Axon president Josh Isner on AI-powered policing, the $630M Prepared acquisition, and taser drones for schools

Well, uh we are pleased to be uh joined by someone who is building in a completely different space. We have the chief operating officer of Axon. The president the president of Axon. I was reading the chat. little misinformation. What's going on, Josh? Welcome to the show, boys. Good to see you.

Thanks for having Thanks for hopping on. What is this setup here? Uh that that looks incredible in space. You're in the you're in the command center. Yeah. Yeah. We have uh some folks at Axon who just on their own come out, they just churn out new Zoom backgrounds.

So, every time I open Zoom, there's like five or six options to choose from that I've never seen before. So, this one looks pretty cool. It's looking great. uh can you take us through kind of the shape of the business today?

Maybe a little bit about your journey just to kind of set the table for those who might not be familiar with the business. Sure. Uh at Axon, we make uh pretty much all the cool technology that police officers use day-to-day.

Everything from their taser gun to uh their body camera to all the SAS software in digital evidence management to all their reporting products, drones, virtual reality police training.

We really believe that uh we can add a lot of value in a lot of different uh spots where really software and hardware intersect in policing. And so uh I've been at the company 16 years, two years or two months, sorry, two weeks after graduating college, you drove across the country to uh to join Axon and here I am.

That's amazing. Uh what's been the most recent uh transformation in the business? I was talking to somebody uh about this category and uh very experienced investor. He was like, "Oh, no one's really made it work. " And I was like, "Well, what about like Axon? They seem to be doing pretty well.

" And he was like, "Ah, it's not that big of a company. " And then we looked up the market cap and it's a huge company and you've been massively successful. Uh, but what what what has been the ma like the biggest transformative moment over the past five years for the business? Sure.

I think uh really for the first 25 years of the company's history, it was really about hardware and then the last five years it's been far more about software. So, we manage more than 40 times as much content as the Netflix library combined. So, we're managing all of this police evidence in the cloud.

Uh, and in figuring out how to make use of this ma massive data set for our customers has been the name of the game. So, most notably, uh, a couple years ago, we released a product called Draft One, which is a Gen AI product.

Analyzes the body camera transcripts and creates the first draft of the police report for the police officer. So cops spend about 50% of their time writing police reports. We think we can make that like 10% of the time. Good use of GI.

So is that something where that an officer in the middle of an action, let's say they're pulling somebody over for speeding, do they talk out loud and say, "All right, I'm you know, how how are you how where is that data coming in from?

" Because I could imagine they're just kind of John actively uses like transcription with chat GBT. So, we'll be like talking about something and then he'll, you know, be prompting it just with voice. I could imagine an officer does the same thing where they're saying like, "Hey, I'm pulling this over.

I'm at this, you know, intersection. I, you know, clocked them going 20 over. " And and then that just gets built in or what is that workflow like? Sure. So, it's a great point. cops are getting better at prompting the AI to uh to, you know, improve the quality of the police reports.

But even without doing that, uh just in the normal conversation, you can imagine that same incident when the cop walks up to the person's window and says, "Hey, you know, you're speeding, you're going 20 miles an hour over the speed limit, still pick that stuff up. " Yeah. Yeah.

And even just like uh like, "What's your name? " name and then the person says that and then that's in the audio file and then that gets transcribed and that's just like so much manual work. You're not putting anyone out of a job. You're just making their life better.

It's it it feels like a fantastic use of artificial intelligence. Very exciting. And you guys have been you guys have been super inquisitive. I think you've done a number of acquisitions but recently completed uh it was an 800ish million dollar acquisition. We got a gong here. We'd love to we'd love to you know.

Yeah, it was about 630 million. It was prepare AI. There we go. Love that. We have a gong as well at headquarters. Big deals. You hit the gong. I love that. We're excited about those guys joining the company. Their founder, Michael Chime, super super talented uh and very happy to have him as part of Axon.

Uh is is is the strategy with that deal uh to kind of diffuse the technology over everything you do, keep it as a separate business line, cross-ell it? How are you thinking about that? Yeah. So, our mission is to protect life.

We want to keep people from getting killed in in policing incidents to include police officers. And the best way we can do that is shrink the time between the incident taking place in help being on the way. Right now, that's like a twominut cycle.

Someone calls 911, they talk to a call taker, call taker talks to the dispatcher, dispatcher figures out the right officer uh to uh to send to the scene. And we think we can automate that entire process uh in partnership with prepared.

So essentially call 911 based on the metadata of the call drone is already taken off uh to head to the incident. Uh you know over time we think we can route the proper police officers uh to those scenes without a lot of human intervention.

And so that's a lot uh of what we have to look forward to uh with with prepared moving into the future here. you guys are at the stage where you're a true platform multi-roduct.

Uh I'm sure you started with you know I don't know the exact uh the the the first product but uh in this category is it how hard is it to get point solutions through I imag if you're just building for other startups like you know a typical YC company for example they can build a really narrow simple point solution sell that into somebody in their batch it's pretty easy you know these aren't um you know super sophisticated operations whereas selling into police departments.

I imagine just the um just just the timeline alone of of building that relationship is probably really hard for for a young company. I guess like talk about the different maybe stages of the companies and getting to the point where you're truly a platform. Sure.

The the biggest part is you know you have to recognize that US public safety is less than a million total users.

So a lot of uh startups wouldn't even be able to get funding based on that TAM you know and so when you go into public safety or uh defense tech or you know this this segment you really have to be confident you can build something for the long term where you can win the majority of the users on your platform.

That's really the only way you can have a viable business model in this space. And so uh you know we look at this like hey you know there we start with the most talented teams we can find and then the most product synergy.

And so with a company like prepared or one like Fus that we acquired last year we really believe we can help accelerate their growth by plugging them into our ecosystem. And you know the products nicely play off each other and uh you know with police they want to outsource as much of the technology as they can.

they don't want to build things internally uh to uh you know to to capture their use cases. And so when we can show up with a really nice endto-end workflow from what we call capture to courtroom uh that that really has a lot of value to our customers.

And so that's part of our our thinking and hypothesis as we look at M&A and which companies to partner with. In terms of that platform, uh, is there demand from local law enforcement agencies for you to play nice with other companies, build some sort of open API?

Is there another kind of enterprise resource planning backbone that they want you to plug into? What's kind of the the software supply chain of a modern police department look like these days? Yeah, it's a great question and the short answer is yes.

You want to build something that's open that at the end of the day, this isn't about trying to help people make more money or make more effective ads on the internet. This is about fighting crime and saving people's lives.

So, when you do those kind of sketchy things where people don't have access to your data or you don't, you know, you don't work well with others, you're ultimately undermining the police department's mission of uh protecting their communities.

And so for us, we try to look at this like, hey, we want to be the Switzerland in this space. And regardless of the hardware inputs or the software outputs, we want to play nice and and leverage open open APIs to to to help our customers achieve what they're setting out to do.

What uh robotic form factors are you guys most bullish on over the long run? I'm sure you leverage a lot of drones. Is there any other we we had Tony from Door Dash on this week who built a really cute robot for delivery L4 autonomy. What are you guys thinking about?

Are drones enough uh or other things on the We want So the reason I'm smiling is I thought like I pitched our founder Rick on this idea like two years ago to do this big hugging robot that like the Michelin man.

Just imagine this thing kind of walks down the street criminal and takes them into custody so people aren't as scared of robots. I don't think that one's in our future. But uh the thing I'm most excited about is the ability to put less lethal on drones. So you can think of a school shooting uh taking place.

And if you have drones in nests in the in the roof of a school, there's no reason those things can't be equipped with taser technology. And they can fly autonomously. They have targeting software for us. A a enduser, you know, a police department uh would always decide when to use force. That would not be autonomous.

But this idea of finding ways where we can keep people safe in these mass shooting incidents before first responders ever get there, that's that's very interesting to us.

And over the over the next few years, we hope to be playing in that space more and more here with uh with some tools that can really uh uh you know again shrink the time frame where these incidents are happening from when they start to when they're concluded.

What about uh some somebody in the chat is asking about uh technology for you know firefighters uh like specifically wildland firefighters. We're here in LA as the uh fires unfolded at the beginning of this year. felt like they were so they they had too little technology. At least that was the perception.

We had firefighters relying on the nonprofit fire tracker that kept going down because they didn't even have capacity. What are you guys doing on that front? It's a tough market because fires are so unpredictable. Um but certainly feels important. Sure thing. Sure thing.

I think the most interesting thing I've seen in that space to date is these drones that carry thousands and thousands of pounds that can stay in the air for hours.

And you can imagine you could actually string those along uh with, you know, with very lengthy fire hoses hoses and fly out into the fire and try to contain the fire from within. Like those types of uh improvements uh through the air I think will help. I think in software it's it's a little bit of what you said.

It's predictive. It's based on the wind. I'm not sure that's like a major game changer. I think you got to find a way to get water faster and more reliably into these fires.

And it feels like drones over time, especially as they can carry more and more weight will be uh you know a reasonable kind of creative way to to to help there. What what are your current thinking around uh facial recognition? Obviously, image generation or generative AI is getting so much better.

Uh, but there's different parties. How are police feeling about it? How are different communities feel about it? Does it differ from township and city to city? How are lawmakers thinking about the tradeoffs there?

It seems like one of those conversations that's like forever ongoing and maybe that's the best process in a democracy, but how are you thinking about it? Yeah, I think you characterize that well.

You know, the United States is one of the few markets right now where facial recognition is not used in policing, and we haven't participated uh in that market yet. I do think we're more interested in it now. The thing we've been waiting for is to be able to demonstraably show that the algorithms are rid of racial bias.

I think that is certainly something that uh is is worth making sure that uh you got that right. and and and ultimately, you know, up until now, we we haven't felt comfortable uh with the technology, but it is improving and we're seeing those improvements.

And so, as we start to kick the tires on what this looks like, ethical design principles and, you know, like we say internally, uh helping make the right things easier and the wrong things harder.

That that's going to be kind of at the center of our thinking as we as we think about facial recognition on fixed cameras, on body cameras, and so forth. Yeah. Where's the uh where uh who has ball control on that narrative? Uh someone in the chat was asking about this documentary All Light Everywhere.

Very small box office. I hadn't heard of it until this person mentioned it. Are there other uh communities or nonprofits or I mean we we talk about AI all day and there's this guy Eleazar Udicowski who's kind of like defined like the AI doom scenario of like super takeoff. Uh and there's a couple thought leaders.

The Dwaresh Patel show is talking about trade-offs and AI building. uh who who are the key thought leaders, influencers, groups, public groups, nonprofits that are actually defining the conversation there. Sure. I think a big piece of it on the policing side is major county, major county sheriffs and major city chiefs.

Those are the large organizations that that um you know that kind of uh paint the path forward uh on on what the priorities are for public safety. But I think ultimately it's it's going to be very political.

Like we we need to get to a place where mayors and city councilors and all you know the the downstream folks they serve, the community is is looking at this saying, "Hey, there's there's far far far more good than harm here in this technology. " And um and it's it's progressing uh albeit slowly.

But I do think we're getting closer to the uh to the point where we can say, "Hey, look, you know, there's there's everybody should be interested in equipping public safety with ways that you can reliably fight crime and keep constituents safe.

" And between uh mayors and city councils and the police chiefs and sheriffs, I think that's the key combination of folks coming together on this one. What about uh on the hardware side?

It feels like that piece of the business is pretty stable, but is there anything interesting happening that you can share on uh you know there's this big American dynamism movement make stuff in America there's a trade war like has anything shifted your thinking about manufacturing just hard hardware devices in America or for you know your business over the past few years.

Yeah, for sure. Um, and and as a a company that's sold to to not only US public safety, but the federal government and federal law enforcement for a long time, um, we've we've always manufactured in the United States. So, Scottsdale, Arizona is where our headquarters is. Um, hopefully it will be in the future.

We got a lot of drama around that right now, but if we can get past that, we'll keep building tasers and body cams in Scottsdale, and we're proud of that. You know, we we we're supporters of the By American Act.

We we think it's in the best interest of the company to have uh all of our defense equipment made here and so u we're happy to participate in that process. How how crazy valuable is an individual person in uh the manufacturing process? There's been this debate going on in Silicon Valley about process knowledge.

this idea that uh I don't know if you've been tracking this, but Mark Zuckerberg has been paying hundreds of millions of dollars to people that know how to train just a very specific set of skills on how you train a large language model.

And once they know that information, once they know how to do it, they're worth so much because they understand how to do it, no one else does.

And there's been this debate about, well, could we do the same thing with uh semiconductors and bring some people over from Taiwan who know how to make that, make it in Arizona where their TSMC is already set up? Uh but does the same thing exist in the world of like hardware manufacturing?

Is is that even something that exists or is it more diffuse? Uh it exists to some extent. I think you know for us uh all of our cartridges the the things that go in the tasers that have the darts in them that fire out those are all made by uh by robots. They're all automated at this point.

So um you know we make millions of those every year. So the idea of people kind of hand winding darts and so forth, that's that uh that gets expensive. And so and robots are expensive too, but they scale really well. And so uh ultimately we've we've really made a big bet on automation uh for for taser manufacturing.

Um and uh we think it's the right bet. We've we've had automation for years and years, but we continue to invest more in it.

And frankly, the people who know how to do that really well, like you said, they might not be, you know, as valuable as some of the folks on the cutting edge of of LLMs right now, but they're very valuable to us.

And uh got to make sure, you know, manufacturing automation is is one of the things we continue to be really really good at. Yeah. Awesome.

Well, thank you for the updates and uh come back on when uh when you guys do your next maybe the next acquisition will, you know, hit hit the1 billion dollar mark, but always welcome in the in the in the 600 million range.