to scroll to the top of the screen. And then you'll notice there, in bold and green, it says that this renewal was closed autonomously. No human involvement, which in this case was our goal. If you've got a sharp eye-- I know it's really small on the screen-- at the very top of the screen, you'll see that the estimated ARR for this particular renewal was $210,000 US.

If you look down in green there, you'll see that it was closed at $197.04. So immediately we know there must have been some negotiation going on here. So we talked about that. We talked about the idea of having a hard constraint being built into the workflow.

So let's see what happened here. So if we scroll down just a little bit, there is that first step that I called out a little while ago. The agent coordinated with an account executive first before it reached out to the customer. In this case, again, I know it's small, so I'll read it for you.

It says, hi, Jordan, who is the account executive. I'm about to send the auto-renew notice for this customer. Here's the renewal date, the ARR that's expected. I'm flagging it before I contact this customer in case you have any notes or anything that you're thinking of where it might not make sense to reach out.

So 18 hours later, Jordan sends a note back to the agent. Looks good. Go ahead. This particular customer is a solid account.

They even mentioned a QBR that I had with them last week that they want to renew, but there might be some discounting that they're interested in. So the notice goes out. The agent reaches out autonomously to the customer, David, and reaches out and mentions the name of the product. It even can bring in other examples of things that the agents learn from reaching out to other customers.

The name of the game here, as we've talked about, is personalization, right? This is not a canned email. These are very, very effective. They get by spam filters very well.

That's actually one of the things that's been most intriguing about some of our projects. Nothing's getting filtered because everything is so personalized in the outreach. And of course, the open rates go up as well because you're starting to see a lot of dynamic interactions between the customer and the agent. Now, if I go down a little bit further, we'll see David respond.

Hi, thanks for reaching out. We're happy to renew. One ask, is there any flex on pricing? We've been a customer for three years, something in the order of 6%.

So if you recall a few minutes ago, 8% was our cutoff, right? So 6% is OK. The agent's looking at that going, two thumbs up. I'm good to go.

I can do this without any of your help. So 6% got the green light. The quote was revised. And then it can tap into DocuSign, CPQ, handle the update and the opportunity in Salesforce, do all the things that can be a little bit cumbersome and time-consuming sometimes for teams.

And you can see that this was marked close 1 for 197.4. So again, the goal here is exactly what you saw. We know that there's lots of edge cases. We know that there's scenarios where there's a good reason to pop out of this sequence and hand it off to one of us.

Those are always going to hold true. I think you can all probably think of scenarios where that might be relevant for your businesses. But hopefully, this brings to life how we think about building this out from the ground up and continuing to iterate on that together. Because again, even with a managed service offering like we're proposing here at Gainsight, you're still very much in the loop.

Your team has visibility through Gainsight CS, through things like CTAs and timeline to see what the agent has sent out to your customers. So you always have visibility. You have line of sight into that interaction. And you're very much a part of that process.

I hope that resonates and was exciting to see a little bit under the hood. One last item I'll mention for all of you. If you're curious to learn a little bit more, I would say in the next two to three weeks, give or take, I'm working with a teammate to bring to life the next version of this demonstration, which will have even more bells and whistles, including a live Salesforce environment, sending the email in real time, seeing some of the mechanics happening back and forth. If that's intriguing, we'd love to get together one day on Zoom or Teams or solution of your choice, and have a great conversation around that.

Do we have a question? Yeah. Let's do it. Yeah.

I'll take that first one. Great question. And thank you all, number one, for being here. It's a big deal to make the trip to Vegas and to be here with us.

So thank you. Number two, we're throwing a lot at you. And we know that. So thank you for being here and absorbing it all and learning and asking questions like this.

Atlas is-- you can think of it like the Innovation Hub here at Gainesite that's very focused on fully autonomous workflows like the one that I just showed you. The other agents, like Expansion Agent or the Risk Analyst or Handoff Analyst, those three in particular that I just listed off, they're part of our core staircase product that's available today. They're available out of the box. You literally turn on staircase.

You plug in your email support ticket and record a meeting data. And then within a matter of, I would say, 12 to 24 hours, you have examples of an expansion analyst report as well as risk and handoff. So those agents, again, they're built into kind of a productized offering that's very much available to you today. And again, to distinguish Atlas, I would consider that more of a project-based engagement that is very bespoke.

It's very much about your particular workflow for your organization, for your long tail, for the most part. That's the target here. And we would build that out with you. And our stated goal is within Atlas and within those projects is to get to know your workflow as well as you know it and to be able to operate within that in mind.

So I kind of touched on the next one of, are these agents, including within Gainesite CS, and if not, are there any agents we can utilize within the core Gainesite? Take a turn. Yeah. They are not-- well, Staircase is now integrated with Gainesite CS.

So when you turn on Staircase, you'll have those agents that are in the Gainesite CS platform. And then some of the stuff that you saw today. But no, as of right now, they're not integrated fully or turned on yet within the classic Gainesite platform. Is the renewal agent connector only out of the box or can custom connectors be created?

It's 100% customized. It's completely configured to your needs. Renewal agent will never be-- it can't be out of the box because it has to actually be built to how you manage your renewal workflows. So whether you want it syncing to Gainesite, whether you want it pulling from Snowflake or another data lake, whether you want to pull it from Salesforce or another piece of software that you're using, that's why it takes configuration and setup is usually 60 to 90 days.

InfoBlocks and PartSource got out of the gate a bit quicker because they had the really clean data. But most of the time, it's about 60 to 90 days. But no, it's totally customized, each one of them. We have a forward deployed engineered motion.

So they actually work with you hand in hand to do all the setup, launching, and monitoring. Do the agents reach out to customers as AI agents? And what's the feedback from customers about that interaction? I'll take that one.

So I love that question because it's one that we're all kind of wrestling with. So some competing schools of thought to be totally transparent. We have one customer who preferred to send autonomous emails using the agent and not disclose, at least initially, that it was an agent. And honestly, the emails are so good you can't tell the difference.

However, if their customer writes back and asks, it will disclose that it's an AI agent. So there are different ways of approaching this. And just as an FYI for the room, we have done some experimenting. For those of you that were here last year, you probably remember that we showed off a phone call in an autonomous scenario as well.

And phone calls are a modality that we do support as well. And so then you do get into the same question, right? Do you disclose? Typically, in our experience so far with some of the project work we've done using phone calls, there is an immediate disclosure, if not immediate, pretty soon into the conversation.

There's that disclosure. So there's flexibility. And as Francisco said, these projects are very much bespoke. So we're looking forward to working with all of you in terms of what makes sense for your organizations.

To use the adoption agent or renewal agent, do we have to buy the Atlas services? Or can we use them as part of our Gainsight implementation on our own? Yeah, I can take that. So renewal agent is separate.

There's a baseline fee package that covers all the setup and think of it as professional services. And then on the back end, we'll take a small percentage based on performance, right? So based on successful renewals, think of it as almost as a commission or sort of a BPO motion. Adoption agent, once live, will be available in Gainsight CS as a self-serve motion.

That's a little bit further out. But that will be available almost sort of as an add-on to your Gainsight contract. Do the agents talk to each other today, the renewal and adoption agent? Yeah, they very much can.

So it's a great question. Again, in the spirit of what I mentioned earlier, getting them to coordinate and not step on each other's toes is important. So we can build out a scenario. We don't have one active yet, but I know we will in the very near future where you have, let's say, a renewal motion and adoption motion active at once.

And then if there is sort of conflict over email outreach for whatever the use case is, then yes. They can keep an eye on each other for lack of a better way to put it and make sure that if one's firing, the other one doesn't, if the time is not perfect for that to happen that way. What are some of the data that needs to be cleaned up before deploying Atlas agents? Yeah, I mean, you have to have basic things, right, like renewal dates, actual ARR, contract data.

As silly as that sounds, maybe stuff that just didn't get updated the last time somebody renewed. But contacts probably are the most critical. You don't want it reaching out to the wrong person with the wrong information. So we'll help kind of go through a lot of that scoping for you during the even pre-sales engagement.

But those are the basic things, is kind of your contract data and contacts. And also, just so everyone knows, too, a lot of these projects, especially-- we're talking about renewals a lot today, right, in this session-- we start small, small scale on purpose. And that also surfaces additional data quality challenges that maybe we all didn't notice during the initial scoping. We do catch a lot, though, in the initial scoping.

We sit down, have an honest conversation with your teams. We get your IT team involved and make sure that we're all well-coordinated with the workflow. But that is another reason why we do typically start with a small scale of your customers that you would initially reach out to. And we're talking about a first week outreach of five, six, eight customers, right?

Let's see how it goes. Let's see if things are firing the way that they need to. And as our business unit leader, Prem, always says, it's sort of like getting water running through the pipes to make sure that everything is smooth and it's where it needs to be. And that'll help us surface other data quality challenges that maybe we can work through together.

Does the handoff agent go from pre-sales discovery through onboarding and then finally to CS? Yeah, it does. It's pretty comprehensive. So the way you can think about it is, Staircase as a solution is the tool that surfaces the handoff analyst.

And Staircase, by nature of how it works, connects directly into teammate email inboxes. And of course, is analyzing transcripts from the meetings that your teammates and yourselves attend. Just to be clear, as an organization, you have full control over who's on that list of teammate inboxes. That's an important caveat that I always raise.

But because of that, if you bring sales team members into that list of synced inboxes, you get visibility into the unstructured data. So you're collecting all those early conversations your team's having with customers, including reasons they bought, expected outcome, any friction-related challenges, what's their tech stack look like, you name it, right? The list goes on. So those are the things that do get carried over into the post-sale motion.

Great for TAMS, PS resources, your CSM to carry with them into the early onboarding phases of the relationship. Can you create custom agents? Are we limited to these three currently? [INAUDIBLE] Yeah.

So the use case we discussed in this session was, again, a project-based set of outcomes that Gainsight owns based on an agentic scenario that we build, of course, with you in coordination with our partnership. We are introducing some tools that you may have seen in the keynote today that are coming very shortly to our Gainsight CS platform that will allow you to build agentic experiences natively within your Gainsight CS product that many of you use and love today. So that's here right now, and more is coming very shortly. But just to be super clear, what I demoed today is not something that would be built independently by one of you.

That would be a project you would engage with us for. Awesome. And let's get to this one. What are the primary data sources Staircase needs to be effective?

Yeah. Yeah. So its primary data sources are email, so either Outlook or Gmail, support cases, and then the recorded conversations. They can also pick up Slack if you're using Slack Connect with customers.

But those are the three. And it's fully out of the box. There was another question about, does it need to connect to anything through Gainsight CS? Staircase works completely outside of-- or independent of Gainsight CS means we'll connect directly to Gmail.

We'll connect directly to things like Salesforce, Sales Cloud-- or, sorry, Service Cloud, Zendesk, et cetera, for ticket information, as well as every piece of conversational software you can possibly think of in terms of bringing in call transcripts. It does not require any connection of a data source from Gainsight CS. And interestingly, just as a quick anecdote for everyone in the room, if you have Gainsight CS and Staircase together, they integrate very nicely from a UI standpoint. So it's one surface for your team.

Not to mention we have the MCP support. So you've got Claude, and you can do your thing, or GPT, and ask questions and build workflows. But Staircase also lives alone as an option as well, which is so powerful because you could, in theory-- we see a lot of our customers do this-- invite, let's say, sales team members who don't necessarily use Gainsight all the time, but they'll use Staircase all the time because they love that conversational intel. And you've seen that in your role as a CSM.

So you're bringing a lot of other personas into this data ecosystem, which honestly makes the whole platform that much more powerful because it gives all of you more to see, and it bubbles up more to your leadership team to drive better outcomes for your business. While I have a captive audience, I just want everyone to be aware. Awesome. Well, we're at the end of time, so thank you, everyone, for joining.

Thank you, Francisco and Justin, for joining us today. Thanks, everybody. And leading. Quick note-- Pulse Party is at 7 PM.

Bring your badge and ID. Otherwise, you will not be allowed in. They are checking IDs. So double check that.

Thank you. Thanks, everyone.