When Failure Isn’t an Option: Using Gainsight in Complex, Regulated Environments

Hey everybody. Again, I want to thank you for sticking around this last. We got a little intimate forum, so hopefully we can have some laughs and learn a few things. Before I get into the topic, I want to talk a little bit about where I'm coming from.

So I got lazy and I asked AI just doing a "about me" slide. And kind of just like AI and just like technology, it fetishized this very young with a lot of hair gentleman, which is very different from the reality that you're seeing up on stage. And the analogy actually resonated with me about the promise of technology and the promise of AI. And absolutely it's there.

And what I'm going to talk about is culture and strategy that is greater at the end of the day than technology itself and AI itself walk you through the journey that we went through in Celebrite. So before I get into that, let's talk about what Celebrite is. I asked my buddy, "How do I explain this company to a room of people that is traditional SaaS, they do enterprise?" And he said, "Just tell them what Celebrite means to you." So I'm going to do it that way. Celebrite is essentially a digital investigations company.

Our technologies are used worldwide in law enforcement to help in investigations, child exploitation, narcotics. I just got a call from one of my CSMs in Europe that our technology was used in an underworld assassination that happened in a European country. And that's really what it's about. So when I think about customer success in our market, I think about it as community success.

And I think about the importance and truly the mission critical tasks that our customers are running. And the challenge of that is, obviously, when you're working with three letter agencies and law enforcement agencies, they're not very high on you recording their calls. It's absolutely, they're not very high on being on the cloud, especially when we're talking about very sensitive things like child exploitation videos. I don't want that at all, anywhere near the cloud.

So that's the type of environment that our customers are working in and the type of constraints that our CSMs are working in. And I want to start in the journey of where customer success managers started in Celebrite. Most of our customer success managers are former law enforcement practitioners, investigators, examiners. Very different than the profile.

We see a lot of people in this conference, more gruffy, big tattooed dudes. And if you don't believe me, Travis is up here. He can stand up. He's exhibit A.

Show them your t- your tat. I'm not thinking gruffy, though. And once you know him, he's a teddy bear. They're coming in without the customer success knowledge.

They're coming in with intimate knowledge of our customer's environment and needs. And they're coming in with that mentality of whatever I need to get the job done. We've had customer success managers in Celebrite for probably over six years, but they were jacks of all trades and masters of none. They were doing everything along the value cycle that the customer needed because they've been in their shoes.

And if a customer is calling you and saying, I have a missing persons investigation, he's not going to say, do you make enough revenue? Are you within my patch? Are you in my book of business? I should hand this off to support.

He's going to do whatever he needs to do to get the job done. What that created was kind of a CSM who did everything. He did services if he needed to. He was tier zero support.

He'd reset your password if he were in a pinch. He was your pre-sales guy in a tough moment. He was doing everything. What this created for us as a company was twofold.

One, we had no repeatability. We had to reinvent the wheel every time. A CSM would come in, understand, come with his experience, and then run his own playbooks without even necessarily calling them playbooks. Two, very reactive.

Always in firefighting mode. If you're everything, are you anything truly? Always taking out the last fire, extinguishing what's on fire, but never being proactive, which is where ultimately churn exists and why we want to implement prediction rather than reaction. The biggest challenge to me as a CS leader was how do I validate resourcing to management?

If they're not working towards one North Star metric, if I can't prove that this CSM is creating above 95% GRR, is creating this type of health, it's very difficult for us when you're having those difficult budgeting conversations at the end of the year to get another headcount for an CSM rather than another seller. These are games and priorities we're all making. Very difficult to tell the story up to management. Everybody that was engaging with a CSM like Travis, net positive.

Customers love them. At the end of the day, you have to tell that story in metrics to management to get more resources. Huge challenge. Another thing in terms of the differences and kind of the reality that we operate in Celebrite is the disparity in product telemetry and data.

I am very envious of accounts, SaaS companies that are fully online. You see login, you see feature utilization. It is an absolute dream. Very envious of the conversations that I can have here, how they're mining, especially now with AI, conversations and understand patterns.

The fact of the matter is, that's simply not our environment. And it's exasperated by the fact that we're running three different products. I'm going to lay them out without names. Product one is sometimes online, sometimes offline, sometimes has telemetry, very partial telemetry, sometimes doesn't have it.

Product two and three are what I'll call our growth products. Any software company knows usually you have growth in a really strong flagship part of the product and now we want to drive new business in the new products. One of those products is fully SaaS. We have data.

The other product is fully locked down, sensitive, no ability to see anything. What this created was two things. One, our CSMs were very focused on our on-prem product completely, even though it only does roughly 10 to 15% of our revenue. A, B, all the feedback that was coming from customers was anecdotal at best if it was coming in.

So you have multiple CSMs in addition to solution engineers, in addition to account executives who can be very loud, telling the product what they need to do. We had no one source of reality and whoever was shouting the loudest was defining priorities. Huge, huge challenge. Silos across products that we couldn't break down.

Around last year I was in Pulse. I was so excited. We were getting up and running with the software. I hadn't used anything.

I walked through every single ops session that was. One of them was about term prediction modeling. When we came back, we started integration and deployment with Gainsight and I said, I'm going to set the table spectacularly. I am going to build with a smart guy and our team and our friend Claude a churn prediction model that we fed into the model, all sorts of different levers and ran it back on what created churn.

We came out and I presented this horrible, I mean, we're ops people. The people that are data geeks in this room will like this. I'm telling you to executive leadership, it means nothing. But if we cut out the noise, we knew five levers that historically were very predictive of churn.

I'm not going to delve into it, but you can read them here. The longer the customer tenure, consistent support tickets and feature resolutions, three adoption of latest version. You got to remember on-prem, companies are going into a portal, logging in and pulling down a version versus SaaS. Understanding that they're on the latest version, how frequently they're taking training, whether free or even more if it's paid, and a very specific feature that I can't name, but we understood that this feature was driving a lot of stickiness.

I thought I figured it out. I know the levers, I know what's drawing churn, and this is the part where you know that saying that culture eats strategy for lunch. I realized at that point that strategy eats data for lunch. And us as ops people and data people love to be able to look at insights.

And what I've realized, especially now where we live in an AI world, is insights are a commodity. The ability to instrument those insights into strategy is much more difficult than understanding. I had five levers that I knew intellectually were creating churn, but each one of those levers existed in different parts of the organization. Great, I have this insight.

Customers need to get trained. Our training director, I said, let's give him free training. Well, that's going to hit his bottom line. He doesn't want to do that.

Now we're getting into meetings. I have five levers that in this content house in corporate are in different rooms, and they're not letting me pull the levers. And that's where data is perfect, but reality is messy. This is how I felt coming out of the first meeting.

And then I realized I did a great job in building a model. I did an absolutely shit job in telling that story to executive leadership, because rather than being able to tell them predictably, this is what we understood. And this is what we're doing about it to prevent churn, to increase risk, to mitigate risk, excuse me, to drive retention. I basically kind of went in there and said, I figured it out.

And I realized that day I didn't figure anything out. I took the first step into understanding where the levers are. And then we took a step back and said, I need to radically take this apart. I need to simplify it from what my data person, my experience inclination is to something that's satiable.

And it's not only about executive leadership understanding it. It's about CSMs who I want them out in the field. I want them serving customers. I want them building relationships.

I don't want them being analysts. I don't want to be able to need to explain what a K means model is in a deep for a dark burst, whatever. How do we bridge that? How do we explain what those levers are without having to overcomplicate it?

So I'm going to show you the cookbook that's been working for us. And it's still a work in motion for the last, I'd say six to seven months, stage one, product journey with milestones. Remember, we have three products. They can have different start times and end times.

But every single product has a subscription start date and a renewal date. RCSMs don't run the renewals themselves like a lot of companies do in SaaS. That means that AE has a lot of interest when this subscription starts. But he's also being responsible.

He's being charged with renewing it, which means he's conked for renewing it. So it's really difficult to say the CSM, you say, the AE owns this, the renewal person owns this, you own the experience in between, but you don't own the renewal. You're not doing that work technically. So how do I get first the CSM thinking about lifecycle stages in his experience and in his engagement with that count?

And more importantly, how do I bridge that thinking to the AE that he can understand the health of that count without quite honestly bothering the CSM, that the CSM can relay this information to him without a lot of friction. We simplified it down to three main stages. Once, first activation, time to activation, TTA and TTFV. All customers have to activate, even if they're offline, they have to go to a portal and activate.

That doesn't matter. If you're behind firewalls, you're going to that portal. And that's the first stage that I can say, you took your first step. You saw first value to adoption check-in.

This is for the high touch products. What I showed before of product number two and product number three that are big line items. The CSM needs to be aware that that deployment happened. He needs to be able to check in 90 days after onboarding essentially to give it the green, yellow, red.

Doesn't mean he necessarily has to engage with the customers because he has a big book of business, but there are predictive indicators that he's looking at in Gainsight to say he's doing well, he's in case of customer support, I can see usage where I can see usage and can give it a green and a nod. But if he doesn't know, he's got to go figure it out, right? And it's not, those are the largest accounts and the ones that are adopting our high touch products. And finally risk review 200 days into the subscription is a place to be able to stop and assess what the health of this account is.

If it's green and we feel confident in the renewal, we can move it down the funnel. If we're seeing predictive or anecdotal risk, we put it up in a risk board. We manage our risk by risk reasons. And that becomes the communal or the tribal bonfire that we're looking against to be able to correspond against product, not only in saying customer A is frustrated by this, but customer A is frustrated by this.

And this represents this much ARR and this much ARR is exposed in the next quarter and calculated across the accounts that Travis owns in the west coast of state and local with European accounts, with Asian accounts to be able to say this common thread of risk is higher priority. And now we're talking, starting to talk about risk in a very, very formulated way that it's not who yells the loudest, it's the numbers that are speaking and it's the common patterns across risk. So we need to understand that and I'm going to show you how we visualize it to simplify it when AEs look at gain site. After that, we ask ourselves, what type of digital activity can I run that will shift behavior?

A CSM like Travis is owning well over 50 accounts and he's being dragged all over the place. There's no way that he can touch accounts at the speed and velocity that it'll impact their behavior and the stickiness with our product. So we dug into the first stage of that customer journey, the activation. We have the data.

We know if the customer has activated or not. We set a threshold 45 days in. If they haven't activated, the meter is running, they're already paying, they're not getting value for what they paid for. We'll run a digital play using Journey Orchestrator, very personalized, very hyper focused, a human face rather than a corporate face that is saying to you, we're here to help.

We want you to get the most out of what you paid. This is something very important that I use a lot when we serve the customers that we serve, the person that's signing the purchase order for our technology, he might be signing a purchase order that's twice as much as he makes a year. I want to make sure that he feels that he's not just buying technology, he's buying experience, and he's buying from parents of people that he can trust. What we realized is twofold.

One, we were able to move the needle without having to have in on-site discussions just by creating and making our CSMs digital personas, the ones that knew them and the ones that didn't. And two, even if they did solve this problem, now they have Travis's as CSM name and say, hey, thanks for that. By the way, I couldn't achieve ABC. And now I'm being proactive.

Rather than being reactive to what the AE thinks is the risk, I'm engaging with a larger number of customers and doing it digitally. And that's a net positive for a lot of accounts that don't have a lot of touches from Celebrite. We want to make them and celebrate them and help them along the way. Even if it means I can help you, I'm going to route you to the right place and support.

I'm going to route you to the right internal expert. That's the intentionality of running journeys and measuring their outcomes if they're shifting behavior. We saw 40% increase in activation time among the ones that we did run it out. And we built it out stage by stage.

I don't want to overwhelm the CSMs, but build them out, prove content, evangelize it back to the organization. Three, manual health scoring. Health scores are hit or miss. I think that most of you would agree that there is more half art, half science.

And we have specifically in our on-prem, there is no telemetry. There's no one metric that will say this customer is using frequently. You have to engage with them through a human. But we had no common language and vocabulary to discuss what the risks and adoption was.

And we undertook a project that we called it the day one baseline, where everybody that was engaging with accounts, that includes CSMs and solution engineers that are also embedded with the same accounts, would go to those and provide a four variable red, yellow, green score by four categories. Frequency of adoption, using a daily sporadically monthly scope of adoption. Is it used in all the addressable units? We could have sold a product and it's being used in the narcotics unit.

But there are seven more detective bureaus in that law enforcement agency that aren't using. That could benefit. And the agency has already paid. So it has low scope or high scope.

Performance, that's a measure of quality. Could be that they're using it, but they're frustrated by certain software issues, technical issues, and value perception. Do they feel that they're getting ROI? I would love to have a calculation that spits this out.

But before we could do it and move it, we had to create the baseline. And a lot of people have been saying, like the four C's of CS or the three A's. So I'm thinking about giving this an acronym, like FSPV, funky Sam plays violin. That's catchy, right?

I'll workshop it. But I'm trying to get FSP. The point is I wanted to create language because language creates reality. It creates culture.

And now when we're looking at an account, and it required also of the CSMs to reflect on their book of business, saying, well, where are they actually? Of all the addressable units adopted our product. And now it required an exercise of everybody reflecting back on it. Now, if it came out bread, it goes on to the risk board that we talked about before.

And our risk categorization, I think, is pretty boilerplate, right? In terms of reasons. You have product performance reasons, you have adoption reasons, and we have subcategories under there. But being able to flag risk as the first part of the journey, or mapping account health, doesn't do a lot, right?

I felt during our first months with Gainsight, we could come and say, there's risk to churn in account A, but we weren't held accountable to mitigating the risk. And we weren't holding other people in the organization accountable to help us in pushing and retaining those type of deals. Which is why the next part is really important, is when you do a CTA, and a risk CTA, and we had the categories, is customizing the fields in terms of the status that it's in, customizing the priority, and notice the types of flags that we've added in terms of the custom CTA of who's responsible for it. So in this instance, for example, our tier three in tech support, which is the most technical entity, if a risk falls under the product reason of product performance, and the CSM says this is a medium to high risk, we had to handshake and shine initials with our tier three support leader, and the leader in the VP of support and organization saying, you now have a 72-hour SLA to pick this up.

It's a product performance reason. Even if our CSMs could solve it, that's not optimal usage of that very valuable resource. If he's going to dive in to technically resolve a technical product issue, or he's navigating through R&D in the product house to find who can help him, it's one, suboptimal usage of his time, and two, I'm just reinforcing that everybody's talking to R&D and product and different voices. I want to create one place that is the tier three that is the gateway into R&D that understands the product the best that a deeper layer than the CSM can understand, and now he's responsible for it, and that creates accountability, both from our support leaders and R&D.

They're also the ones that can raise the flag and say, I need to send somebody out there, right? It doesn't happen a lot in the SaaS world. In our world, it happens all the time that you need to send an engineer on site for on-prem products. He had this issue with a very significant client in Spain.

We had to fly an R&D person over there. Tier three are the ones that have the lever to be able to say, this is needed now. The CSM is the one to orchestrate it to say, this is important enough for the business, and his responsibility is to say, this account is strategic and there is x amount of revenue at stake versus something that's lesser priority. He needs to be able to vet that, and the way I think about it is creating, we sit in the customer experience organization.

I'm head of CS, and there's a leader of support, and there's a leader of services, and there's a leader of training. Each one of us is responsible for that account. That accounts are not the CSMs. The CSM is the orchestrator of that, and he's not going to be able to solve everything.

How do you create accountability and collaboration? First of all, within our own home of the guy who knows how to do the job the best, the one who knows the product the best in tier three support, how do I create SLAs for him to solve it? The other product reason, and this is we're kicking off right now in terms of mitigation pathways, is our second product risk category, which is lack of adoption or lack of ROI. Here, the problem isn't technical, it's educational.

They haven't found the right way to use this, and especially in our market, every organization moves slowly in law enforcement. It's even slower. These are huge, very bureaucratic organizations. A lot of time with people have been doing the same job for 30 years, and you're coming in and saying, "We can do this for you." You know what?

He doesn't want to change. It's not the same space of enterprise. There's good to that and bad to that, and a lot of the education has to happen on site with me showing you why, because they're working. They're stressed.

It's not going to happen overnight. How do I map that categorization and send the right person who knows how to do onboarding more professionally or training more professionally than the CSM? Again, I don't want to be the jack of all trades. I want to be the quarterback, but I want him to have very tight processes and accountability in terms of mitigation pathways by risk.

Another thing that we run Gainsight for, and I think this is a little non-traditional, is we track in timeline on every risk once the mitigation is handed off. Our tier three support is saying, "I did this. I did this." The reason that we do that is, again, we want one infrastructure full visibility, and all of that stuff in terms of the CTAs and timeline pouring into our customer data warehouse where obviously we're running it against AI. If anybody asks, "Well, where's the risk at customer A?" I can spit out a report very quickly that says, "This is all the actions that we took." Finally, I want to loop it back to the account executive and educating.

Everything I've been talking about is how we adjust language and culture. The account executive is very important in the relationship, but he's not as in the details around the performance, nor do I want to explain, if I don't want the CSMs having to deal with data at a deep level, I certainly don't want the AE using his time to understand things. But I do want to evangelize what we understood. If we understood that the adoption of feature B drives retention and growth, we can isolate that NRRS larger when accounts are using it, I just want him to be able to understand, you can look at here, he went from 34% to 77%.

Then that gets him having more data-centric types of conversation with his CSM and gradually without a lot of effort understanding. The way we did this is, again, we stream everything into Snowflake as our customer data warehouse and ran the AI multi-source there. But we didn't do it in a generative AI kind of way where it's call and response. We wanted a very structured output to AE where they couldn't get lost in the data.

Because the biggest risk I see from AI is not the technology itself, it's the inundation of it. I don't know what to do with all this information. What we did is we put an interface in Teams, Microsoft Teams is our chat platform. A box, they go in there, put an account name or SFID in, they get a report.

They can't customize it. They can't get an ask it questions, which may seem counterintuitive. But I believe that in order to build cuts correctly, you have to start structured and then expand the funnel. What we're doing right now is a much more Gen AI on our Snowflake platform that the CSMs can ask questions, it can auto prep them for QBRs with analysis of what that account is doing, so on and so forth.

But I don't want us drown the account team very early in the life cycle. Finally is the unification of the timeline. So I walked you through the process of understanding, explaining language, understanding the levers of retention, driving to a G.R.R., driving to what creates N.R.R., creating playbooks. And then we pushed Gainsight out to R.A.E.s.

And the way we did it was through the Salesforce integration on a tab. And this widget, I found it very effective, is you can get lost in the system, right? CSMs should be able to navigate it. It's not correct of us as CSM and ops people to expect account representatives to parse everything that's available.

But I do want them looking at a timeline and thinking in terms of stages that happen, thinking in terms of time to value, thinking in terms of there's a health risk here. And those things will be populated on a timeline that you can look and those create the milestone events that slowly but surely I don't have to get in front of every A.E. and E.C. in the organization.

I create the GUI that it will create the language that we're embedding at a deeper level. So that unified timeline is really important as we build out more and more milestones, it goes into it without overwhelming the end customer. Finally, I'm just going to go into retrospect about what my key learnings were in terms of the process that we underwent. First of all, define the product journey milestones before you even get started.

Make them simple. If you don't think they're simple enough, you can make it even simpler. Three, four, we can get inundated by steps and data and ops people tend to have that tendency. Two is targeted intentional digital outreach.

Do your automation of digital journeys and journey orchestrator very lock step. Be able to effectively communicate to the A.E.s and S.E.s how you're touching the accounts at scale and the outcomes of it. The outcomes are what you want to attain and all that you ladder back into GRR to show that you're impacting retention. Two, how do you create health scores out of tribal knowledge?

Level set it, right? There's tribal knowledge and especially in the age of A.I., people instinctively want to guard their knowledge. We have to be able to show them that we're creating this to be able to level ourselves up and not have them protected. Once it's there, we all enjoy and benefit from it.

And four is mitigation playbooks. We have to be very precise with the business of we're not just out here preventing or reinventing the wheel. There are repeatable playbooks because not every customer is unique in its world but there are patterns. And it should be clear and transparent to them when this playbook is running, what stage it's at, and be able to report back into it.

Then they understand the benefits of a robust corporate wide customer success organization. My last kind of ultra learning, and I think about this a lot, is every culture in the organization is different. And we like to think in terms of playbooks, but taking a moment to reflect the type of sales culture you have in the organization, the type of support culture you have in the organization, is very important to be able to evangelize. Because when I come later and say, "This is how much we impact the GRR," I want to be able to couch that in the culture that is above me but also the culture that is below me.

So thank you very much for your attention. Appreciate it. Awesome. Thank you, Adam.

I appreciate that. It's also really interesting to hear what happens in like these businesses that really matter. You know, you're protecting people, protecting kids, so that's pretty cool. I forgot to ask everyone to ask Slido questions, so thank you for proactively doing it for us.

So I'll go through the list here. I don't see any Game of Thrones questions, so you're welcome to throw some in there. I'm very disappointed. Yeah.

Or what else would you like to hear talk about? Game of Thrones. I've watched Severance recently. Severance, okay.

That's a mind cluster. Okay. Holly, any Holly questions? Let's see.

Okay. So first question from Tom. I have a large TLA government account, which are all self-hosted. Interesting.

Collecting telemetry is not going to happen, right? They don't provide telemetry. So how do you get this kind of usage information from your customers if they don't allow you to have it? So two things.

One is talk to your legal. If they're online and you put it in your EULA, it's okay. But in all respect, you're going to go to legal, say, hey, do this. Why should they?

The most important thing is to start by a point collection that is under legal framework, that everything that's sensitive in PII is blocked out. But the important thing is to come back to them. If they give you logs to say, I'm going to show you how you can use the system better, right? And today in SaaS environments, we expect this.

And our customers, they may be government employees that don't want to share data, but they're also Facebook customers and Instagram customers and choose SaaS in their day to day. They understand today the value of sharing data will help them use the product better. The question is how you can do a one-time collection and then provide value to the customer and being to have a conversation with them that all of a sudden you're providing a personalized guidance to him as a CSM or whatever in the organization that it's not about him trying to be creepy and slip out anything. I'm trying to help you use the product better.

From there, you can move to a stage where the collection is more automated. But even in our on-prem product, customers that are very, very supported, we do collect logs occasionally that are under legal confidence. And we commit coming back to them and explaining them what we learned and how they can get more out of their product. So that's what I would recommend on that angle.

Okay, awesome. I love all these really interesting questions. So number two, Mr. or Mrs.

Anonymous, in a highly secure environment, how do you decide which data is useful enough to centralize the gain site versus which data should completely stay out of the system? Interesting. My inclination is centralize it in gain site and visualize it selectively. I don't see any restriction to passing the data from a system admin's data architect type of thinking, but surface it the way that you use your tabs in C360 very selectively.

I'll give you an example. Our NPS data is good, very robust. We found that it's absolutely not predicted. It's not like data wise.

It's the last tab over, right? I don't want to draw their attention to it. So the question isn't what you want to centralize in gain site. It's what you want to call attention to the CSM.

And the more proactive CSM will go out and they'll dig what they need. I want them to be able to have that type of functionality and access, but I do want to wire the GUI to where I know that there are levers that they need to pull. And I need that interaction with them to say, hey, actually, this you thought it was really you, I need it. Those are the type of conversations that we continually have within the CS organization to make sure that we're not deciding for them.

But I would recommend as long as you can, there's no reason that all of you shouldn't live there. It's a question of what you make accessible. Okay, you're coming in hot with the questions. I love this at the end of the day.

The next one's great too. How do you balance the need for predictive and proactive engagement with risk of over alerting teams in a mission critical customer environment? So the CTA fatigue, I'm thinking here. Yeah, yeah.

This is so this is this is a very interesting question. And I think that's the reason that my my my learning about the journey orchestrator was to be very systematic and conscious of it. Because we have historically blasted our customers from salespeople from marketing, I think every marketing department is kind of as send in over communicating to customers. So we did walk into this with a sentiment from our customers of celebrate, you're always trying to get me to do something or buy something or go somewhere is what my thinking was is how to not over alert teams is to start creating the code that when a CSM is the one sending it, he's not trying to sell to you.

He's not trying to market to you. He's trying to give you value in the product, right? That doesn't mean that marketing and selling is bad, obviously. But it means that this institution in celebrate is not going to get you to demo you.

It's really creating that acquisition. And in the brains of users, that when they see Travis's name, they know that it's about adoption. So they don't get that fatigue of selling that fatigue of over communication, but know that there's value in it. And our customers, I believe, I know the stats are that our open rates through Gainsight Journey Orchestrator are 60%.

Our marketing open rates are 2%, right? And it's not because they're bad. It's because we're talking value. And that's what users want to see versus driving new business, which is more wider audience.

So we want to be selective, run a handful, make sure that it's valuable. And every quarter we build out probably two more Journey Orchestrators, evaluate them, run it back, see if it worked, scale it back if it didn't have the resonance that we wanted. Okay, awesome. A couple more.

How about the milestones that you showed earlier? Do you share those adoption milestones with your customers? Like, is it part of your value conversations? I think it's right now, not really.

It's more about aligning our internal language. And we're in the early stages of that. But I am hopeful that the AI agent that we're developing now off of Snowflake will reflect a CSM and generate a deck that he can share with the customers. Because we've learned over time that initially we were very hesitant about sharing even the data that we do have saying, hey, you have six licenses that haven't been activated.

Now we're very transparent. So we will begin sharing that selectively with customers to get them thinking, hey, this is what's successful in other agencies. Because just like any enterprise, right, is saying, hey, your competitors are doing this. Well, they're not competitors, but they still want to be the top dog, right?

Small police agency wants to know what's working for NYPD and saying this is how it's working for other companies. So we will share them selectively in automated business review decks that we're preparing. But right now, no. Okay, last one.

What was the biggest trust building moment with your internal teams who were skeptical about using AI and skeptical about these real time signals specifically in this regulated environment? I think we just came to market with a new AI tool. And initially, every new technology that comes out, law enforcement is very skeptical of it, running it and being able to see what it does for you. It's just that simple.

And then if the value is there, I think we're at an inflection point in AI that there isn't a person that doesn't understand the productivity boost that you're going to get to it. So the pressure is so strong that it's coming from the bottom. It's just that simple. You did it, Adam.

All right. Thank you so much, Adam. Thank you all for staying. And I will see you all next year.

Thank you, guys.