The CS Ops Playbook: From Data Chaos to Decision Intelligence

Hello everyone. Good afternoon. How are y'all doing? How's the energy after lunch?

Another lunch was delicious. Great. So my name is Revant, or you can call me Reeves. I go by that as well.

And it's been over seven years. I've been in the ecosystem of Gainsight. And my first five years of that was actually at Gainsight. And I've started as Gainsight's own Gainsight admin.

And then I had the opportunity to get into their CS team. I got in. Who wouldn't say no? And then I was one of the founding members.

I started as an associate CSM, and by the time I left Gainsight, I was managing a team of CSMs and or looking to 100 customers. It was a fantastic journey. I was a consultant, helping clients like you who are part of the Gainsight ecosystem, help things out, build your CS, tech stack, and all of it. But I really wanted to do something of my own and be the chef in the kitchen.

I like to cook, so that's my hobby as well. So if you have any nice vegetarian or vegan recipes, send that over. So I got an opportunity to be full-time in CSOps, and I took that over. So I was a responsive for like two and a half years-ish.

And I own the CS tech stack and got in CS, Maddox, staircase, and built all of the cool things. Chef in the kitchen. I learned so many things, and that's how I sort of wanted to curate the CSOps playbook. And right now, I'm a savant, taking care of the business systems, Gainsight, Maddox, and Lachmani going forward.

I'm really excited to be here, share this with you all, and take it ahead. And this is something that's just going to live forever and will be built on top of some of my learnings that I have as I have sort of worked and ventured through the jungle of CS and CSOps, as you all are in right now. So excited. Are we excited to get on this journey?

Awesome. Wonderful. I love the energy. And before I get in, just wanted to know by the show of hands, how many of you all had like no experience at all with CS or Gainsight products here?

Anyone here in the room with no experience or haven't used Gainsight? Oh, cool. This will still apply to you, by the way. So it's CS agnostic or system agnostic, so you can sort of feel free to use these, you know, things that sort of we share going forward.

And how many of you all have been with Gainsight for over like three, four years? Yeah, a lot of pros. Wonderful. So that's great.

So if you have any tips, please feel free to share with the teammates or who are on your desk, you know, who are new to this thing. So with that, let's go to the next slide. Yeah. Every CS team, you know, has data.

A lot of us have a lot of data right now in different forms of systems, different products from different teams, and this just data is mostly not connected or somewhat connected, but doesn't really make sense. And this isn't easy to sort of take it forward, so we need to sort of bring this all, tie this all together to make sense and, you know, take it forward. So being in CSOps is just all of this, because CSOps is the center of all of it. We get data from product marketing support.

You just save it, you have it, and CSOps is asked to use a magic wand to create a beautiful deck two days before a board meeting. So that is not easy, and it... all these things, you know, pitfalls, they're sort of not really a ladder, but like some of you who all watch Game of Thrones, this chaos isn't a ladder. This is a quick stand.

So we'll just sink into this, because if this is not tied the right way, if there's no meaningful playbook that's being written, there's no processes that are being developed, it's just going to slowly sink the system. I have seen hundreds of instances where there wasn't enough care that was being taken care of, the data that's being set up, and the data governance and all of those things, and it just took a downhole from there, and they probably had to restart and reimagine their CS strategy. So in order to just avoid that, like let's just see how to sort of get it all organized, get it tied, and sort of take it forward. So let's use this map...

let's use this time to build that map that'll help you take it forward. And as I said, the chaos is real. And when I say that, it just means when there's data, it's just available in all different formats. Sometimes you're not just aware, you need to work with a lot of cross-functional teams to sort of understand what this data is all about.

So do you even trust your data? So you need to work cross-functionally with your teams to sort of understand how much reliable the data is, sort of plug it, play around and see what comes out of it, and then are your systems actually connected? That's the next big thing, CRM being the core centralized application. You need to know how, first of all, how clean that is or how integral that is, and then how connected it is to your other business systems.

Think of it as different ingredients while you're cooking. So you need to quality check each of your ingredients before you start cooking. And then there's a method to the madness. So you need to sort of connect the systems one thing after another, follow a fashion, sort of to make it all connected the same way.

So knowing if the systems are actually connected really matters to help you understand what's the right way to connect the systems and take it forward. And last but not the least, when we get the data, when we all have the systems, what we do is we set up fancy processes on top of that. We start off easy, like a simple risk framework or a health scorecard or like a simple 360 attributes page, right, just to start off with when we start. But how efficient are they?

You know, being in CS ops, I feel it's such a nuanced role. We need to switch the hats multiple times to sort of know, you know, as a TSM, is this, you know, C360 being helpful or not? You know, is it even serving the right details for me to do my work? For a manager or a leader, is the health scorecard actually, you know, showing the right signals, showing the right score to take the actions?

And is the risk framework actually helping us retain the customers, right? As we progress and as we launch new products, as, you know, things move on, these processes need to be evaluated and they need to be, like, updated. And there needs to be some sort of governance that needs to be set up around this to do the quality checks, ensure there's ROI coming out of, you know, these things that, you know, CS ops team deploys. And I feel this, not just CS ops, but it's just the ownership of all the other cross-functional teams that work with CS ops should sort of collaborate to sort of, you know, make this happen to, you know, build better systems and better processes.

So, with that said, let's go to the first slide poll. So, what kind of chaos are you dealing with right now? Is this slide poll live? Cool.

This must be a word. Oh, yeah, obviously. Too many tools, obviously. Okay.

Cool. I see there are, like, everyone's coming up with, like, way too many tools. I mean, that's the reality of today. There are way too many applications trying to serve different use cases, and sometimes there, you know, there's an intersection to all of it.

And inaccurate data, tools having its own data, improper identifiers and all of that stuff. Updating data, there's wipe coding. Oh, wow. Wipe coding is a headache?

That's cool. Yes, unclear health, patience, decentralization, no telemetry. Yep. A lot of it's just around, you know, not having good data or just data that's sort of not married to the right systems or married to, you know, the other data points that sort of actually make sense and sort of tell a story.

So, with that said, like, probably by the show of hands, how many of you are actually from, like, CS ops dealing with this chaos on a data, wonderful, like, everyone, almost? That's great. Because you all understand, you know, how it is sort of being in the thick of this and sort of taking it forward. So, as I said, it's just not the data, you know, like, how it's sort of being used.

Because there are way too many metrics. When you look at product data, they're like, the product team will share you, like, 25, 30 different features and, you know, their usage and stuff. But how much of it is actually helping your customers stay in the product, like your product, and be with you, you know, to continue the journey. So, you can't have 25 different, you know, metrics things on a dashboard, right?

So, it's way too much for CSMs to comprehend and then also for a customer to understand, are they actually on the right track or even sometimes to sort of have a healthy competition to do benchmarking, like, how many things are you going to actually compare to show all of that? So, too many metrics. So, one needs to always simplify the metrics, divide them into specific tenets. To divide the ownership, have a clear vision on who is, you know, fetching this from there and, you know, what is it actually leading us to.

So, and the next is disconnected signals. Obviously, as many of you said, there are way too many applications trying to say way too many things and communicate way too many things to us. We need to sort of centralize all of that and to have a single system of truth or a few system of truths to actually rely on specific signals, like your RevOps tech stack should sort of own your renewal signals. Or sometimes when you have the entire CS tech stack properly built as a one-stop shop, it can fuel your risk signals, your renewal signals and your product intel if you plug that in and various other things.

So, having disconnected signals is a pretty common thing. Actually, that's not really a very bad issue. You actually have something. So, you just need to wet the signal to see if it's right or valid and then sort of tie it and take it forward.

Having some sort of signal is better than having nothing. So, that's sort of a good problem to have and, you know, something we can actually fix and sort of take it forward and, you know, take an advantage off. No action taken is basically hundreds of CTs just lying around in your instance, open success plans that are just basically living, dashboards that are untouched, reports that just lie around, hundreds of rule-querying, millions of records just lying around and, you know, exhausting energy. So, that is all of that.

No actions being taken. That's where I feel like we need to sort of develop some governance on a quarterly or a bi-annual or an annual basis to understand how we can actually make those assets, you know, accountable to provide its own ROI, like a dashboard. You need to visit it every year to see, you know, is that actually being helpful? Conduct, like, a roundtable with your CSMs or, like, the CS leaders to see actually if it's being valuable.

Revisit your risk framework. Go back and see, you know, how many of those actual risk CDAs that open did lead to a churn in X days in advance and stuff. So, we could take a lot of those actions and we have all these issues that we spoke about, right? And in this age of AI, everyone has AI or some sort of AI in your tech stack, right?

Does anyone not have AI here in their tech stack in some shape or form? Yeah, that'll be, yeah, it's like not having internet at this point, you know, so everyone needs AI in some shape or form. And when you run AI on top of these signals, it's just going to show things that are not really actually useful. Sometimes, you know, it will show things actually that are useful, but, you know, bad data in is bad data out.

It is not really scalable. So, we need to fix these foundations of data, the signals and the processes for AI to have a proper context around, you know, like, what it should look at, how it should behave, the kind of output it should give. So, that's what I say, you know, AI doesn't fix broken signals, it just, you know, makes them worse. So, and with that, this is something I've sort of came up with.

Having been at Gainsight, you all must have been familiar with the dear framework Gainsight, what I recommended and a value framework as well. So, this was something that I came up with because I had a bit of an entrepreneurial idea myself and I was thinking like, what exactly should I sort of track, you know, if you set up a business, that's how I come up with consumption, outcomes, relationship and experience. Consumption is basically customers using your products, not just your products, but your services too, right? If they're just using your product, but if they're not certified or if they're not engaging on your community, they will not be your raving fans or your thriving customers.

Or if they're not even engaging in your professional services, they might not be able to implement a lot more things on an accelerated way. So, that's how I sort of summed up consumption as a metric. I also have a framework as we go down the slide, we'll do a crawl walk run approach so that we can take it things slowly. And as I mentioned, right, like for consumption, it starts off with, you know, knowing are they actually logging in?

Where are they logging? Where are they getting to? What are they actually using? And then beyond the product, you know, what other things that they're actually using from us.

Outcomes comes to pretty simple and straightforward, like, are they even getting the ROI from us with the things that they have purchased? Tracking the sales to, you know, the onboarding handoff, knowing if the metric has been converted, having some framework there, having a first implementation check in to see, you know, how things are, you know, being built, at least in the phase one and, you know, having the CS team sort of pick the customer the right way to get them on the right adoption. Put them on our ROI success plan and track all of it. So, outcomes are pretty important.

Relationships, it not only just, you know, comes with, you know, the cadence or the strategic engagement, but it also helps us identify different kinds of multithreading that we are doing in this age of AI and increased products that we are sort of coming up with to sell to our customer to boost our NRR, right? So, it'll really help us understand who are we actually connected with to sort of know and sell things the right way or, you know, move things the right way. So, gauging the relationship is pretty important to save risk, bring more money and just get more outcomes to the customers. And last but not the least, experience sort of covers all of it, right?

From NPS to CSATs to in-app surveys to like knowing how this quick deployment was. So, knowing how the customer feels at different levels is actually important because if an admin loves your tool but your exec doesn't want to buy because they didn't have a good experience, they'll probably churn and vice versa. So, if the exec lives with the admin sort of, you know, that is something I've seen with GainSight too, like pretty much it was a complex tool to build and, you know, as like admins, we sort of sort of confess that through different mediums. So, I'm glad they sort of hear us and, you know, the product in that way.

So, that's where, you know, the experience signal comes in. And with that, now we know what the core signals are or the core framework is. Which tenant do you feel is not, you know, very mature as of now in your org today? So, outcome seems to be the biggest one that's not very mature.

So, you can probably just speak out loud what specifically with respect to outcomes, like, is it lack of having a closed loop ROI success plan or anything in specific that is sort of, you know, being an issue in the system? [ Inaudible question ] Customs success plans, yeah. And that's for all customers, right, across the entire customer base. Okay.

Anyone else? Yes. [ Inaudible question ] Do you have any product telemetry that could probably sort of? Yeah.

Okay. Yeah. Yeah. There are some instances like that.

It sort of doesn't really get easy when the customer would, you know, want to commit to say yes because they would want to get more from you, maybe. So, that happens. I was at that place, you know, when I was at response, I was like, you know, I need more ROI from Gainsight. I was not satisfied.

Yes. But, yeah, we'll get through that. And that's how I sort of developed, oops. Is it back?

Cool. I've developed this crawl walk run methodology so that we can sort of take it slowly because there's only some stuff that we can build and in an iterative way that just gives us the ROI. So, we need to do that. So, let's just go by one tenant to another.

I know there's a lot of context there. So, let's go to outcomes because that was the first one, right? The crawl walk run methodology, as I mentioned over here, there's, you know, you basically have some success plans, no standardized goal framework. Outcomes are not part of Scorecard.

That's usually like, you know, the basic setup, like one sort of has when they start off with Gainsight and try to use success plans or the goals functionality for tracking the outcomes. And then we need to get to the place where, you know, the success plans are actually tracked. There are some smart goals that are being defined. There needs to be some enablement that needs to be done to the CSMs to identify the goals the right way and flag them the right way and then sort of keep them accountable to, you know, update them on a recurring basis.

If that's the case, on a one is to one basis, right? If there's a human touch involved. If it's a digital touch, CS, I would probably recommend having a survey sent out when they sign up to understand what their core goals are. Convert those to your success plan objectives and then eventually sort of track as they progress through the adoption journey.

So, that's how we can sort of get to the run methodology where you sort of set it on an autopilot when they sign up with you. They tell you what they want to achieve that automatically gets locked as success plan outcomes. And then as the product adoption happens and the metrics show up, you can send a survey to sort of understand how they feel about, you know, achieving XYZ metric, you know, or you can sort of notch the CSM letting them know they have hit a green adoption. Like, why don't you just go and check on, you know, if they are satisfied with this particular feature that got released and ROI and things like that.

So, that was outcomes. So, it's the same thing that was on the big metrics that I've created just easy to read. Back to the consumption, right? As I mentioned, like, not just product adoption, but beyond just, you know, your product adoption consuming your services and, you know, other solutions that you offer, you need to sort of be able to track that eventually.

So, that's what I've put in the run methodology. But you need to start off with knowing what your basic, you know, adoption related data points are. I have cool AI workflows and tips eventually down the line as we get to the workshop content to sort of know how you need to identify these as you work with your product teams or data teams to uncover the right metrics. Tie it back with your retention data to validate if that's actually like, you know, a worthwhile metric to look at and weigh on.

And then, you know, plug it into scorecard and then have scorecard run things, you know, automate things like CTAs or a journey and stuff going forward. So, that's consumption. So, you can take a look at these examples to sort of understand where you sort of actually stand and plan the next steps. I know everyone will be in different journey for each of these tenets.

And especially for relationships, right? Like this timeline is just used consistently. There's single threading, no structured relationship tracking. Basically, GainSight acts as like a logging tool to push meeting records and just a data entry tool.

But there's nothing really that happens to know who are you actually in touch with, developing a personas mechanism to know are you actually meeting with the right people in the right cadence? So, you need to get there. That'll be like a walk approach is what I would say having some cadence set up. That is really important because if you're not speaking to the right people at the right times in the right frequency, it just derails the relationships.

And in the world of AI, I think having a human relationship, which is check on if they're doing well in doing, you know, getting the ROI, I think that really sort of takes the CSM's job to a next level and sort of puts them at the right spot, being the actual human and being accountable for the account. The run methodology is all pretty cool. I've used staircase before. That was really wonderful for me because I've defined the personas.

I've defined the cadence at which each personas have to be, you know, be meeting at with the CSM for different tiers. Enterprise differently versus the digital. It can sort of take a look at the delta between, you know, the days since X meeting has happened and it'll trigger a CTA for CSM's to meet with the calendar link and things like that. So that's where you sort of once you have these, you know, good contacts, having personas defined, you know, having a good framework set.

This will help you run on that. And what I've just mentioned is probably like a jet thing is what I would say, which I haven't added here, but you can get to there where it is not really at a point. It doesn't really get a problem and things sort of automate and take its flow. Last but not the least experience.

It all matters. We just do the basic NPS feedback is review. There are disconnected actions. It's just a lot of bunch of data around how customers are feeling a CSM sentiment manual measure and NPS that's done biannually just for a sake.

And then if these are not tying back to tell a story, that's not really a good metric. So you need to sort of elevate from there and then go to like, you know, understanding how the experience of the customer was as they sort of began their journey with us and take it forward at different levels. Of course, as I mentioned, so having that said, right. I know most of you all felt the outcomes tenant was not really up to the mark or needs more improvement.

Where do you feel now that you've sort of understood water crawl walk run methodology could look like? How do you feel where you are in the journey with these tenant? Early call. I'd love to learn from people who are fully running.

Yes. Yes. Totally. Yes.

Yeah. Those are those are wonderful examples of, you know, walk and run methodology where you just set a journey up that does a buy on your LMPs and will just branch your promoters detractors and passives promoters are obviously they get into the advocacy club you pull, you know, that's where the CSF actually is the central thing that will help your marketing and advocacy team to rely on that pool of people who love your product. And then take it forward with the ones who are at risk, you know, you have a risk framework that sort of gets triggered and the passives people sort of sometimes ignore passes, but they are actually more important ones because it's pretty easy to turn them over into promoters. And that's where the CSM should be focusing a lot more on to sort of understand why they're in that mode and immediately have a playbook triggered that's either associated to like, you know, a product risk or like a rental risk or experience a risk or something like that.

So that's a wonderful call of thanks. So yes, so those are wonderful methodology. So seems like there's people a lot in early crawl and then they're walking. Oh, there's no one in for running.

Some of them are walking and running. That's great. Because this is great. Because as we go to the next slide.

Cool. As I mentioned, right, these signals and these methodologies could be different across different tenants and being in the CS ops, you just don't cater to just the CS team, but there are different things, you know, that you cater to the same health scorecard that you're going to get. The same health scorecard that you deploy could help you show how well you're retaining customers. As the stops can take that to show and related to retention methodologies.

CSM can take it to show how well they're maintaining the accounts. A manager can take a look at to sort of show how well are they maintaining their portfolios? Are they investing their energies and sort of driving it forward and for, you know, the CS leadership, it really helps them to showcase, you know, how healthy their customers are and where to invest their energies in the right way. So there are different signals that sort of, you know, power these different users who are in, you know, in the CS ecosystem.

The core users, obviously, but as I mentioned, it goes beyond this. You can send all the product feedback to the product team, the support experience signals to the support leaders, any good content that people, you know, share about, you know, the customers raving about a particular feature can go to your marketing team and advocacy team and take it forward. So CS ops is literally truly the center heart engine of a company. So you can relay different signals based on what you get.

And as I mentioned, right, all these signals and deploying AI on top of this needs to be done with the discipline because bad data is very bad data out. So as I mentioned, these examples over here, right, like if you have a bunch of yellow customers and if you want AI to crawl on top of that to help you identify what exactly are the issues, it will hallucinate and sort of give the examples that would probably eventually be real. So as I mentioned, good signals don't just power AI. They power actions across your systems in organizations.

So with that, how did you feel interacting with AI? Because everyone over here has had AI usage. So. Pre-sensitized sort of being a brain around like taking logical decisions, I feel it has really helped argument anyone' s sort of technical abilities.

So that's actually a wonderful thing because having that in the arsenal of any particular user be it CSM or a CS— that'll actually help them, you know,ill bend AI or use AI in the right ways to get the right outcomes. outcomes. AI, of course, will sort of miss the real story if you have not the full context around it. So it doesn't surface the insights, the insights you don't trust.

So yeah. Thank you so much. With that comes almost the end of my talk track. We'll get into something more dynamic.

I've curated a set of these AI workflows, tips, et cetera. Before we get there, I would probably recommend you to sort of run this core signal self-assessment prompt. There should be handouts on the table. Feel free to scan it, run it, take like five, six minutes.

This will actually help you understand where you are at right now and sort of understand your signal, understand your company, and help AI sort of be guru to sort of help you the right way. So this prompt has things structured in that way. I would recommend to go one tenant by other, as most of your informed outcomes is the issue. So you can probably focus on one tenant, inform the kind of company you are.

The kind of process you're trying to develop, the challenges that you face, and then wrap it up in the prompt and send it over. I had to use it a few times to avoid token maxing in Claude. So yeah, this sort of has helped me sort of understand and position the issue and sort of know where I can actually get to. It will not solve your issue magically, just a reality check.

It will at least help you get there to sort of know where you need to start. It's a very good compass. As you're in this jungle, it will definitely help you navigate and know the pitfalls. As you sort of help AI know what kind of data or you're dealing with.

Yeah, sure. [INAUDIBLE] Good question. So it took me like an hour or so. But because I have sort of dealt with each of these tenants individually when I was working, in order to sort of wrap it all in one major prompt, it just took some time and sort of efforts.

I do sometimes still feel it doesn't really get the entire picture in one single prompt. You just need to feed the AI and inform the solution or the problem that you have and then look at its output and then probably interact with it accordingly. So if it works in the first prompt, that's wonderful. If it's not, then probably you need to tweak it and then probably ask a better question, navigate it forward.

So have you all started plugging the prompt in your LLMs? Cool. Anyone with the outputs? We'll head to share what did AI suggest going forward with any of these tenants.

Cool. I would recommend you take a picture of the thing and you can just try it forward. And with that, let's switch over to the resource page, where you can actually sort of visit the website. Can we switch over to the demo?

Sweet. I made this for you all, so some of the resources that I've helped, you can go to this page, amiraves.netlify.app. Vibe coded. I don't know coding, so yeah.

It was a GSheet that I sort of transformed into this. I hope this helps. So you should land in a page that looks like this. Ta-da.

And then I've curated all these core, tenant, and different prompts or workflows or tips that you can actually use in Gain. Again, as I mentioned, this is CS agnostic. You can use it within Gain site or Staircase, LLMs, or any of your existing tools to just refine your process. So feel free to take a look at it, as most of you all mentioned, so you can feel free to filter down the signals where you have the opportunities to improve, and then where you feel you are right now, and the kind of different type of solutions that you would want.

So let's pick on this one. Identifying what adoption actually means for your product, right? So you can just go here, click. You can just mention you manage customer success for a particular product type.

And this is where you need to be specific with AI to help it understand the specifics of your product or a particular feature to make it simple. And your customer has access to these features, help you define these meaningful adoption versus what looks like a vanity metric, so that it actually shows you things that you can track that would help you correlate it to retention or sticky usage behavior that would drive a positive NPS or something like that. So again, this was an AI prompt, as it's mentioned over there, so you can use your charge APT or a clot for that. And there are some tips that I've baked in.

So pick the single most important feature, so know how to use it, and then clearly define adoption, like what it actually means with one feature, and take it forward from there. So with that said, now that you all looked at the core framework, I would love to open it for the folks to understand which tenets do you feel are more-- let me actually go to the slide. I think there's a slide. Can you switch to the last poll, please?

Yeah. Yeah. What is the first thing that you're looking forward to? Now that you have the resource, what do you feel based on the journey that you are in with the kind of compass that you have, what would you plan to do next?

And then we can pick on questions later after this. [INAUDIBLE] Where are the teams? Cool. So a lot of you would love to understand where you actually sit in the framework, a crawl, walk, run.

Every stage has its own progress. It's actually important to understand where you are at right now, to plan the next steps ahead, and have an iterative strategy for your deployments that you do, and have an iterative approach so you can document it, enable your teams, understand how things are being adopted internally as you deploy these new processes. Reflect on that. Take it back so that you can improve your CS ops.

I was just one person when I started in CS ops, and I just kept deploying new things, one thing after another. And I realized I never really checked the dashboard that I made was actually useful, or were there any things that can be improved on it. So that's where reflect back on the things that you have done. Know how useful they are, document it, and plan and iterate them accordingly.

So use the core framework to align. Yes. So this is really important. I think this has to be like a collective exercise that you need to do with your CS leadership.

Marketing other product teams to sort of understand where are the priorities, and then which tenant you need to sort of focus on to identify what metrics you're trying to bring in, or what signals are you trying to produce to drive different kinds of outcomes. So that's cool. We can probably switch back to the demo. I hope you all have access to this site.

So again, I know we won't get a lot of time, so this is live so you can browse the site, take a look at it. If you have any questions around identifying different signals and core tenets, and how to sort of take it forward, I would love to help and address those. So that's me. [APPLAUSE] Thank you.

Are there questions, anyone? I can come. Yes, please. [INAUDIBLE] Most of our products, we don't get any telemetry on, and that's because they're products that are on premises, and don't send any-- there's no web data to get back from it, or anything like that.

So it's not something we could develop. We don't have a way to get it, and we wouldn't be getting it. That's kind of the design of the product. So we always kind of struggle with that area.

How else would we know if the customer-- I mean, we could have a customer that's using our product all the time, or a customer that's never even-- I mean, we can see that they activated it, but that's about where it stops. Yeah, so that's what I would say, right? When systems or data cannot help you, humans can actually help. So I would probably recommend having CSMs have a structured call with these defined agenda items to sort of take a look at and gauge these things.

And sort of transform the call as evaluate the recall to understand how they're feeling about the product. And then having that plugged with staircase, or your AI engine, or Gong-like tools, right, it'll actually understand the conversation is related to a particular product and a particular feature, and the sentiment associated with it, which will at least help you translate the vibe of the customers associated that particular feature. You can deploy the tokens and stuff, especially in staircase, to look at that particular keyword to associate to a particular product feature, and then tie a sentiment to it. So that's how I've sort of done that as well.

We didn't really have, initially, good product data, so we were relying on CSMs meetings and calls to sort of understand what are they actually talking about with the customers, and how are they feeling about it. And another quick cool tip, actually, on this is, if you're launching a new feature, I would probably recommend baking the keywords of that particular feature into a Gong or staircase, so that when people start talking about it, you actually know if the CSMs are actually talking about the new things that are in the keynote, to know if there's a right adoption happening, and then the sentiment associated with it to sort of track how are your customers receiving that particular product. So this is some way that you can sort of generate product adoption intel, like the sentiment on the product, through this AI system mechanism. So I hope that helps.

Thank you. We are at time. So round of applause again for Ravant. Thank you so much.