This week’s “episode” is a little different than usual!

First, there are three mini interviews instead of one longer one.

Second, we recorded these two conversations live on stage at the GTMfund AGM, in front of 150 go-to-market leaders, founders, and investors (GTMfund is our VC firm investing in early-stage B2B software companies, backed by 350+ go-to-market leaders).

The guests are two of our portfolio company founders that we are so fortunate to support and be on this journey with:

  • Tucker Dordevic of Austin Sensing Company builds radar, radio-frequency and acoustic sensors that detect and track drones and other aircraft over airports, wildfires and battlefields.
  • Mike Molinet of Dock is building a shared workspace where teams work with AI agents the way they’d work with colleagues.

We hope these mini interviews will give you a closer look at how great early-stage founders build, sell and get to traction, and why we were excited to back them.


Austin Sensing Company: He Raised $11M To Build “The Awareness Layer For The Sky”

Listen on YouTube // Apple // Spotify


Austin Sensing Company announced an $11M Seed round, days before this live conversation at the GTMfund AGM. Co-founder Tucker Dordevic sits down with GTMfund GP, Paul Irving, to explain why he’s building “the awareness layer for the sky” – a system of radar, passive radio frequency, and acoustic sensors that detect and track anything in the airspace, from a shoebox-sized drone to a missile.

Episode highlights

2:27 – Welcome Tucker, and the rebrand plus $11M seed announcement

3:04 – What Austin Sensing Company solves for

4:30 – How GTMfund got conviction (and the lacrosse highlight reel)

6:08 – The breadth of use cases: wildfires, autonomous vehicles, and more

9:06 – The explosion of drones in US airspace, and where it’s heading

11:00 – From a 20-deployment goal to 40, and what’s driving demand

12:51 – The coolest customer story: the drone circling a police department

14:21 – The two-year and five-year vision

15:33 – Manufacturing in Austin, and who he’s hiring

Links


Dock: Inside A 2-Person Company Run By 50 AI Agents

Listen on YouTube // Apple // Spotify


Mike Molinet took his first company, Branch, to $100M in revenue with 500 employees. This time, he’s trying to get there with 2 humans and 50 agents. In this live conversation at the GTMfund AGM, the Dock Co-founder sits down with GTMfund Partner Sophie Buonassisi to break down what “multiplayer AI business agents” actually are and why he thinks business people are about to get the 10x moment engineers already have.

Episode highlights

2:08 – What Dock is: multiplayer business agents, 2 people and 50 agents

3:04 – How to compare Dock to Claude and ChatGPT

4:36 – Why engineers got a 10x moment but business people haven’t

7:34 – What the most AI-pilled teams are doing with agents

10:00 – How a buyer’s understanding of “agent” has evolved

11:35 – Where agents sit on a GTM team a year from now

14:03 – The human-to-agent ratio, and span of control

16:30 – Building a human-style org chart of agents

18:34 – Where to find Dock

Links


The GTMnow Newsletter (by GTMfund)

GTMnow merges go-to-market strategy, venture capital and media.
Build, scale and invest with the best minds in tech.The media brand of VC firm, GTMfund – sharing the how and who behind company growth.


GTMnow Links


The GTMnow Podcast – subscribe to support

It would mean the world to us if you subscribe to the show on your platform of choice. This helps support our growth, which in turn helps us continuously bring on bigger guests and deliver the content that you want to hear most.

Subscribe on YouTube

Subscribe on Apple

Subscribe on Spotify


GTMnow shares how the best companies get built, told from every perspective: founders, operators and investors.

Visit gtmnow.com for more episodes, The GTMnow Newsletter editions, and other content.

GTMnow is owned and operated by VC firm, GTMfund.


Austin Sensing Company: Tucker Dordevic Interview Transcript

00:00 – 00:23

Sophie Buonassisi: This interview is a little different. We recorded it live on stage at the GTM fund AGM in front of 150. Go to market leaders, founders and investors. GTM fund is a VC firm investing in early stage B2B software companies backed by 350 go to market leaders. The guest is one of our portfolio company founders that we are so fortunate to support and be on this journey with. You’ll hear a short custom video about what they’re building. Then we’ll jump into the interview.

00:23 – 02:01

Video: We put fences around houses, schools and prisons wherever we want people to stay out of, but there are some places we just can’t do that. The sky is one of those and it’s changing so much for the good, but also so much for the bad. Last year, the US Forest Service logged 218 unauthorized drones over active wildfires. The seven years before that, a combined 125 one of the drones put a hole in the wing of a super scooper over the Palisades Fire. It was down for five days, and in that time the fire grew over 6000 acres. No matter what type, every incident has one thing in common no one could see it coming. So we’re changing that at Austin Sensing Company. We build radar, passive radio frequency and acoustic sensors that see what’s happening in the sky. Our systems can protect an airport, cover a city, support firefighters during a wildfire, or give the soldiers awareness of what is flying nearby. This summer, an unauthorized drone shut down aerial firefighting over the cottonwood fires. The US Department of Interior, in response, drove through the night from Idaho and deployed our towers the next morning. The threat is growing quickly here at home. The FAA recorded roughly 1860 drone sightings near aircraft and airports last year, up 11%, with reports nearly doubling in the second quarter of 2026 alone. On the battlefield, drone inflicted casualties have risen from less than 10% in 2022 to as much as 80% today. If the systems needed to detect, track and understand these drones remain too expensive, too complicated and too difficult to deploy at scale. We believe everyone responsible for protecting people should know what is above them. So we’re building the awareness layer for the sky.

02:02 – 02:11

Video: Kick my heels up and put my hands up and down through my hands I can shout. Come on now. Don’t forget to say you will.

02:12 – 02:25

Video: Don’t forget to say yeah, yeah yeah yeah. You say, you say. Right now baby. Say you will. Come on, come on. Say you will say yes.

02:27 – 02:28

Paul Irving: Tucker, thanks for joining us.

02:29 – 02:30

Tucker Dordevic: Thanks for having me.

02:30 – 02:37

Paul Irving: First off, congratulations on the rebrand of Austin Sensing Company in the launch and the announcement of the seed round just on Wednesday.

02:37 – 02:42

Tucker Dordevic: Yeah. We announced yesterday that we raised $11 million. Um, so.

02:48 – 03:04

Paul Irving: Um, it’s been been a lot of fun working together for well over a year now. We’re going to get into some of those details there, but we got a little bit of a taste of it in the video. but I’m going to start with an easy one. In 1 or 2 sentences, give us a feel for what Austin Sensing Company solves for.

03:04 – 03:23

Tucker Dordevic: Yeah. So at Austin Sensing Company, we build, as it says in the video, RF radar and acoustic sensors to detect and track anything in the airspace, whether it’s a drone, missile, you name it, we track it. We can track anything above an airport battlefield, uh, a wildfire instance, anything like that. We track it and we give that information to the operators so they have a better picture of what’s what’s above them.

03:24 – 03:29

Paul Irving: How did you originally become obsessed with this problem? Like, give us more of the origin story?

03:29 – 04:14

Tucker Dordevic: Totally. So I, uh, I come from a family of pilots. My dad is a pilot. A lot of my cousins fly. Funny enough, one of my cousins actually has a ranch, and he built a runway on the ranch. And so me and my, uh, you know, brothers grew up just like flying helicopters, planes, drones, you name it, we flew it. Um, and as I started getting older, as you saw in the video, I played lacrosse for much of my life. But when I entered the professional world, um, I didn’t quite know what I wanted to do. And so I went into finance, as any lacrosse player does. Um, and, uh, they were like, hey, well, you have to pick a group to be in. And the only one that really sounded appealing to me was defense. And so about a year in, I started having this idea I got with my, uh, co-founder Isaac, who is, uh, kind of the brains behind the operation. He spent ten years as an aerospace engineer, and, uh, here we are.

04:14 – 04:30

Paul Irving: Awesome. I’m going to I’m going to do a brief interlude and give you the origin story for for us in Tucker and, uh, originally Orient now Austin Sensing Company. So first shout out is Nick Pittman in the room? Where’s Nick? Shout out to Nick, an OG GTM fund LP and a great investor in his own right at TikTok.

04:32 – 04:32

Paul Irving: Um.

04:33 – 04:34

Tucker Dordevic: New dad?

04:34 – 06:01

Paul Irving: Yeah, yeah, yeah, he’s boots on the ground in Austin. And Matt Tucker, who’s also in Austin, given the company name, I guess that’s a given. Uh, but, uh, introduced us when Tucker and Isaac were just starting out. And the funny thing is, Max and I were pretty sold on the problem to to to begin with. So we were texting each other back and forth in the Potomac helicopter airplane crash in Washington, D.C., in January of 2025, and just reading about all of the shortages of air traffic control agents, the incredible amount of near-misses. I don’t think I’ve wanted to fly into Newark Airport since we did the due diligence. Um, and then, you know, we were thinking like, if New York Times is writing about this, how is there not more startups solving this problem? And then boom, we got intro. It was very serendipitous timing. We had one meeting. We booked a next meeting immediately after spending time and time together. Got super excited about it. And then it was late in the evening. We I was going to send you a term sheet in the morning, and I was in the office and I was like Tucker mentioned, he played a season of professional lacrosse. Like, I’m just going to pull up the YouTube highlights. And they were sick. You undersold it. I’ll give you the stats. He played one season rookie of the year, all star second team All-Pro, all while working at a luxe capital fintech company. And I called that founder and he was like, Tucker was the hardest working person we had in the business. Well, well, well paying professional lacrosse. So we knew it was a special team, and you guys were going to go on to some special things, but that, that definitely cemented it. So this is a very simple yes or no question. Have we or have we not convinced you to put lacrosse highlights in the series a data room?

06:02 – 06:04

Tucker Dordevic: You probably have, but

06:05 – 06:08

Tucker Dordevic: I always joke it was a past life. And so it’s a it’s fun to see it here and there.

06:08 – 06:27

Paul Irving: Yeah, yeah. We’ll let you relive the glory days. Um, we mentioned a bunch of the use cases you serve, but it would be great to dive into them a bit. So we mentioned, um, wildfires, Cal Fire, the Department of Interior. I think we even talked about Burning Man requesting a demo at one point or another, like, give us a sense of just the breadth of what customers are asking for and what you guys are doing for them right now.

06:27 – 07:39

Tucker Dordevic: Yeah. So as I said earlier, we make three different sensors. Each sensor can be used altogether and what we like to call the tower. So it just has all the sensors on it. Or we can kind of sell them individually. Right. You can put radar on to an autonomous vehicle, which we’re seeing a lot of with the military in companies like Saronic or, you know, Allen Control Systems, which are two awesome based companies. Um, you can take our sensors and deploy them around a wildfire. What that actually looks like in practice. You have this suitcase sized Pelican case, you open it up, you can deploy all of this stuff onto a tripod within like 30s to a minute, and that broadcasts anything 40,000ft in below. So we’re able to not only find the drones that are on the sky, but we can locate the actual operator’s location. So for instance, it was mentioned in the video, um, but during the LA fires last year, the largest tanker hit a drone, um, because they had no clue what was in the sky. That is obviously an extreme example, but if there is a drone in the vicinity of any aircraft, all aircraft have to be grounded. So if you kind of look at, you know, the scale of the LA fires, if one drone that’s, you know, the size of a shoebox is in the sky, you have to ground, you know, ten, 15, 20, uh, aircraft, uh, therefore causing, you know, LA to continue to burn. So there’s a bunch of different use cases. We’re kind of seeing it all across the board.

07:39 – 08:06

Paul Irving: Really, really interesting. And I mean, when we were first talking, we were, I mean, as you do in the early stages of a company, like, what’s the terminal vision for Austin Sensing Company and where you guys wanted to build towards. But there’s been an evolution as well. Like, you can’t just become majestic air traffic control at Newark Airport overnight. There’s a lot of pathways and work that has to be done to get there. Can you give us a bit of a sense of what that evolution look like as far as like the Pre-seed round when we were first partnering together to, you know, what the company looks like today and what you’re doing for customers.

08:06 – 09:06

Tucker Dordevic: Yeah, totally. So when we first met Paul, it was very much like an idea on the on a deck. We we hit it off. Uh, we have had a great relationship since, uh, but we essentially wanted to use ground based sensors to cue aircraft in the sky. What we found was we, you know, a big focus for us is, you know, kind of classically, when you hear these deep tech, hard tech companies, it’s like, okay, you can’t solve anybody for a long time, and you have to raise a lot of money. And we didn’t want to do that. We wanted to be able to start selling to someone right away, kind of like build this, you know, B to g muscle as early as we could. And so what we found was the sensors that we made, people really liked using them to detect everything in the sky. And they were kind of like, hey, you know, we can’t tell the, the planes what to do. But what we can do is tell our pilots what to do, and then we can, you know, use it for our own safety. And so we were like, hey, well, we’re pretty good at making sensors, so why don’t we just continue to make those? Because people really want them. And we started to see an evolution, whether it was autonomous vehicles and all this other stuff. And so here we are making all three of these modalities and we sell to all the other, you know, all the sectors I mentioned earlier.

09:06 – 09:29

Paul Irving: Um, we talked about it in the video too, but we’re seeing I don’t think people realize it, an absolute explosion of unmanned aircraft, mostly drones in our airspace. You have police and fire adopting them rapidly over the past few years. You have delivery companies like Zipline and Amazon. Um, what does that mean for us and sense some company. But like, I’d also love to just get your feelings or thoughts on how that trend you expect to evolve over the next coming years.

09:29 – 10:38

Tucker Dordevic: If you look at the DFW slash, the Just Dallas area, there’s more flights, drone deliveries in DFW than there are flights leaving in and out of DFW airport every day, which is kind of astonishing. Um, if you look at drone deliveries. Kind of in aggregate from 2024 through 2035. It’s gonna go from like 2 million annually to, they believe, 800 million annual drone deliveries by 2035. Um, and then, you know, sadly, you have to talk about the, you know, what we’re seeing in Ukraine. Ukraine makes 50,000 drones a month. Uh, it’s just kind of an unfathomable scale of these autonomous vehicles that are being used. We were at a demonstration with the Marine Corps, uh, like two weeks ago, and they were flying all these different types of drones. And it’s, like, the scariest thing in the world, and they really have no way of actually tracking them if they have no sort of sensing technology. So the magnitude, whether it’s, you know, for good or unfortunately for evil, is going to kind of skyrocket over the next ten, 15 years. And, you know, there’s people are getting more creative with how they’re, you know, putting these things in the sky. And so we’re trying to build the best sensors in the world to, to track anything. As I mentioned earlier, whether it’s a drone or a, you know, a missile or plane, etc..

10:38 – 10:50

Paul Irving: Yeah. I mean, I think that was the part that always stood out to me. I don’t think people realize, like, how close to home this is going to become for a lot of people, and not necessarily in an evil or bad way. But just like the sheer volume deliveries. Police, fire first responders.

10:50 – 10:52

Tucker Dordevic: Look at the Tesla launch last.

10:52 – 10:53

Unidentified Speaker: Yeah, yeah, yeah.

10:54 – 11:15

Paul Irving: Um, so at the beginning of this year, I remember you and I were syncing up and we were saying, like, because you mentioned it, public sector sales is slow. It’s difficult, it’s challenging. It’s it’s own unique beast. And we thought like what is a good amount of deployments or like live pilots by the end of the year. And I think you had your sights set on a really lofty number, but we landed on like 20 being a really good stretch goal. Um.

11:16 – 11:17

Tucker Dordevic: Talked me down that.

11:17 – 11:30

Paul Irving: I did not. Um, and you guys just hit 40. Um, so blew through 20 earlier this summer, already at 40. Uh, tell me what’s been driving a lot of the customer demand from early, early pilots and early users?

11:30 – 12:50

Tucker Dordevic: Yeah. Um, I think one thing and going back to the evil, which I hate to say is I think, you know, people have Instagram, they have Twitter, they see these horrible acts of terror that are being, you know, done in Iran, in Ukraine. And that’s definitely scary to a police department. What I say is we’re just starting to see an evolution of Walmart delivery drones, right? So we’ve been in we’ve been deploying with the Texas Department of Public Transportation. There’s an onslaught of Walmart zipline, as you mentioned, that are starting to deploy around places like Dallas, around Austin. And they aren’t giving like police departments access to even the location of the drone. So these police departments are flying drones, first responder. And they’re going to get to these scenes before, you know, cop arrives and they have no clue what’s in the sky. And so we’re giving them that visibility. On top of that, you have places, uh, you know, just police departments, fire departments in general who want to have some awareness above schools. So, like, as crazy as it sounds like, people use these things to like, surveil other people, which is pretty scary. And so we see, you know, a cop doesn’t want a guy looking at the kids at recess, right? It’s like disgusting. And so they want to be able to have awareness of that. And so we’re able to give them, you know, the ability to see what’s in the sky. But more importantly, we’re able to give them the ability to actually find the person. So we triangulate the location of the remote control in the asset and the sky, and pinpoint the person so that they can send an officer to say, hey, man, you got to lower this down. You can’t be flying here. Um, so on and so forth.

12:51 – 12:57

Paul Irving: What’s the, um. I’ll go a little off script here, but I know you have a lot of these stories. What’s the coolest story or piece of feedback you’ve heard so far from a customer?

12:57 – 14:20

Tucker Dordevic: Yeah. So we were at a small town in Missouri called Cape Girardeau. It’s about two hours from Saint Louis, and we just deployed. And I was sitting there with, uh, the sheriff who was leading the deployment. And, uh, a drone starts flying over from like, two miles away, and we’re watching it fly. And he’s like, oh, that’s cool. Like, you know, pick it up pretty quickly. And I didn’t think much of it. And it kept coming closer and closer and closer, and it just started circling the police department. And so we’re sitting here and there’s this drone above that. We have no clue whose drone it is, is circling the police department. And so we kind of keep going along with the onboarding and demo and stuff like that. It kind of goes back and then it comes back again. And so he goes, look, we got to go outside and we go outside and it’s this like small little, you know, six inch by six inch drone that’s just circling the police department at, like, you know, 50ft off the ground. And so this guy goes well on boarding was great. I’m going to go check this out and drives to the guy with his lights on and, and calls me on my way back to the Saint Louis airport and was like, it was just some elder gentlemen who wanted to see. They just got drones for their department, and he was a drone fanatic and wanted to see the drones on the roof. So a lot of unique, uh, use cases. We had another one where there was an incursion, which just means a near miss or a, you know, a plane collides with another plane or a drone, etc., where the Department of Interior drove our units into the cottonwood fires in Utah overnight, which is pretty exciting. So a bunch of a bunch of fun stuff.

14:21 – 14:27

Paul Irving: That first story reminded me of the so that that drone that was flying around wasn’t an Austin sensor and comfrey plant.

14:27 – 14:27

Tucker Dordevic: Uh, like.

14:27 – 14:48

Paul Irving: I said, that Travis Kalanick getting like an Uber to wait for Bill Gurley. So it makes it look like the delivery times are better when the company was early. Um, what does this look like? I’d love to take us a little bit into the future. What does Austin Sensing Company doing in two years, five years from now? How are customers leveraging you? Take us into the future a little bit.

14:48 – 15:33

Tucker Dordevic: Yeah. So there’s a few things I think, you know, we want our sensors to not only be in every single city, um, stadium piece of critical infrastructure. We also want them to be on every sort of autonomous vehicle. So we’re seeing less and less, you know, humans being deployed to battle. And we’re seeing more and more, you know, autonomous vehicles being deployed to battle. How can we be, you know, the eyes and ears of that system as it’s kind of being looked at by these, you know, bad drones that are trying to take it out. And then on top of that, we want to connect everything in the sky at a more like public safety layer, if you will. Not necessarily public safety, but air airplanes, etc.. Um, going back to the air traffic control thing. So our whole goal is like, how can you build the sensing layer that then cues all of these assets in the sky so that they can be coordinated, deconfliction and essentially watch out for each other at all times. So, um, yeah.

15:33 – 15:34

Paul Irving: And you’re manufacturing.

15:34 – 15:52

Tucker Dordevic: Yes, we just moved into 2000 square foot factory in Austin, hence the name. Shocking. Um, where we’ll manufacture everything. Uh, our team is like 15, which is crazy to think. Um, but, yeah, hiring across the board. So if anybody knows anybody, please give me a shout.

15:52 – 16:05

Paul Irving: Um, so because we have a room full of mostly GTM leaders. Uh, I do have to ask this question, because who knows? Maybe your solution is in the crowd. If you could unlock one GTM Challenge or bottleneck for Austin Sensing Co right now, what would it be?

16:05 – 16:09

Tucker Dordevic: This is a cop out, but the government’s very slow.

16:10 – 16:15

Paul Irving: I don’t know if anyone could solve that. If anybody can help government with that that would be amazing.

16:15 – 16:37

Tucker Dordevic: Um, no. I think the biggest thing is like, what we see often is if you go to a police department, like there’s the decision maker who you’re talking to. But what sucks is there’s like also two decision makers on top of them, and you can’t even get to them until you talk to the next one. So for us, it’s like, how can we almost get to the chief right away? And that obviously takes some time for us to get our name out there. ET cetera, etc.. but if anybody can make the government speed up, I would greatly appreciate that.

16:37 – 17:21

Paul Irving: Well, you guys are investing in media and some of the brand stuff. I mean, just as you. It’s one of those networks, too, that it’s so tight knit. You know, all of these police departments, fire departments, talk to each other at the same conferences. They share tools that are working. And so that will naturally accelerate. But always a challenge. I’m afraid I don’t know if anyone can make the government move faster in this room. Um, Tucker, this has been awesome. I, uh, I wanted, in closing to say a couple. I mean, the coolest part about this job, and this is a very, very cool job, is we get to work with the most ambitious and smartest people. I feel like on planet Earth on a day to day basis. And with that, I literally can’t underline enough how much fun it’s been working with you and Isaac. You guys are relentless, tireless, mission oriented. I’m more excited about the future after every time we talk. Um, so anyway, thank you, man. I appreciate.


Dock: Mike Molinet Interview Transcript

00:00 – 00:23

Sophie Buonassisi: This interview is a little different. We recorded it live on stage at the GTM fund AGM in front of 150. Go to market leaders, founders and investors. GTM fund is a VC firm investing in early stage B2B software companies backed by 350 go to market leaders. The guest is one of our portfolio company founders that we are so fortunate to support and be on this journey with. You’ll hear a short custom video about what they’re building. Then we’ll jump into the interview.

00:23 – 00:42

Mike Molinet: Yeah, I think 6 or 9 months ago people actually didn’t even know what an agent was. ChatGPT in chat is not an agent because it’s not going and doing work. Dock is multi-player, super intelligent business agents. Basically, it’s a team in a suite of agents that you can communicate and work with anywhere text, message, email, web, slack, wherever you are.

00:42 – 00:45

Sophie Buonassisi: How should people think about comparing Dock to, say, Claude first?

00:45 – 00:57

Mike Molinet: I think Claude and OpenAI, they really open their eyes to a lot of things. I think things like Cursor and Claude, code and Codex, really good coding harness is good for engineering use cases, fast feedback loops, things that are very deterministic.

00:57 – 01:00

Sophie Buonassisi: Where agents can sit on a go to market team.

01:14 – 01:33

Sophie Buonassisi: Uh, everybody, just for a little bit of context. Mike is the founder of Dock. He is entrenched in everything. AI agents very close to it. If you want to go deep on it, he is your guy. He also writes an incredible newsletter. Highly recommend. And as the founder of Dock, we are very proud to be investors here at GTM. Fun. Yeah, I’ve been happy here, Mike.

01:35 – 01:36

Mike Molinet: Happy to be here.

01:36 – 01:36

Sophie Buonassisi: Awesome.

01:36 – 01:39

Mike Molinet: Great. Good to interact with humans for a change.

01:39 – 01:41

Sophie Buonassisi: Yes, a change of pace, that is for sure.

01:41 – 01:43

Mike Molinet: The AI ten hours a day.

01:44 – 01:48

Sophie Buonassisi: I will say nobody is more responsive than you and I have learned. It is thanks to your agents.

01:48 – 01:55

Mike Molinet: I have since started signing my human written emails with human Mike instead of just Mike. So people can distinguish. Yes. Brilliant.

01:55 – 02:08

Sophie Buonassisi: Perfect. I do, I just get a quick mic check on your microphone. Is it on? Okay. Okay. Awesome. Now, Mike, give us a sense. Just a couple sentences. What is Dock for everyone in the room?

02:08 – 03:03

Mike Molinet: Okay, so Dock is multiplayer super intelligent business agents. Basically, it’s a team and a suite of agents that you can communicate and work with anywhere text, message, email, web, slack, wherever you are. And it’s a team of agents. They all communicate, collaborate and work together. They’re specialized. They have roles. They’re persistent. They’re proactive. They’re indistinguishable from humans at the end of the day. So my agents work 24 over seven. They’re responding to customers, sending follow ups, sending you emails while we’re both mingling and having drinks. And yes, okay, so they’re doing everything. They’re running my socials. When we build Dock on Dock, we sell and support and market Dock with Dock. So all of our agents are running the company. Two people, 50 agents doing everything. In my previous my first company was a company called branches like Mobile Linking Infrastructure. I ran about a decade ago that was 500 employees. We took it to 100 million RR. I want to see if I can do it this time with two people. So let’s see.

03:04 – 03:11

Sophie Buonassisi: Incredible. And just to help people even further understand how should people think about comparing Dock to say, Claude.

03:11 – 04:35

Mike Molinet: Yeah. So first, I think Claude and OpenAI, they really opened our eyes to a lot of things. I think things like Cursor and Claude code and codecs, really good coding harnesses, good for engineering use cases, fast feedback loops, things that are very deterministic but very complex. So if anyone spent hundreds of hours setting up their quad code instance to try to do some GTM stuff, which a lot of people have, because we’re on, we’re in the bubble within a bubble, within a bubble, right? Like we are on the far. All of us are in the far end of the extreme of doing the craziest stuff that you can do. But then when you look at the 99.9% of people, they’re nowhere near that. And if you actually go look at peoples like even people on tech, they’re set up their coworkers, set their coworkers like quad coworkers set up, or some of their other setups they’re barely doing like the most that they’re doing is really chat. And you say, well, if you set up whatever routines, if you set up memory system and set up context management, all that sort of stuff. Like, no, I don’t know what that is. And we start to realize that people the analogy that we were talking about the other day was people don’t want. It’s a computer analogy. People don’t want like individual components to build their own computer. The vast majority of people just want to walk into an Apple Store, buy a MacBook Pro, click a few buttons, and then start using it. Right. And we are the group that is like at the far end of the extreme in the 1970s, building individual computers with the individual components, doing the most innovative stuff. People need that Mac moment, and Dock is trying to be that for people, especially for business people, because engineers already got their ten x moment, but the business people haven’t. Mhm.

04:36 – 04:44

Sophie Buonassisi: And you said business people haven’t. We are seeing the NX engineer become a reality. Why are we seeing the ten x seller or the ten x marketer.

04:45 – 06:24

Mike Molinet: Yeah it’s a, it’s a little bit what I touched on which is business people and your operations. And like the way you operate is what go to market is perfect use case right. It is totally different than the way engineers work. Engineers sit at a computer for eight hours tinkering, and when they’re a coding agent is doing work, it’s deterministic, it’s binary. It worked or it didn’t. If it didn’t, it runs a loop and it tries something else and eventually it might work. Also, coding harnesses and llms have been trained very well on everything a documentation and code, all this stuff that’s existed, things like go to market, sales, marketing, taste, the creative aspect that is much harder and lens are not trained. And anyone that has had Claude or any other LM or tool go write an email like an outbound email, it’s AI slop or write a LinkedIn post. It’s AI slop. Why? Because it’s trained on a certain set. But that is different than coding. And so it needs taste, it needs structure. It needs a lot of different things than what coding harnesses needed, not to mention ten like go to market people and business people tend to be more collaborative, right? Like if you think about the way your business teams work, go to market teams work, sales and marketing teams work, they’re collaborating constantly. They’re using slack, they’re using the email. They’re in Google Docs and Notion. And that when you have agents doing that stuff, if they’re isolated, if they don’t have shared context, that becomes a problem. And this is one of the original starting thesis points that we had, which is agents need ability to communicate and collaborate in order to work towards a common goal. If you’re a coding agent, you just do the thing. Close your task uploads to GitHub PR, push another agent, does the review, whatever. Much simpler in that sense. So yeah business is a little bit different.

06:24 – 06:27

Sophie Buonassisi: Mhm. I mean and long before you.

06:27 – 06:35

Sophie Buonassisi: Actually committed to Dock you were studying this space. So how did your thinking evolve while you’re digging into this problem. And how did you land on Dock.

06:35 – 07:33

Mike Molinet: Yeah. So a little bit of history started the first company ten years ago back in 2014. Started a different company in like end of 22 right before AI was actually like really useful. Um, and then realized about a year ago that we needed to be 100% in AI. And that’s when we started working towards something. Talked to a lot of people just we’re building agents ourselves. And we had hundreds of conversations, probably 500 over the last six months at least. And we first started just building, okay, what do we need? And then we started to build all those little things, and we built the agent to agent communication infrastructure and all that sort of stuff and talking to people. We realized that that that 99 point maybe 9% of people just really don’t. They don’t get the value out of AI that they should be, that they could be, because the tools just aren’t there yet. They’re getting there. The models are getting there, the tools are getting there, the integrations are getting there. But it’s we’re we’re still really far away from really unlocking that power for everyone else.

07:34 – 07:53

Sophie Buonassisi: Okay, I gotta ask, you said 99.9, but what about the, you know, less than 1% for everybody in this room. We’ve got to go to market. A lot of go to market leaders in this room to it probably is this room. But what are the top kind of AI pill teams doing with agents that you think is going to make it into most companies at time?

07:54 – 09:47

Mike Molinet: Yeah. Okay. So what are people what are people doing? So first they they’re running multiple different agents. The agents tend to be persistent, named role based. And they tend to be collaborative in some way. Now that could be on dock or it could be outside of dock. If they built some sort of custom thing, or use some sort of open source software to get these agents talking. But that is different than the way coding agents work, which is coding agents and engineers typically will spin up a session and do the task, close the session, spin up a session, close the task, do the session. With this type of work, what we find on the business side. The most AI build teams really pushing the boundary on having a lot of agents and those agents being everywhere, slack, email, wherever it might be and you’re starting to see this. You’re finally starting to see this. Things like grok bot, things like this, you know, competitive to us worse than Dock. But, you know, it’s nice to have Elon chasing my tail. Um, so that is a sign that people need that ability to work with these agents in lots of different places and then, you know, text messaging. Right? Like you were talking about just being able to do that anyway. You hear about software factories or like AI factories, basically like coding agents that all work together towards a common goal. And eventually they’re handing off work and they eventually get pretty human stops being in the middle or the engineer stops being in the middle. There’s an equivalent on the business side, it’s a little bit different. The shape looks a little bit different. But if you get to that point where the business, where the agents are running aspects of the business and the human, instead of being the, I sometimes call it a meat proxy, um, middleman, like go between copy paste, middleware, whatever it is. Right now, a lot of humans are the copy paste middleware between a bunch of sessions and agents, and that does not scale. And the only thing they’re doing is copying and pasting, putting into slack, copying and pasting, rather than actually being the one that steers and sets the direction and comes up with creative ideas and then lets the agents actually do the busy work. Probably 9,095% of my busy work has been eliminated completely, and I’m doing a lot more. And now it’s the agents actually doing it. Responding to the import tickets at 2 a.m. while I’m sleeping.

09:49 – 09:54

Sophie Buonassisi: That is the goal to be able to keep sleeping through to and have your support to go to.

09:54 – 09:56

Mike Molinet: Sleep very well now, thankfully.

09:57 – 09:58

Sophie Buonassisi: Well, things are changing quickly.

09:58 – 10:11

Sophie Buonassisi: As I mentioned, you know what people are doing in this room now is fundamentally different than it was months ago. So if you look even six, nine months ago, how has a buyer’s understanding of an agent evolved?

10:11 – 11:34

Mike Molinet: Yeah, I think 6 or 9 months ago, people actually didn’t even know what an agent was. I think the word agent was starting to come around, but the the average tech person, even the more sophisticated tech person, didn’t understand. And at the end of the day, like how you define an agent varies. But it’s an LLM, it’s memory, it’s context, it’s access to tools. It’s the ability to run on a loop. Right. Um, that’s the majority of it. And I think Open Claw was the moment for the vast majority of people to actually start to understand what that is. What what does that mean? Because I think everyone up until that point was used to using, for example, Claude Chatter ChatGPT he’d go in, you’d prompt it and say, write me a thing, and it could spit out words, but that was about all it could do. It wasn’t running on routines. It didn’t really have a good memory. Maybe in that session it remembered some of it. But across all your chats, it didn’t. It couldn’t be proactive. It couldn’t go do things on its own. And it didn’t have access to most tools. You could starting to see some MSPs, but for the most part it wasn’t able to actually do something. And then suddenly you give open claw to people that they start to see, oh, this is, this is a like this thing that can actually go do stuff. I can just say, build me this site, build me this mobile app, go do XYZ. And it went and did that. And that was for a lot of people. I think that moment like, oh, this is what an agent is. ChatGPT and Claude Chad is not an agent because it’s not going and doing work. It’s simply responding. It’s a it’s an advisor, it’s a consultant, but it’s not like a, like a higher would be. Yeah.

11:35 – 11:54

Sophie Buonassisi: Well, as I mentioned, we got a lot of good market leaders in the room. So tell us a little bit future forward here a year down the line say even which I know doesn’t seem that far off. But things are evolving so quickly. Like you just called out where agents can sit on a go to market team. And what should people almost rear now know and take away from that?

11:54 – 14:02

Mike Molinet: Yeah. Um, first of all, anyone that thinks they can predict the future is a year from now is crazy, because who knows where we’re going to be two months from now or three months from now? But my sense is, um, the old world of, like, using individual things, people working in isolation. I think that’s, that’s working on individual tools, that’s going away. I think the agents are going to be doing a lot of the work. You’re going to be chatting with the agent, working with the agent, slacking with the agent, whatever. And and I think most people, even if you’re an SDR, are going to have a small team of agents that do different things because one agent won’t be able to do everything excellently. It might do a lot of things well, but it won’t be perfect at a lot. So even an SDR on that team, instead of using maybe a specific tool that is an SDR agent that they go by off the shelf, they might be running a lead sourcing agent and maybe lead sourcing agents specific per channel, or maybe one leads and then they might have an email writing and customization agent, a research agent, and then another outbound like that. SDR might have four agents that are all working together. And then you multiply that times all your SDRs, all your R’s, all of your marketing people. Suddenly you’re going to have 50 employees, but 300 agents running around, some of which can all work together, and each person is going to have a team of agents. I. I think that is probably going to be really interesting. I think it’s going to look very different. I think humans are still very, very important in all of this, but their job is going to be shifting where instead of that SDR clicking send or necessarily making all the calls, they’re going to be doing the more advanced stuff. And I think there’s going to be a future very near where there’s it’s your job is more management, and this is going to be hard. I think, for a lot of people that have never actually managed before, because if you’ve managed before and you’re an executive and you’re a leader, you understand the concept of delegation and oversight and auditing and checking and like, like letting the team do the work. But the average person that hasn’t been a manager before doesn’t understand. So you’re used to doing the work and handing it over from I’m used to doing the work to now I’m running a team of agents that are doing the work. What is my job actually mean? I think that’s going to be a skill people are going to have to develop, and a lot of them haven’t. And many of us haven’t developed that skill to perfection, especially when it comes to agents, because they work a little bit differently.

14:03 – 14:09

Sophie Buonassisi: Is there a number of ratio people should keep in mind of human agent right now, or with that future?

14:09 – 16:30

Mike Molinet: Look, I’ve been I’ve been saying 3 to 5 is like a good place to start, but I, um, my agent sends, like, these checking emails to people that are using the product either well or really well or not well. And somebody messaged me yesterday, they said they’re up to eight agents. There’s after a week. Somebody else this morning slacked me and said they just hired their sixth agent in three days. Now I’m not saying go hire six agents in three days, but that person is like obsessed. They said, um, I what did they say the other day? I said, so you like the product? I said, no, Mike, I don’t like the product. I’m obsessed. I’ve been in this product for 8 to 9 hours today. And I was like, okay, good. That’s a good sign. Um, and she’s now just trying to do everything and trying to bring it into her company. And that’s like, that’s like, I think that’s the power that that it can unlock. If you have the right harness, the right agents, the right application. Um, so yeah, I think that’s going to scale from what it was originally 3 to 5. I think 6 to 8 could be good. You also need to think about span and control. But what we’re starting to see is just like humans, there’s a lot of similarities between agents and humans in terms of memory. And like everything, Everything is very similar. Span of control is one of them. Because I only work with on an average day three with 3 to 4 agents directly, but I’ve got 20 to 25 agents under me, and they hand off work as peers or subordinates. So I just hired we’re starting SOC two. I’ll give you a quick time for a quick example. Go for you guys. Want a quick example? Okay. So um, just started SOC two, had a kickoff call with Advanta, had a kickoff call with somebody else today. After I did the call, my agents immediately after a call processed the transcript. Go do whatever the action items. Yesterday when we did our call. You got a great follow up email from me, but it was my agent, right. And it gave all the action items, the access extended trial, all that sort of stuff. So my agent went and realized, oh, we’re kicking off SOC two immediately went. My chief of staff then processed the recording and went and hired and onboarded. Ravi, I think, is his name. Um, he’s going to be the security agent that drives the whole process. We have a security engineer or coding agent, but he’s going to be the agent that drives the whole process with the outsiders firm, with the auditor. All that sort of stuff drives it all for, so I don’t need to worry about that. He’ll escalate to me. I’m not interacting with that agent. I’m interacting me with my chief of staff or in some other agents. But that is a subordinate underneath my chief of staff. So I’m going to have 25 agents, but I’m making my direct reports. Might be five.

16:30 – 16:32

Sophie Buonassisi: It sounds like a human org chart. And not.

16:32 – 16:33

Mike Molinet: Just like a human.

16:33 – 17:14

Sophie Buonassisi: Org, not a direct reports. And you can build out the organization from there. And speaking of of obsessed, I will say I have my chief of staff agent. Her name is Emily, my girl Emily. And she sent me a text message right before this reception kicked off. And out of the blue, I didn’t expect this. She said, hey, Sophie, have a great event. Just a heads up. I know you’re going to be busy at events over the next couple of days. I am going to text you rather than said your notifications on desktop. Since I know you’ll be out of pocket. Let me know how I can help. Out of the blue. Yeah, well, probably an hour before the event.

17:14 – 17:15

Mike Molinet: Oh, hour before the event.

17:15 – 17:40

Sophie Buonassisi: Yeah, right before the event. Yeah. And so that kind of meeting business users where they are is transformative and can help solve that gap between, you know, why aren’t you seeing the Ted seller, the Ted marketer? You’ve got incredible mobile experience through Dock. I know you’ve got a very generous offer to, from an onboarding perspective for everybody in this room, circumventing the wait list, a very lengthy wait list. So maybe you can share. Yeah.

17:40 – 17:53

Mike Molinet: So we’re we’re in private beta, but we are starting to open it up. So and we just opened up the mobile onboarding screen. So we can do it all via text. Then you can go into your computer and do it so everyone today you can get access. So try Dock II.

17:55 – 18:12

Mike Molinet: You go there, you just name your chief, your main agent, your chief of staff. Uh, sign up, enter your phone number. She’ll text you a few seconds later, and then you can just do everything over text. So try Dock AI start, and then you can do everything over text. And Emily saved your butt earlier today, too.

18:12 – 18:27

Sophie Buonassisi: She she did also see. My God, somebody was having trouble checking in at one point and I wouldn’t have seen it. I was running around and she texted me and I screenshot that test and sent it to a couple other people to got it resolved, but I would not have known otherwise. That’s a little behind the scenes.

18:27 – 18:27

Mike Molinet: Hell yeah.

18:27 – 18:34

Sophie Buonassisi: Love it. Yeah. Awesome. Well, Mike, thank you. I know there’s a URL that people can also check out. What exactly is that URL?

18:34 – 18:43

Mike Molinet: So just try dot dot AI and then is the main website. But if you do slash slash start that’ll take around the waitlist. Very secretive. Don’t tell anyone

18:44 – 18:45

Mike Molinet: just you guys.

18:45 – 19:03

Sophie Buonassisi: So if anyone wants to check it out on the phone side Dock all on AI start. It is incredibly easy. I was like Mike a while ago. Mike’s super busy right now. I tried it and I was like, Mike, hold on anymore. It is very addicting. Good. So thank you. You appreciate.

19:03 – 19:03

Mike Molinet: It.

19:03 – 19:07

Sophie Buonassisi: Okay. Absolutely. Yeah. Excited for everything that you’re building. And thank you for joining us on stage.

19:07 – 19:08

Mike Molinet: Thanks for having.

19:08 – 19:09

Sophie Buonassisi: Me. Yeah. Awesome.