Thank you so much, David and Andy, joining us here. A few housekeeping items here. Everyone probably knows the Knicks are playing tonight. You got San Francisco, my apologies. You got Boston, apologies. We're in a fantastic-
Will not New York have its time in the sun?
27 years. We're finally there. Listen-
Win it for Patrick.
There you go. We'll have this session in three different sections. One is a basic introduction, the second one is some questions, third is a fireside, sort of like a quick round of questions, then we open up the forum for everyone to ask you, if that's okay.
That sounds great.
Just as a real background, I'm in N.Y. I'm one of the managing directors in the TMT team. David, for those of you who do not know him, he's the CFO of Cellebrite, he's had over three decades of leading finance organizations at various public companies including his last role at New Relic, where you build a business from a very small business to almost $1 billion.
Yep.
You've run many other organizations in finance roles as CFO of C3 AI and others, run profitable businesses.
It's a pleasure to have you here. Andy, for you, congrats on your new role to take on more responsibility running treasury in addition to the IR.
Thank you very much. Yeah.
When we think of Cellebrite, the lore is that every time something very public happens, somebody's phone sort of gets involved in a potential government investigation or other investigation, whether it's an iPhone, Android, or another locked device, there's two companies that show up, Cellebrite and MSAB. Any thoughts on how is that all make-belief? Is it real?
It almost takes you back to a world that we're maybe almost pre-social, everything was anecdotal. You're right. A disaster happens in the world, there's something that's really hard to be unlocked, and suddenly somebody says, "Well, okay, let's go to the experts and have them unlock it." Whereas the vast majority of the time, the devices that people put on our platform are already consent-based and unlocked. There's this weird reality of, yes, there's something with an elected official, or there's an issue with a child, or there's an issue with a high-profile adult, and suddenly the experts swoop in and they almost think of Cellebrite, and there's a circus trick, and suddenly the phone is magically unlocked, and wealth and insights eternal spring from it. It's pretty far from how the company works.
I say that from the standpoint that the vast majority of what we do is somebody has a SIM card, an SD card, could be a tablet, could be a computer, could be a mobile phone. The ability of the company to run a full file extraction, including over 500 applications, 50 cloud services, to take that which is encrypted and suddenly decrypt it, and take a device that could be anywhere from 60 gigabytes up to a half a terabyte and extract a lot of sensitive data you might have even think that was deleted at one point, all of it is suddenly back and alive and fully connected. That I think is the true power that people miss, which is step one isn't the unlock.
The unlock is certainly a capability, extracting the intelligence out of any piece of device and then bringing it together with call detail records, search warrant return data, maybe it's license plate reader data, could be a variety of data elements. It all starts with the superpower of the company. Even when somebody might use the competing platform to unlock, many customers come back to our extraction capability because it feeds the investigative life cycle. Let me pause there.
Yeah. No.
That might've been a different answer than what you were-
No, this is great. I think you obviously sort of become a very important element to protect a lot of people, frankly speaking, and uncover a lot of stuff that could go wrong. Thank you for doing that. There's also this whole notion of we had the Apple WWDC event that happened Monday. Android's going to come out with a new software. There's a constant sort of call it a cat or mouse, to frame it-
Yep
in a sense that platforms come in, software evolves, new features get developed, new security protocols get implemented. How does that work from your R&D team's side? Help us think through how do you plan for it? How does it work internally-
Yep
at Cellebrite?
Well, I'd say the good news about every OEM and every platform is bugs are a way of life. Bugs, vulnerabilities, those are the opening points. Why do they happen? Well, frequently they happen for intentional reasons, and the intentionality of it is really just as simple as when you look at a device, there are trade-offs that are made between a device as it's architected for security and a device that's architected for the user experience or the application performance. It could be usability in the form of every device now is cross-platform, where whether it's Android or an iOS-based device, could be a BlackBerry, could be one of the Chinese manufacturers. They are designing a device so that it can work with the Windows ecosystem, has to work with the Apple's ecosystem, has to work with the Google ecosystem, has to connect into Meta.
These broad platforms all involve a series of trade-offs. What Mythos brought up, I think, was the idea that certainly, yes, there was a variety of bugs, and it certainly had the ability of fixing bugs. I think what you'd hear from our research team is when we gain access, whether it's an encrypted app, could be an encrypted device, it frequently, one, sits at one of those intersection points that I just shared. Also it frequently tends to be logic-based or reason-based, and we are typically taking eight, 10, 12 different bugs and vulnerabilities, chaining them together through what becomes the exploit, and the exploit takes advantage of them. That's ultimately what gets us access to a platform or gets us access to an application or a cloud service.
The research in and of itself is one where we will continue to use Corellium, the company we acquired last December. We will continue to use AI. We will innovate faster because the world continues to innovate faster. Good news, as the world innovates faster, they are building more applications, more capabilities. There's certainly trade-offs with security. All of these create bugs, vulnerabilities which lead to exploits. Our view is, yes, the world has certainly hastened. If you were to look at a treadmill, maybe the treadmill's spinning a little bit faster, maybe it has a little bit more pitch to it. That's the nature of the environment through which we are ultimately competing and working.
Maybe just a couple of points to build off of what Dave's discussed. When you think about the Cellebrite solution, it's a hardware-enabled software solution. You can't just simply connect a USB-C cable to a PC that, or to a PC that's got Mythos loaded on it or has access to Mythos. That won't get you access to a phone. We're really looking at the intersection of the network, the hardware, the firmware, the OS, and the application. That's where our solution is pointed. The hardware has capability in it. It also helps to protect and preserve and promote the IP built into the software. As you think about the work that our mobile researchers do, they are using AI tools themselves, to make themselves more productive.
I think that we also have invested more in mobile research, and that is not only in our teams and our people and their tools, but also in third-party organizations that complement the skill sets and trade craft that we bring to bear. As we do that work, undoubtedly, we will get value from Mythos and/or Mythos equivalents coming from frontier models, coming from the open source community. As Dave said, the treadmill runs faster. I think that treadmill requires a higher intensity investment from companies like Cellebrite that also raises the barriers to entry into our market. You started with the concept of there are only a couple people out there who can do what you do. We think we're the best at it, but we have good and credible competitors in this space. It does dissuade the high cost of entry, does dissuade others.
I think sort of related to that, the value of what we deliver to our customers also increases.
No, thank you. I wouldn't have thought of it as the barriers are so high. Outside looking in, it becomes sort of very different versus when you are explaining it the way you are.
Go back to David, obviously you had built many businesses. You were one of the top CFOs. You could go to any company. When Cellebrite came to you or you went to Cellebrite, what exactly did you see in this business that sort of gave you the confidence, here's the one that's for you. Here's the one that is not a point solution?
Well, maybe the nature of the relationship really probably started with one of my favorite board members from a past life who had worked closely with Tom. I'd say the secret ingredient maybe in any tech company really comes down to the people. For me, that always is a key differentiator just in terms of the human capital, it's vital to how we approach innovation, go to market. I think as I got to know Tom, and certainly getting to know Tom through the eyes of somebody I respected and trusted and two people who just worked closely together, that was the starting point.
Maybe I had a personal bias, I'd say, coming in the sense that even if I think about my own community, law enforcement was something that's even active at our church and the school where we send our kids from the standpoint that one of the great detectives or members of the police force of San Mateo is somebody that I saw in that community. As I talked about with Tom How, all crime is becoming digital. You could almost rewind the clock to when the iPhone came out, and you started to see that real fundamental shift in technology.
I could think about the challenges of somebody that I could see in our community, and then the challenges associated with all digital technology becoming part of crime, and how that has just changed over really the last 15 years in a very fundamental way, where all crime has a digital component to it. I took my experience, I'd say from maybe my past job of thinking about digital transformation and thought, gosh, the government is probably behind financial services. They're probably behind life sciences. They're probably that last frontier of our economy that is not digitally transformed.
I thought about digital crime, digital transformation, and I looked at all the products that were on the Cellebrite platform and said, "I think we can do something special." Particularly given the opportunity with AI to be able to bring together potentially a platform of six or more solutions for law enforcement where you could fundamentally change that experience for how a detective works or how does a DA work. That was kind of my thesis of if you start with human capital, you go with a macro tailwind of digitization and what does digital mean to crime, and then how do you respond at a local level or a national level? It seemed as though we had a lot of real estate to pursue in it. I'd say the last piece is it was a mission that was worth going after.
Crime picks on people, it kind of picks on people in an unsuspecting way. I looked at this as a company that we need to succeed. We're better off to have Cellebrite successful in all of our communities because the crimes pick on children, and there's exploitation of classes of people, and, well, we have that opportunity to make life a little bit better.
Yeah. Thank you. Thank you for sharing it. I want to go back to Andy's point. You touched up a little bit on Mythos. Not a single conversation would be complete if we didn't discuss AI.
Yeah
Foundationally, a lot of investors, some of them sitting here, probably got trained on this. We spent about 10 years thinking about per-seat model, right?
Yeah.
As we enter this, call it agentic era, some of this per-seat model is getting challenged, right?
TAM's getting challenged. How do you view the agentic opportunity? Is it a compression because intelligence is being layered in and you could potentially need fewer people analyzing this information that you are trying to sort of extract out?
Yep.
Help us contextualize all that for us.
Oh, for us, I think AI is really a very natural extension of where the business was. If the business, again, kind of rewind and think about when you got your first smartphone, it may have been 2008, 2009 - 2010. Our business started off with an ecosystem of about 50,000 or less. That was the examiner population. It's a super small ecosystem, and it started off not logically because, call it 2010, there weren't that many smartphones or other media that was associated with crime. One of our customers had shared a photo. They showed a photo of where the examiners. God bless you. Where an examiner sits, and it was somewhere between where the janitor's closet is and the payroll department. There were no windows. There was cinder block.
Since then, our TAM continues to open up because the examiners can't keep up with the amount of digital evidence. Detectives, investigators, the analysts, the entire team that supports a District Attorney, including all the ADAs and all their researchers. Everybody wants access to the digital evidence. AI for us opens up more and more TAM, just like our case evidence platform, Guardian, where we have an AI viewer built into it. We have a full chain of custody solution. Connecting operational data with investigative data as it leads into judicial data is a super valuable place where we touch more and people. For me, AI is just another way to engage a larger opportunity.
Yep, the company you brought up, once upon a time when we were a much younger company, we had perpetual licenses, then we went to term-based, and we've moved into consumption a while ago. Everything that we sell wants to have a consumption component, if for no other reason than there are people and there are cases. The cases really drive the workflows for everybody. There is a natural consumption component that says, "I need storage, I need tokens for compute, or I just need raw compute." As we've been moving our model forward, there's a nice way to continue both serving our customer well and very affordably, but doing it on the basis of consumption.
Maybe just to further frame some of that color and commentary, we've had AI embedded into our offerings for the better part of a decade. We couldn't monetize off of some of that, but it was sort of table stakes. If you are extracting data from a phone, there could be 10,000 pictures. How do you very quickly synthesize and pinpoint the images that are going to be most relevant to an examination? Proprietary machine learning helped us do that. As we've grown and expanded and extended our capability, we then had an on-prem analytic product called Cellebrite Pathfinder, and we've seen some really nice success with that product, and it continues to see really good adoption with some of our larger customers, particularly as we look at defense and intelligence use cases. An extension of the downstream consumers, the analysts, the detectives, the prosecutors that Dave mentioned.
At any given point in time, there are 100,000 of those who are actively reading the reports coming out of the labs through Cellebrite technology. Giving those individuals better tools, better capability to very quickly not only understand what's on a device, but correlate that against a dozen other artifacts, license plate reader data, call detail records from the cell towers, warrant returns on social media sites, et cetera. All of that becomes incredibly powerful and a force multiplier if you can build the right agentic capability. Actually, timing is awesome in terms of earlier today we announced general availability for Cellebrite Genesis.
Congrats.
We think that's a product that opens up our TAM in a very tangible, and significant way.
In light of that, congrats on the launch of the agentic products. Cellebrite Genesis Guardian, those are, I would assume, small relative to the overall business. How do you educate the marketplace on ARR guidance? How does all that sort of work?
I think for an agentic AI product, I think we've been very open with folks saying adoption is what's most paramount. The fact that we have customers from 15-plus countries now logging into a Google Cloud in the U.S. for investigative purposes, I think went a long way of saying product market fit, very much there. We'll figure out if we need to do regional clouds. With more likelihood, we probably do. Our first test was almost that ChatGPT moment that everybody had at the end of 2022. Is the market ready for agentic AI? Are they ready to solve crime with agentic AI? Are they willing to go through that type of experience? I think the feedback has been incredibly positive, and it's almost a referendum. Yes, the detectives need it. The investigators need it. The DAs, they need it, too.
Our entire orientation was this is an engineering pod that's been hard at work. We weren't sure if they were going to be able to bring product to market this year, so we kind of left it intentionally to the side. Now that we've started to see customers use it, we are pushing on adoption, and we'll continue to push on it. It will certainly contribute to ARR this year. The thrust of it, though, is, again, continue to drive adoption, take the ARR that comes to you after customers have been using it for several months, and then just keep pressing on the adoption because you know the monetization will happen.
Yeah.
I almost think this is one of those wonderful products. You're starting to set up 2027 right now just by virtue that you're pushing on how if you serve communities well and you serve your customers well, in the same way that Chat or the Anthropic team got to monetization later, I think you get to healthy levels of monetization, but seed the market really well.
You've had experience in the past in turning businesses into a consumption model and educating the marketplace, so appreciate those thoughts. Just shifting gears more strategically, when you think about SpaceX potentially having a big day tomorrow or Anthropic.
Sure
OpenAI for that matter, why can't they do what you do? I'm a detective, I'm police department, I'm trying to figure all this out. Why Cellebrite?
Why Cellebrite versus the frontier model?
Yeah.
Is that the heart of the question? Well, it's a good Andy and I happened to kind of randomly be at a place where we had the benefit of having Dario from Anthropic come talk to a variety of folks, his approach was, "I'm building a platform." It was very platform-centric, then I thought he offered a very thoughtful perspective, saying, "Coding. That's a great application on top of the platform." That makes a ton of sense, or Copilot for productivity made a ton of sense. In his conversation, he drifted over and he said, "HIPAA." Well, HIPAA gets a lot more challenging. HIPAA's kind of a difficult thing for us. I think when you think about law enforcement and the CJIS standard of enforcement, even PII data falls under the CJIS standard for how investigations work.
If you approach the idea of a search warrant, well, a search warrant has many dimensions to it. A lawful investigation, understanding the construct of a search warrant suddenly becomes more important. The other element is how you go about training and running all your prompt engineering. The prompt engineering for a frontier model that's working as a platform versus, hey, this is designed for all of the crime types that you run into. Some of it could be financial crimes, Medicare, Medicaid fraud. Frequently, it's crimes against children.
Right.
Some of it is crimes, exploitation of certain classes of people. That's difficult material to train on. You're not using a model to train another model. The team that went through this went through a very specific level of training, including how to avoid hallucinations. How do you get a model to stop working when the model doesn't have enough context is equally important. Honoring the chain of custody and knowing when you've gotten to the end of what AI can do is just as important. I guess maybe the last thing I'd offer is, there was a wonderful podcast with the CFO from Anthropic, and he talked about how a lot of what they do gets them to 95%. Well, 95% is great. I agreed with him.
For a lot of the work that the finance team can do, 95% gets me a lot of productivity. 95% doesn't get anywhere close to the standard for a lawful investigation. That may be the difference between a platform and platform-level AI versus vertical AI and what you need to be able to serve certain markets, whether maybe that is where Dario was going on HIPAA or where we're at on law enforcement. Our customers are doing side-by-side comparisons. Having a vertical orientation is aligned with the level of specificity and the way they need to work in order to ensure that it's lawful.
Well.
I hope that's helpful.
No, thank you.
That's probably the best way I can describe it.
Thank you. Andy, you mentioned Corellium. Congrats on closing that. You have SCG as well. All of a sudden, it appears to us outside looking in, now you have the defense arena open up. You are in the drone theater now. Take us through when somebody comes in with an idea of here's an asset that you should acquire. What happens inside the boardroom? How do you think about it? How do you think about the TAM? How do you think about evolution of Cellebrite beyond what you do today?
Well, Andy, do you want to talk a little bit about drones, and then we'll kind of pop up and talk a little bit more broadly?
Yeah. I think what's very interesting is if you think about where much of our business, and it's probably in the upwards of 85%, and that's where we grew up in the digital forensic area. Ultimately, you're looking at device types because those are sources of information. We don't necessarily do the extraction for every device that's out there. There are specialist tools that are used. Now, we're oftentimes the decoding engine of choice. 75%-85% of the time you'll extract from a device, for example, an automobile in the infotainment system, but then you'll do the decoding on Cellebrite. Increasingly, we'll look at opportunities to own the extraction tool of choice.
In the case of drones, not only is there a very conventional, you don't need to look beyond the front page of any newspaper today, whether it's Ukraine, Russia, U.S., Israel, Iran, to know that drones are a huge way of conducting combat operations. There's a lot of very rich information to be obtained from those devices. What's perhaps a little less understood but should be becoming increasingly common is, you're going to have events like the World Cup in the U.S., it starts next week. You don't want an unauthorized drone flying over a stadium. The responsibility for mitigating those drones now can be done at a local level, whereas prior to this year, particularly in the U.S., that was done at a national level.
There are not only military use cases and value to be mined from those devices, but increasingly, there's going to be value mined from the use of drones in everyday life. Think about prison systems. Contraband is dropped in. Drones are being used for it. Think about what happened, I think it was two summers ago in New Jersey, where there was a lot of hysteria over drones that were being flown. Airport in Texas gets shut down because of drones being flown by cartels. Drones have the equivalent of a black box on them, similar to what a plane would have. 8,000 data points per second. You'll know where it originates. You'll know where it stops. You'll know where it dropped a payload, perhaps. How the altitude changes, how the speed changes. All of that becomes a very rich and robust source of data.
If you think that phones are ubiquitous, and from an investigative standpoint, they are. 95% of cases that have digital evidence, the phone is the most common. From the standpoint of other device types that are taking on increased relevance, you have computers, and a very fast comer is drones. That helps inform the backdrop and the strategic thinking, which really from a Cellebrite perspective, was sort of a no-brainer. Owning that technology makes a lot of sense to us, and it's also instructive in terms of the form factor of that technology. Right? You have soldiers who are out in the field. They need information at their fingertips about where that drone may be originating. They can't wait to collect the drone, take it into a lab, and get the information back days later.
Having more information at the point of an event, the point of an incident, the point of an investigation matters, I think it's instructive for where we're going with our technology to give that capability to a broader range of users. I'll pause there, then Dave, if you want to build off of that in terms of other thoughts on M&A.
Well, thank you. It was a great double-click into drones. I think just more broadly, when you even look at the company we acquired in drones, almost all of their customers, if not 100%, were our customers. Leveraging our customers, whether it's at a national level between civilian and defense, or even at a local level, I think our customers are doing a great job guiding us towards technologies that are actually wonderful to add to the Cellebrite platform. I tend to be a big fan of let your customer kind of guide you because I think what we're building out organically in terms of the investigate product or what we're building out in Genesis, all of that is great organic innovation. Then I think customers are kind of pointing us naturally, either do investments that will allow us to scale the company.
They may be more mature products with great cash flow margins. Or it could be more nascent, like Corellium. We will continue to be very much at the customer, for the customer, letting them feed our views on M&A.
Thank you. You're one of the rare companies, half a billion of cash, no debt, real cash flows, right?
Financially, I would assume everything is hunky-dory, it goes right. You wake up, what worries you?
One of the areas that I probably pushed on as soon as I came in the door was, are we really running the company on AI? I had spent time, between wrapping up in New Relic and ultimately joining Cellebrite, really thinking through how would I think about the finance and accounting vocation being an AI-native. When I came in, we started off experimenting with AI. I think we had had Copilot, then we went to ChatGPT. Lately, we're working with Claude, but it really comes down to, I guess I look at it two ways. From a return on invested capital perspective, I think the best busines ses will be AI-native. That's kind of, I'd say principle one is how we operate.
That I think has been, I'd say feeding my energy of are we running fast enough? I would say from a human capital perspective, I think about it the same way from the standpoint that anybody in their 20s or 30s, I mean, they are going to be living in this world for a long time. I guess I get the benefit, I'm in my 50s, so I won't be living in the AI world for another 20 or 30 years. It's probably one of the best gifts that I can give the team, where if you're in your first 10 years of your career, I want you to build the business the right way. I always look at it and say, there's got to be somebody in the world, and this is, I guess I look at it of a global economy.
We've got to pick up the pace. We've got to run faster. If nothing else, the criminal element is picking up AI so quickly. In the world of if we serve public safety, we've got to be at the very best of our game because there's always somebody out there picking on somebody that we care about. We do exist. We talk about our mission. We're here to protect nations. We're here to protect communities. There's always a threat actor. We always think about cybersecurity as full of threat actors. Unfortunately, there's always a predator out there picking on somebody, and if we're going to be our best, our product team takes a lot of The research team, Evy's team, who's building Genesis, they take great pride in being able to innovate AI-natively, but we need to be building the business almost in all dimensions with AI.
If there's something that keeps me up at night, it's knowing we're not at our best. There's always room to go faster. There's always room to do it a little bit better.
I appreciate that. What's the biggest threat to your thesis? Is it the phones, the new devices? Is it AI? Is it the regulators? Is it some new company that sort of comes in?
Well, it's funny. We were discussing Mythos. It was Chris Wade, who founded Corellium, he's been in the research world a long time. People frequently come back, gosh, Mythos is going to come across the valley, it's going to, I don't know, AI will eat every software company. Chris had an interesting point. He said, "Changes the game, picks up the pace. There have been locks for thousands of years, yet somehow there's always somebody who can pick a lock." He said, "I'm pretty good at what I do. I feel pretty good about our research capability. I think I'm that lock picker." I thought that made me smile. I was like, "Well, probably true." We've had locks on doors for thousands of years, yet somehow everyone can still break into your house.
Chris, from a research perspective, I think that's probably one of the things that's misunderstood. The thing that's on my mind lately is I've probably realized, maybe I'm just on the slow train in this job, I've realized how networked the world is on crime. That is probably the thing that bothers me maybe the most on the job is the pedophiles seem to know each other. The cyber terrorists seem to know each other. Almost in the same way Andy and I were discussing it. Well, portfolio managers all seem to know each other. Hedge funds seem to know each other. The criminal networks, they actually know each other. The thing that kind of, again, it almost goes back to the AI statement. We got to be at our best. It's not because I'm worried about a competitor necessarily.
I'll say, sure, I always think about competition, that voices me. Our mission of serving nations and communities, we've got to rise up.
No, we are very grateful for the work that your company does. I'm sure those who are impacted by it are grateful as well. With that, I know I've extended my time, but I do want to open up if anyone's got a question or maybe two.
Can you talk about Genesis versus Guardian?
Sure.
I'm a little confused about where Genesis fits in. Is it an add-on to Guardian or a replacement?
It's a great question. Sorry.
Do you want to repeat the question just so it is picked up on the mic?
There was a great question about two of our brands. We have a brand called Guardian. We have a brand called Genesis. Sorry, maybe we should have invested more in brands. That was the cheap side of frugal CFO's fault. Guardian is everything for case evidence. Once you run an extraction, an examiner or anybody in the law enforcement will move that evidence into Guardian. Within Guardian, you have full chain of custody. You have storage. It could be either AI-enabled or cold storage, but you also have the ability to collab and share. It also uses AI to actually tag all of that information. Once you have that evidence stored within Guardian, you have the ability to move it into the agentic product.
You could also, within the Genesis, the agentic AI product, grab data elements from a records management system, a case management system. You could import license plate reader data. You could import minutes from prisoner calls from the prison to any personal calls.
You can input all that data to Guardian as well.
You can use Guardian to store all of your data. Genesis, though, doesn't depend strictly on Guardian. It's easier and it's more efficient to move it over from Guardian into Genesis, but Genesis has the ability to import from anywhere.
From a sales motion standpoint, it's intended to be an add-on to Guardian.
It's 100% of an add-on. They're both consumption products because our case evidence was already priced on ingest levels, and the Genesis product will be based on your tokens or your credits that you need.
Generally, you use S3 storage for that?
Most of Guardian is on S3.
As you think about that opportunity, I think either way, Cellebrite's well positioned to gain more wallet share with the customer. They want a system of record. They want to formalize and standardize how they store evidence, whether it's phone data, whether it's other evidence types, how they manage cases.
Delegating tasks to other detectives.
Effectively, we're saying that the query tools on top of Guardian were relatively weak, and this is a state-of-the-art query tool on top of Guardian.
I wouldn't say that at all because Guardian was never. Guardian had an AI viewer built into it, and that viewer was really designed around the mental health of detectives to say, "You don't want to look at pictures of rape, exploitation of children, or anything else." We will use AI to meta tag all of that imagery so that it can actually move through the detective process, ultimately to the human in the loop. It was never designed to have an analytical. The agentic AI product is actually that addition.
Think about when Guardian was first introduced, it was focused on the forensic workflows of customers.
Correct.
Had single phone analytic capability, AI capability, Gen AI capability built into it. As you broaden that lens and now you're supporting the full investigation, not just the examination workflows.
Now you have the opportunity to apply agentic AI. The AI has progressed alongside the birth of Genesis. Think about it as, you could use Guardian Investigate with an agentic AI. All of this is sort of grounded in Cellebrite AI. Think about Cellebrite AI inside of Guardian Investigate, Genesis powered by Cellebrite AI. Regardless of the choice that the customer makes, we can benefit. You want a standalone analytic tool across multiple evidence types? Genesis is your option. You want to standardize the way you conduct investigations? Guardian Investigate will be the best path. In some cases, within the same agency, they'll use both.
How do I understand price points for both today? Sort of, what's the seat-based pricing and what's the consumption amount and what averages look like?
Yeah. It's a great question. On case evidence, it's already on ingest levels, and so for every agency, they have an idea of how big their cases are. As an example, for some, a mobile phone extraction might be 60 GB. In some cases, it can run a half a terabyte. Each agency from a city all the way up to the federal level, they have an idea of how big is a case file. That drives, for case evidence, how big of an environment you buy. For a product like Genesis, again, some cities might only have one, two, three cases that they're prosecuting in any given month. That will drive them into a subscription tier that is much smaller, where they will buy only a certain number of credits/tokens that they need. Other agencies, they have 30 cases per month.
That will drive them to a higher tier of tokens. No different maybe than how you are buying anything for Claude or OpenAI. You're buying tokens into tiers.
Did only how many of the 800 seats that were in beta bought a light? Have you given any sort of metrics on what average pricing could look like and average using?
So-
The 800 and so on. That was an awful call this morning.
Okay.
That guy. Yeah.
Look, I want to be mindful. I know they have other meetings here. I'm sure that they're happy to answer the questions. I do want to be respectful of their time. Appreciate the questions. As a moderator, I'll say, "Okay.
Thank you so much for having us. This is a real treat to be here.
Yeah. Thank you.
Thank you.
Thank you so much.