How AI Drones Could Replace Oil Field Pumpers / Unmanned Operator

Peter Brecht00:00

All right, welcome to week 33 of Local Energy. I’m Peter Brecht with my co-host Wade Spear. This is actually our third interview — with Travis Johnson here from Unmanned Operator. Excited to dig into it today. Travis, how are you doing?

Travis Johnson00:17

Pretty good, pretty good. Thanks for having me.

Peter Brecht00:20

Super stoked. We talked last week briefly about 3D models and all the fun stuff AI is doing, and you started telling me about your company. I said, man, we’ve really got to get you on and learn more about you and what you do. To kick things off — where are you located, and tell us the backstory of how you ended up starting your own company.

Travis Johnson00:42

Yeah, so we’re in Stanley, North Dakota. I was a pumper for ten-ish years prior to starting Unmanned Operator. We’re probably three years old now, I suppose. I know you guys have had the pumper-versus-lease-operator chat before — it’s called Unmanned Operator because of what I used to be.

Wade Spear01:00

Yeah, I was going to say — that was going to be my first question. Which one do you fall on there?

Travis Johnson01:11

I was a lease operator — pumper. I think pumpers call each other pumpers, and “lease operator” is more of a hiring term, in my experience.

Peter Brecht01:18

Mm-hmm.

Wade Spear01:20

That settles it.

Travis Johnson01:23

That’s just what I’ve seen. If it’s a “lease operator” posted out there, you know.

Wade Spear01:30

Run.

Travis Johnson01:32

Pumpers always talk about being the low end of the food chain — even lower than truckers sometimes — because we get the really bad jobs. Sometimes very easy, but we get thrown under the bus a lot.

Peter Brecht01:47

So you’re the guy who could answer our questions about glycol pumps, I’m guessing. If we had questions, you’ve probably had your fair share.

Wade Spear01:53

Did you work with dehys very much and all that?

Travis Johnson01:55

You’re talking about chem pumps, more or less, or what?

Wade Spear01:58

Well, more like dehys — using glycol to strip water out of your gas stream and all that. Dehys aren’t used everywhere; it’s not a super common method.

Peter Brecht02:14

Yeah — every pumper I’ve talked to seems to know everything about them. And I’m thinking, okay, we’re talking North Dakota, South Dakota, up in there — it gets a little chilly up there.

Travis Johnson02:23

No, I’ve never actually seen those.

Peter Brecht02:27

Yeah, okay. Very cool.

Travis Johnson02:28

Our gas seems to drop the liquids out pretty easily, so we end up with big giant scrubbers on the flare lines or wherever. It drops out quick, and we also have these big temperature fluctuations — some days you can get crazy amounts, and some days there’s almost nothing.

Wade Spear02:31

It’s different everywhere.

Travis Johnson02:58

But no, I don’t think I ever used a glycol pump. The pumps we used were like methanol, to keep the damn things open, but that was it. That’s all I can really remember. They might have them now — I don’t see them, because I mostly walk sites doing my job now, and I mostly just see these big, big scrubbers.

Wade Spear03:10

Right. Keep it from freezing up, right? Yep.

Peter Brecht03:31

Gotcha. How long were you a pumper for? Is that where you started — the first thing you did and the last thing you did?

Travis Johnson03:36

Yeah, I was a pumper for a couple of different companies, and then I was a contract pumper too. I started out in Williston, and I ended up losing my place to stay — that was in 2011, 2012, summer in there. From there, I wanted to move to a place where the Walmart wasn’t so crazy every day and there was a place to live. It’s settled down a lot since then.

Wade Spear04:16

I was going to say — was it as bad as what they said it was back then? When everything was really ramping up? Just no place to stay, tons of people coming in from everywhere.

Travis Johnson04:24

Yeah. My first experience in Williston was going to the Walmart and looking at the bread aisle — you guys know the bread aisle at Walmart, right? It’s huge. There was nothing but rye bread. The only thing left was rye bread.

Wade Spear04:37

Right.

Peter Brecht04:43

Gosh.

Travis Johnson04:45

And I go, okay, I guess we’re not having bread today. They wouldn’t even put stuff on shelves, because they didn’t have enough employees. They’d just drop it on crates in the middle of the floor. Yeah, it was nuts. It’s a lot different now — a lot more civilized, I guess.

I’ve been to places recently, though, in the last couple of years, that look like that did back then. Like your Carlsbad area — it’s kind of a mess out that way. They don’t have the housing or infrastructure for people. The roads are terrible, just like back then up here. There’s nowhere to stay — hotels are sometimes hard to get, and stuff like that. But here, now, it’s not that way anymore.

Wade Spear05:26

They can’t support the influx. Wow. Man.

Peter Brecht05:48

And you’re ramping up your business too, if I remember. You started this company a couple of years ago, and you’re getting into some of the most fascinating stuff when it comes to AI — just what we were talking about. Maybe share a little bit about Unmanned Operator and how it turned into what you’re doing now.

Travis Johnson06:05

Yeah. Over the last three years, we’ve been, more or less, a data acquisition and data processing company. We basically go out and fly a site — make a 2D/3D model, map the site — and sometimes we’d walk it with a 360 camera, so you could pretty much visit the site from wherever you were in the world if you wanted.

Now we’re really transitioning into using the data we already have to create AI systems for the next gen of oil and gas. Think about pumping sites with drones instead of people having to drive all that way every day. Companies are starting to try it, especially up here, and we’re building the software to support that.

Peter Brecht06:59

Okay. Go ahead, Wade — you’re good.

Wade Spear07:00

So how would that work? I can imagine you could use one drone to check a couple of different sites if they’re really close to each other. Would you literally put a drone on each pad, activate it, fly it around, have it check for leaks or heat spots or whatever, and then it goes back, redocks, and recharges itself for another check — every day? How does that whole system work?

Travis Johnson07:32

That’s what operators, I believe, are trying right now — that’s what they’re going to try first. We, as my company, are going to give them another option in probably the next six months. The Trump administration really opened up a lot for the drone community and what we can do. So what we’ll have in the next six months to really go after this kind of stuff isn’t just your regular drones — we’re talking 14-to-18-foot-wingspan fixed wings that can go nine to fifteen hours a day if they want.

What we’re building is what they’d call edge AI. Instead of the data going from the drone to a cloud, it’s actually on the drone, in real time. We train it to recognize things like leaks, and to go through your equipment groups — your treater building, your treater facilities. As soon as it sees a treater, it says, “Okay, I need to check temp, all these things, look for leaks.” If it’s good, it’s green. If it’s not, it’s red, and a pumper needs to come out. That type of thing, and it’s real time.

Wade Spear09:08

Right.

Travis Johnson09:08

And then we can stream that too, with equipment we have, with a couple-second delay, fully encrypted. It’s pretty cool what we can do now. We’ll apply that to pipelines too.

And then we have another piece of software that’s actually coming first — it’ll be out in a couple of weeks — for your smaller, Bakken-style pads. I don’t know how big the pads are down there all the time, but we usually go from four to eight acres, and the frac and drilling crews are always having headaches with the size. So we’re building an AI — it’s called PARA, and it’ll be out pretty quick — it’s a pad planning software. You can put all your equipment on the site — all your service guys, you can create all this equipment — and it goes on the site at scale, so you can plan more efficiently. We found out they’re still doing it on a notebook. We went to look for software and there wasn’t one, so we just decided last year, actually about this time, that we’d try to build one. I built about 90 percent of it and then hired some AI engineers to finish it up. The finished product will have more of an “optimize” button — you tell it all the equipment you want on the site, and it’ll place it for you. All your sand’s delivered in an efficient manner. Everything’s efficient.

Wade Spear11:11

That’s really interesting. I know that years and years ago, I’d heard of a company getting some of those Six Sigma firms to come out and evaluate things like rig moves, to try to optimize them so you’re not moving the same piece of equipment three times trying to get stuff around it. So using AI on the topographic map of your pad size and your location and everything — man, that would really accelerate that process.

Travis Johnson11:42

And it’s a drone map too, so you’re good to a few centimeters, instead of looking at Google Maps, which is like five feet. We’ve got centimeter-level accuracy, basically, so we can place things accurately. On the other side of it, we’re even going to build some devices that go on the sand trucks, so we can figure out how long a delivery is taking, or when that sand truck’s coming back — ETAs and stuff like that.

And on top of that, we’re looking at a safety AI on a drone-in-a-box with a tether, so no pilot is needed. It’s constantly watching the more restricted areas of the frac site. If somebody gets in there who’s not supposed to be, someone’s alerted. And we’ll build something around that too — we just haven’t decided what yet.

Peter Brecht12:49

Gotcha. How many terabytes of data are you collecting on every site? And is the opening scene of Interstellar just a major inspiration for you?

Travis Johnson13:01

I don’t remember Interstellar, to be very honest.

Peter Brecht13:04

The drone — it’s okay, it’s okay.

Wade Spear13:07

It’s amazing.

Travis Johnson13:07

Terabytes, I would say not — not a terabyte typically. But gigs, yes. Many gigs.

Peter Brecht13:21

And it’s all solid state? Maybe you’re transmitting it to the cloud, but on the drone, if it’s being collected on the drone — are you having to build these out custom, or are they pre-made somewhere?

Travis Johnson13:36

You’re talking about the ones in the next six months? Yeah — those will actually be built in-state, probably.

Peter Brecht13:49

Okay, very cool. So basically, you’re building jobs here in the country — in America.

Travis Johnson13:56

Oh yeah, everything will be 100 percent American-made. We’re done with that. I love DJI drones — love them, because they’re so nice to use — but we’re done with that. That ship sailed. Everything’s going to be American-made, because you won’t get the work otherwise. You won’t.

Peter Brecht14:19

Wade, have you ever flown a DJI before? Those little ones — they’re not that big, maybe eight inches square. Have you flown them?

Wade Spear14:25

Nah — always wanted to. I’ve always looked for an excuse to buy one, but I’ve just never pulled the trigger.

Peter Brecht14:34

They’re pretty cheap now. My son wants to do all the things under the sun, and they have one on Amazon for like 300 bucks. But I don’t know if I’d fly it enough — and I’m pretty sure he’ll crash it into somebody and cause a problem. But they’re fun to fly. They’re wonderful.

Wade Spear14:54

Oh yeah — I always looked at the ones with object avoidance and all that. The ones that could tell where the tree branches were.

Peter Brecht15:00

Yeah, I had one. We went to Puerto Rico about five years ago — my wife had a surf trip — and I got to fly it out over the ocean. You push the little locator and it flies back. It got so windy, though. I got some amazing shots, but it got so windy that the thing crashed in the sand and I couldn’t fix it, and I was bummed. I said I’d never do it again.

But where it’s come in five years — that’s really what blows me away. That was not an inexpensive drone, and that was a very painful mistake, but now AI has really come online within the last year, and there are these commercial applications where you’re actually mapping things three-dimensionally. And not only that — it’s no longer “we’re just generally mapping things,” it’s “now we’re specifically doing this for oil and gas.” There’s a safety component, there’s a data component. I can’t think of an insurance application I go through that doesn’t have a checkbox: do they fly drones, yes or no? And it’s cool, because if the company is flying its own drone, it’s a consideration — but if you’re coming in as an aid for them, it takes that off of them, and they’re getting a ton of information and help out of it too. I see a huge path forward here in so many different aspects — not only productivity, but from a risk standpoint too.

Travis Johnson16:19

Yeah. The real goal of going to autonomous pumpers is to take the guesswork out of when equipment’s going to break, right? You’re getting data on that daily, so you’re able to reference that daily data coming in and go, “Okay, there’s something wrong here,” or “Why are my weights this much different today on my pumping unit?” Or even diagnosing a stuffing box change — that’s in the realm of possibilities right now. How many threads have you got left?

Peter Brecht17:02

What did you just say — a stuffing box? What is that?

Travis Johnson17:07

The seals right at the wellhead where the rod goes in — there’s a seal, it’s called a stuffing box. What do you call it, Wade?

Wade Spear17:17

Oh yeah, that’s a stuffing box.

Travis Johnson17:19

So you could diagnose when that needs to be changed just by the threads left, on whatever type of system is on there. You could have it reference the manufacturer’s —

Wade Spear17:44

Specs or whatever.

Travis Johnson17:45

Specs, yeah — and say, “Okay, it’s time to change.” And you could have it literally say that in real time.

Peter Brecht17:53

Okay. Is this system doing gas leaks, flare monitoring, tank levels, stuff like that too, or no?

Travis Johnson18:01

Eventually, I would love to have that. Gas leaks — big deal. We haven’t gone after these big gas cameras because they’re so damn expensive. Hundreds of thousands of dollars on those things — just the camera.

Peter Brecht18:14

Hmm. Are they too heavy to fly on a drone?

Travis Johnson18:20

No, not at all — some companies do it. The other thing is, there have been hints that the Trump administration might be rolling back some of these things — Quad O, I think it’s called.

Wade Spear18:35

Mm-hmm. Because they use drones right now — don’t some of the state regulatory agencies use drones to go around and monitor flare stacks and locations, looking for leaks and all that?

Travis Johnson18:45

Absolutely. Yeah, they do. It’s just the cameras — you know how TVs come down in price? We haven’t seen that yet. They’ve been around for a while, and the tech has not gotten any better in the last three years. The price also hasn’t changed. So I’m not really sure where that’s going to go, or if someone else is going to come in and blow it out of the water.

There are lasers that can do it now, but it’s not as visual, I guess. They’ll take a laser and shoot it through a plume, and they can figure out what gases are in there. That can be flown on a drone as well. It’s an option, but the tech hasn’t gone very far in the last couple of years.

Peter Brecht19:40

I wonder if that’s a supply-and-demand thing — there just aren’t enough people who need that product. I certainly wouldn’t need one at my house compared to a regular camera. Maybe that’s part of it.

Travis Johnson19:52

Yeah. I think it’s also a thing with the companies who own the tech. It hasn’t really gotten out — the company that owns the internal sensor everybody builds around, nobody else has that tech, and nobody’s really gone after creating a better version, I guess. For example, Sierra-Olympic owns that tech, and then FLIR owns the handheld version on the ground. So you’ve got two companies, basically, and they don’t compete against each other because they own different sides of the tech, and nobody else has it. They’ve just been sitting on it, I guess.

Peter Brecht20:40

Is there an advantage — I mean, with a drone you can get it up in the air quickly, it’s quickly stowed for transport, probably a smaller footprint too — but what’s the benefit of putting it on a tethered drone versus putting it on a post or a pole? Is it just the maneuverability and having more flexibility in what you’re looking at?

Travis Johnson21:05

Putting it on a pole — that’s happening. On big emitter sites they’re worried about, they will do that. That’s 100 percent being done right now.

Peter Brecht21:17

On something stationary — like a production site, or a midstream refinery, et cetera.

Travis Johnson21:21

Yeah. And positioning those things is another thing — you’ve got to figure out your normal winds, all that. It’s got to get high enough. And honestly, I think the best way to do it right now is satellites.

Peter Brecht21:40

Really? That far up?

Travis Johnson21:41

Yeah — that’s how the government’s actually doing it, for the most part. The EPA. They have — I forget what it’s called — an emitter satellite. I’ve talked to a few operators, and they’re like, “I don’t ever want to be on that thing.” And I’m like, well, how do you avoid getting there? Because especially up here, our tank hatches — you could have them leaking like a sieve one day because it’s cold, and the next day it’s hot and they’re not. It’s a daily difference. You could change a perfectly good seal on those things and it would still leak sometimes. It’s one of those things that’s hard to fix in our environment. I don’t know where this whole thing is going to go.

Peter Brecht22:36

Yeah — that’s probably the fun part of your job, right? You’re on the cutting edge of everything. I think the first time we talked was the day ChatGPT-5 was released, and you were like, “Man, I’ve been in ChatGPT-5 all day long and doing incredible things with it.” And I’m just like — this guy. I logged into ChatGPT once, and it’s part of my workflow — I like to check stock prices with it because it’s easier than Googling. But you’re like, “No dude, I’m building stuff with it already.” That’s got to be exciting, but also maybe stressful. Wade, your job feels more manual, almost, whereas Travis, you’re kind of in the future already. If I wanted to ask somebody where this is going, you might have a better idea than me. And I don’t know — do you like ChatGPT-5? Do you think it was a big enough improvement?

Travis Johnson23:35

It was massive. Yeah.

Peter Brecht23:36

Really? Okay, what am I missing? I’m just curious, because I tried it and Claude still crushed it for my purposes.

Travis Johnson23:46

The Pro with Python — I think it’s maybe Max, or whatever it’s called — can really do some amazing things. For example, yesterday I did a presentation in town, kind of about this same idea. We have a thing in Minot called Startup Minot — and we’re kind of a startup, you could consider us, I guess. It’s an early morning thing; there’s coffee and donuts. So I actually had ChatGPT make me a piece of edge software — I think it spent a couple of hours on it — where I could plug my GoPro into my computer, have the GoPro looking at the crowd, and it detects coffee, donuts, and cell phones in real time. If it saw coffee or a donut, it said “coffee” and “donut.” If it saw a cell phone, it said, “Get off your phone.”

Peter Brecht24:45

That’s insane. Wow, that is impressive. I would pay 20 bucks just to see that work, you know what I mean? Just for fun. That’s cool.

Travis Johnson25:07

It was just a fun experiment. But that’s where you can go. We had a guy in town — one of my buddies, actually an old pumper with me, and he’s in a whole different industry now, I think he’s in banking — and he was looking for a good spot to drill a water well on his ranch land. I put in his information, and it started popping out dots on the map. I’m like, oh my God. And it’s giving reasons why — “drill here because of this and that.” It actually went and found county data and stuff, created a map, gave you your coordinates and everything. It’s nuts.

Peter Brecht25:54

Yeah. Wade sent me — we were talking about files — he’s like, “Hey, this is kind of what a well would look like if you drilled one. Here’s some data sets, theoretically.” So I extrapolated: I took his list and said, okay, let’s run this out another hundred rows. I had 190-something data points, put it in to model a well, dropped it into Claude, had it create the 3D model, converted it into a Blender file, put it in Blender, and it populated. That stuff’s cool. For me, as someone who wants to educate people on various aspects of oil and gas, I think that’s a really cool factor. Wade, I haven’t heard back from you on whether it was correct or not! But visually, it’s impressive. And I’m almost scared to find out where things go in five years, because it’s rapidly increasing.

AI is such a big topic — what are your clients looking for when they say, “We heard of Unmanned Operator, we want to bring you in”? What are some of the first conversations you’re having, or why would they be calling you to begin with?

Travis Johnson27:08

Well, mostly they’ve been calling for the service side. Most of my clients don’t know anything about what we’re doing — not yet — because we’ve been building stuff in the background, kind of stealth mode. We’re really coming out of that now; we’re going to go after funding for it, and we’ve got some pretty good traction on that going forward.

Most of my clients on the service side call me for LiDAR, if you guys are familiar with that — making point clouds for engineers, essentially. Doing a retrofit on a site, or whatever they want it for. Sometimes flares — adding something into a flare line: where’s the best spot, how can we do a drag-and-drop? The engineers don’t like to go out in the field anymore, which is great for me — the design guys, at least. We can make a point cloud, send it to them, they drop it into their CAD, and all of a sudden they can just add something in.

Peter Brecht28:33

Hmm. Wade, do you see an advantage with this stuff for what you’re doing? I’m sure there are numerous applications, but I know you’re still going out to drill sites actively — your pictures are like, “This is what I’m doing today,” and it’s completely different from my life. Do you see applications where this could enhance what you’re doing on a new drill?

Wade Spear28:58

Yeah — in the rig move, rig up, rig down, I definitely see that it could. Most of what we actually do when we’re drilling is subsurface, so I think that’s probably outside the scope of some of this. But if we’re talking about trying to make more efficient rig moves — or you even mentioned earlier about being able to track the sand trucks — for us, one of the challenging aspects is casing deliveries. You’ve got 12 trucks rolling in full of casing, and there’s just not room on location for those. Being able to stage them and efficiently get them in and out — normally you’d hope you’d have somebody there to direct those people in and out, but I’ll be honest with you, oftentimes that’s a challenge.

So I could see it helping on the margins on the drilling side. Ninety, ninety-five percent of what we do is subsurface, so I don’t know that it falls in there. But midstream, production — like you’re talking about on the pumper side — one of the things that immediately hit me, and Peter, I think this pulls in where you come from too: you’re starting to reduce the number of miles driven, which is probably one of the highest-risk activities you have in terms of coverage and all that. I could definitely see it reducing the need for some of those miles and helping on that side.

Travis Johnson30:50

Yeah, that’s kind of the whole goal. Especially since, in North Dakota, we have about 700,000 people, and we consistently hang at that 2 percent unemployment rate — which is essentially zero, because the people who aren’t employed don’t want a job. If we keep drilling at this rate, we’re not going to have enough people eventually. When I started, we used to cover 20, 25 wells. When I was done, it was like 60.

Wade Spear31:30

Right. And they’re not as closely spaced, either. Some pumpers’ routes are hundreds of miles at times.

Travis Johnson31:37

Yeah, they can be, for sure. You get in the truck at six in the morning and you don’t get home till six, most of the time.

Wade Spear31:45

Right. And the truck never turns off the entire day.

Travis Johnson31:49

Yeah — well, some companies make you now. God, that was the whole thing right when I was getting out of it: they didn’t want idle time, they didn’t want this or that. And it was like — if it’s 20 below, I’m not turning it off. I don’t care if I idle.

Wade Spear32:08

Not going to risk it, right.

Peter Brecht32:12

So Travis, I have a question for you — which is going to cause a lot of comments, I think, because guaranteed you’re going to get it wrong even if you’re spot on. When you park your truck as a pumper and you turn it off and get out — what are you doing on a well site? Give us a quick overview of what you’re doing on site, and then we can follow it up with how Unmanned Operator is relieving that specifically.

Travis Johnson32:41

So you’re parking — it depends on the site — whatever equipment’s first, you’re hitting that first. Going around in a circle: pretty much your treater, probably an MCC building now, because mostly everything is automated. Sometimes you’ll have an iPad with all your production on it. If production’s off for the day, you’re going to go try to figure out what’s going on first, probably. Walking your site — that doesn’t happen every day, for sure — but essentially just making sure everything’s working. You’re getting eyes on everything, getting in the treaters.

I don’t think companies work their dumps much anymore, because they’re a lot more automated now. If it’s an old site, you’re probably working dumps. Up here, we have freshwater injection, because we have real heavy water — so if that freshwater goes down, your dumps can get pretty sticky and messed up. Essentially, you’re walking the site, and then hitting the wellhead — it depends on how that’s set up as well, but there’s a lube box on there, most likely. I don’t know how they do it down south — I see videos all the time from down south and there’s no lube box or anything, and I’m like, how do these guys do this? Grease? Old-school grease or what? That’s a tough answer, I guess, because it’s site-specific.

Peter Brecht34:24

Okay. Are you ever grabbing tools and turning wrenches, or no?

Travis Johnson34:31

Oh yeah, yeah. You fixed everything you could do yourself. If you didn’t know how to do something, you’d call another pumper and have them help you — and then the next time, you could do it yourself. You’re pretty self-reliant.

Peter Brecht34:45

Got it. Okay — so it’s a job where you’re learning.

Wade Spear34:46

So let me ask — on a decent route, an average route — like you said, it’s extremely variable, but how many locations in a day are you normally going to where you’re having to turn valves or run dumps or whatever? Because my real question — you hit on it a little earlier — is that by using Unmanned Operator’s service, you can figure out, “Okay, everything’s good,” or “Hey, we need somebody to come out here and do this.” What percentage of wells do you think would need a pumper to actually go out and intervene?

Travis Johnson35:28

It’s going to depend on how well that pumper takes care of everything. If they’re checking things pretty regularly, you’re probably going to keep stuff pretty smooth. It seems like things happen in waves — if one thing goes wrong, there’s probably going to be a lot more that day. That’s just how it worked.

But as far as flying them — with all the automation that’s happening, there are very few manual processes left other than fixing stuff. A SCADA system can tell you if a treater’s having issues — most of them now have real-time data and numbers coming through. You don’t really need to manually work valves anymore. Basically, you’re walking in there, checking if things are covered or not, and looking for a leak — a gas leak, a quarter-inch line that’s leaking. Something like that is probably what a drone would miss versus a person. But there’s not much anymore, just because of automation. And that’s what makes it a really good time to get into this.

The other side of using a drone is the amount of data collected. You can actually start to be proactive instead of reactive on a lot of things, because you have consistent data on that site, and you’re able to see: there’s something wrong here. Or there’s a belt on the ground — part of a belt on the ground — maybe we’ve got to get out there right now. Pumpers miss that stuff. There are so many things a pumper does in a day, and they typically take the same exact route every day, and drive the same way on the site every day. It builds complacency a little bit. And the thing about data is that you’re constantly collecting it — eventually it will spot more than a pumper can, 100 percent.

Peter Brecht38:09

Would you imagine using this tool for the whole lifespan of the well, or more for the first year, when you have heavier production? Do you think it’s important for the full 30 years?

Travis Johnson38:29

Right away, it’ll be the heavy producers they’ll want to use them on, because you might be flying a couple of times a day — and basically you’re using that to free up your pumpers in the field to go do other stuff. Keeping an eye on things, right? Eventually, it’ll go to everything.

I don’t know how everybody does it, but a lot of companies have gone to either big pads, or satellite locations that all flow into a central tank battery — you’ll have a well out here and it goes into a central tank battery somewhere else. So it’s a quick flight around, making sure you’re not leaking, you’re not making a mess, you’re not pumping backwards, the vibe switch didn’t fail — something goofy like that. Just eyes on. And that’s all you do anyway, for the most part, other than filling lube boxes on the polished rod.

And then also checking your mechanical gauges versus what your controller sees. The controller sees the Rosemount or whatever electronic gauges they have — we can see the mechanical gauge. With AI, you look at a picture sometimes and can’t make out what something is. Well, AI can go way deeper than a pixel and figure it out, especially when you train it to.

Peter Brecht40:14

Yeah. Awesome. I’ve learned so much from the oil and gas industry — some of the most brilliant people are working in it, trying to make things better and improve it, and I’m constantly encouraged. Wade, do you have any questions as we get toward our 45-minute mark?

Wade Spear40:36

I guess my question is: what do you see as the primary area for growth for your operations at Unmanned Operator? Where are the places —

Travis Johnson40:46

Probably pipeline, right away.

Peter Brecht40:51

I’m sorry, what was the answer on that? I want to make sure we got that one.

Travis Johnson40:53

On the pipelines. Pipelines, 100 percent.

Peter Brecht40:58

Okay, all right — pipelines.

Travis Johnson40:59

Yeah. They’re typically flown in operations right now, and they take very little data — if they take any. And that’s a big thing. The saltwater pipelines are the big ones. When they break —

Wade Spear41:18

More prone to failure, due to the saltwater content and all that. Or maybe volumes, too.

Travis Johnson41:27

Yeah — and they’re more of an ecological mess. When they spill, if it hits the surface, it’s a million dollars. And if it’s big — if you make a puddle — it just gets exponentially worse every time. We get one every year up here. Whatever that saltwater — brine, whatever you want to call it — touches, it essentially sterilizes. That dirt isn’t usable for many, many, many years.

So we want to start taking more data on these things and figuring out if we can predict failures more regularly — if there are cave-ins. And it’s not steel pipe, either. A lot of the pipe in the ground up in North Dakota is old poly or a fiberglass-type mixture — more fiberglass — and it is just terrible. Identifying things quicker is basically the name of the game, and we can do it at the same price or cheaper.

Peter Brecht42:53

Gotcha. Well, I want to keep talking with you, for sure, on the risk side of it — I think that’s a huge part. In my head, I’m thinking through the possibilities of how that would work, and for everyone who’d have this product, it’s almost like an endorsement of safety, in my opinion. Any data we can get is huge. I look at it as a huge opportunity as well. I’m not in your business, but from the outside in, that’s a big need from a risk standpoint — can I make this less expensive to take care of?

And the other fun part for me is the education side. If you’ve got cool videos and 3D stuff where you can get a high schooler to look at something like that and play around with it on their computer and learn about it — I think that’s huge. I don’t think that’s going away.

Travis Johnson43:54

Yeah. I mean, if you can take a pumper and instead of having to drive 200 miles a day, they drive 100 miles a day — there’s probably a significant insurance deduction. Less insurance.

Peter Brecht44:02

Huge. Massive. Auto is horrible across the board — my personal auto, and I’m sure you guys’ as well. And then you go, “Let’s go do this for 300 trucks,” and you’re like, oh my gosh, what does that look like? It’s rough. So it’s cool to see. Very excited.

Well, I want to be respectful of everyone’s time — I know we’ve all got other things to jump to. Travis, where can we find you? Where can people learn more about you? Give us your shameless plug.

Travis Johnson44:34

Unmanned Operator — it’s unmanned-operator.com. You can find me on LinkedIn. You’ll start to see a lot of videos pop out soon. Like I said, we’ve been kind of in the dark about all this stuff, but once our first software is out, it’ll be pretty quick — probably every-other-day content, I’m thinking. We just haven’t made anything yet because it’s not quite done — it’s not how I want it. But the frac planning — the site planning software — is going to be out real soon.

Peter Brecht45:16

Very cool. Well, we’re excited to keep in touch and learn more as you keep building this out. Go give Travis a follow on LinkedIn and check out his website. And obviously, I think you’re on TikTok as well somewhere — I think that’s how we found each other.

Travis Johnson45:30

Yeah, yeah — we don’t do a lot on there. That’s more for fun.

Peter Brecht45:36

Yeah, well — we didn’t expect to be on TikTok either, but here we are. Anyways, great. Thank you guys for being here. That will conclude this week’s episode, and we’ll see you next week. Thanks.

Travis Johnson45:46

Thanks for having me.

Wade Spear45:47

Good to see you, Travis. Thanks, guys.