Subscribe
Share
Share
Embed
Welcome to Energy Bytes with John Kalfayan and Bobby Neelon. Your essential guide to the intersection of data and energy. This podcast dives deep into the world of energy, shedding light on how data, AI, and technology are revolutionizing this sector. Each episode equips listeners with insights into the most efficient tools and resources, paving the way for a data-driven future in energy. From technical nuances to broader industry trends, Energy Bytes offers an unparalleled perspective on the evolution of the energy industry. Join us as we decode the algorithms of energy, one byte at a time.
0:00 Welcome back to everyone to another exhilarating episode of energy by some super super excited about our guest today.
0:10 First of all, we've got my my trust were the co host Bobby Neland coming to us live from the Southwest from
0:21 the Southwest side. And today's guest is the owner of Caso land and imaging Scott Caso Scott. Thanks for joining us, man. Thanks for having me guys. Always appreciate chatting with you guys. I
0:34 look forward I look forward to it, actually, you know, there's some people in this industry feel like you got to kind of like, Oh, man, I gotta go talk to this guy, you know, I enjoy chatting
0:44 with you guys, actually, I appreciate it. Likewise. Yeah, I'm an interest.
0:51 Always helps for sure. They, uh, Yes, we all know those people who you're like, if I have a conversation, it's gonna turn into an hour long thing, which can be good or bad, depending on what
1:03 you're trying to do. But no, for sure, man, I'm excited to chat with you and have you guys on. Especially now that we've kind of worked together and a little bit of a capacity makes it, makes it
1:15 a lot easier
1:17 to contextualize everything But kind of tell everyone who you are, what you guys do, and we'll kind of work backwards, I think. Cool. Of this one. Yeah. So Scott Costo, I own Costo Land
1:31 Imaging Services. We're primarily a land service companies. We'll help landmen gather data from courthouses so they can run title reports and do those types of
1:46 reporting that they need to do
1:50 We go out to courthouses, we take photos of houses, we take photos of. these land documents. We organize them and send them off. An old land man used to drive out to these courthouses, sit there
2:00 and flip page by page, and chain out their title, standing there in the courthouse. And with technology, we've been able to expedite that process. That's kind of our primary business. It's how
2:11 we started in the industry was helping land men get the data that they need from these remote areas. A lot of stuff in Oklahoma, that's where I'm from. Oklahoma does not have a lot of data online
2:23 You guys are down in Texas. Kind of opposite scenario. Down there, you guys have a ton of access to online records from courthouses, all of that type of stuff that you would need to run your
2:34 reports with, prove ownership, things like that is all readily available in Oklahoma. It's not, there's a number of other states that it's not, and we operate in those states as well. The other
2:44 side of our business is the side that, John, you and I have worked together on, that's our scanning side something that's really been growing. our business and in our business. It's been really
2:57 cool to see people like the operators that we've been working with, really want to harness and leverage the AI. And to do that, they have to have their data digitized. You can't apply AI to a
3:13 warehouse full of bankers boxes. You've got to digitize that stuff and have it available then to go into the system. So that's been really the push lately for me has been the scanning side of stuff.
3:26 I see that as the future, I think shortly Oklahoma will be like these other states that has everything online. And that'll kind of shut down that side of our business. And so preparing for what the
3:40 future has is driving the scanning division that I've been developing the last few years. Started with some work with the Archdiocese in Oklahoma City. We were managing some of their just paper
3:52 legal files. planted a seed in me to go out and start finding people either in oil and gas or outside of oil and gas that have just paper files that need to be digitized. It seems like such an easy
4:05 process. And it is for a technician to take a document and run it through a scanner, it doesn't take a super high level of competency to do that. But we realized that there was a model here that we
4:20 could present that wasn't just your nephew or steps on or someone else coming in the office during their summer break or something like that and coming in and running a scanner in their office. There
4:34 was a higher level that was actually needed and that's at least understanding how these documents affect each other, what these types of documents are. The same type of things that I think you guys
4:43 are working on the back end. How do you classify these documents? How do you organize them? And it's so cool seeing what AI is doing to help our process. you know, our process is so heavily
4:55 manual still today and seeing what some of the, you know, even basic things like stopping having to manually index things. You know, that's a really huge thing that we're seeing on our side that's
5:07 like, Oh my God, this is going to take projects that you should take a month to get done. We're going to get that done, you know, in two, three weeks. You know, it's something that, you know,
5:15 real simple things like that are changing, you know, what we're seeing. You know, so that's what we do. We take paper and we digitize it, whether it's for a landman or an operator or a bureau of
5:30 prisons or a Catholic church or whoever you are, we like touching your paper and dealing with it and, you know, getting it in into whatever system you guys need. You know, years and years ago,
5:41 it was just a hard drive and now it's all of these cool collide systems, you know? For sure. No, I think you bring up a really good point that I'm going to probably unpack a little bit here,
5:51 which is one. Pretty much every operator I've ever been to has at least one probably dozens to hundreds of bankers boxes full of files that are, you know, well related that most of them have no
6:07 idea what's actually in them, right, they know that this box is for this asset, but that's about it. And so like the first project that we did, you know, I don't even remember how many boxes
6:18 they sent over to you guys, but
6:21 yeah, there were quite a few boxes, there were like 3, 000 documents almost, I think it was 75, 000 pages plus, I believe. And so it's, you know, to me, I honestly look at it as like tech
6:37 debt, right? It's like as an engineer, I always want as much data as I possibly can, you know, can have, which is great But then that ends up getting us in these positions of like, okay, well,
6:48 we kept all this data, these files on these wells. but they're just sitting in boxes. So it's literally useless dark data that we literally have no clue what is actually in, in those boxes. And
7:02 so, but, you know, that's one piece of this, which I think is a big piece because there's a lot of, you know, really interesting information about that data or there's a lot of interesting
7:15 learnings, so to speak, to come from this kind of dark data that everyone has sitting in their offices or in filing cabinets and stuff And so,
7:24 you know, I think the struggle historically has been, okay, we digitize it, so what? We still don't know what's in there. We still, you know, it's still just files. And so the cool part, you
7:35 know, the cool thing that we're working with y'all on is not only taking those files and digitizing them, but then digitizing them and then putting them into a language model that you can actually,
7:45 basically have a conversation with those documents, find the important things those documents or even just discover. all the documents that you didn't know that you already had. And so there's a
7:56 lot of really important things there because that problem's not going away, right? Like the longer they sit there, the harder they probably end up being to read. And then, you know, there's the
8:07 whole digitization effort and then there's the whole digitization to the language model or, you know, to truly digital effort where you can actually do something with the data and the documents and
8:18 stuff. And so I think this is a big part of it But I think that's also something just to kind of interject there. I think that's also something that's interesting that you mentioned, you know,
8:27 it's not just about getting the data into the system but having the system be able to function, you know, and I think that's maybe where some of the disconnect is right now where AI is such this
8:40 overarching term that, you know, you talk to a landman, they're like, okay, well, I want an AI system where I just drop this in and it tells me this, and it's like, Well, I mean, that's
8:50 possible, but. you've gotta develop everything along the way to get there. And there's gonna be trial and error. There's not just something that I see out the box today that you just dump in and
9:01 it does exactly what you need. You've gotta have conversations with real humans and say these are the environments we want built out. This is the type of information we're wanting pulled from these,
9:12 and it used to be draw a box over this area and extract this type of information from it. And that's changing and it's really cool to see I was listening to something the other day. I think it was a
9:23 video I found on Reddit about the dude that created the QR code, which I didn't realize. I think I was watching one of your guys podcast a while back and you're asking people, like, what does LLM
9:36 mean? Or what does GPT stand for? QR was one that I was like, oh, I use that term all the time, but I didn't know what QR stood for. I believe it's quick read I'm not sure if you guys are
9:50 familiar with that or just know it as a QR code, but it was a transition from barcodes that were like single-type deals to these QR codes that have faster reading capabilities and you can put more
10:07 dense data into them and things like that. So it's that development of these things that are existing No, absolutely. I think the most impressive thing about working with you guys so far has just
10:22 been not only did we get the data back, but it's organized, it's structured, it's labeled correctly, which makes our job infinitely easier on the back end because I can start classifying and
10:35 categorizing right out of the gate instead of having to identify each document, then classify and categorize them. I know
10:46 that sounds very simple to the person listening to this, but that is a tremendous amount of work. that is ultimately required by somebody along the line as you're doing this process. And so I just
10:59 wanna point that out that like, yes, you can have your nephew or your cousin or your grandson scan documents, but there's two things there. One, you're not gonna get the quality of document or
11:11 scan that Scott's gonna offer. But most importantly, you're not gonna be, you're still gonna be in the same exact spot that you ended up with where it's like, these are the files about as well.
11:22 Unstructured data, it's a bank, it's a digital bank, Exactly. It's still sitting there, and while that, I think still has utility to some folks, you know? I think for others, you know,
11:35 that's not the solution. It's just, you know, compiling. But, you know, some of the people that we've been working with with you, another thing, you know, another component of this is, Those
11:46 people are still and continuously generating more paper. some of their drawings, they said that they will never stop doing by hand because that's how they prefer to do some of the drawings. And
11:56 that's fine. That's not something that needs to be changed, but it's also something that needs to be understood as a reality here that like, this paper is not just going away. We can scan
12:05 everything from 1960 up until today. You know, if your business started then, but you're still going to be continuing this. And so what's the plan down the line, you know? For sure. Yeah. And
12:16 I think some of this, you guys are saying simple, and I'm one of my favorite sayings like simple, but not easy.
12:23 So I mean, again, it was easy. There'd be a lot more people out there doing it. But also, let's maybe step back real quick too, 'cause we've said digitize, and I think you guys are starting to
12:33 hit on it or if hit on a little bit, but like, can we talk about what digitizing it means? 'Cause again, I think some people, it may just mean, you know, going from a paper file to, I mean,
12:42 are these going into PDFs? Are they going into, you know, JPEGs, PNGs or so on? You know, so what does digitization mean? Yeah, a number of different formats. So, it's kind of up to clients
12:53 discretion. It's up to people like John letting me know how Collide System is going to ingest this best. And then even document by document, you got well logs that a lot of times you're scanning in
13:05 as like a TIFF document. It's more compact.
13:11 And so it is, it's taking a document and running it through a scanner. Now we just so happen to have high volume scanners We've got these large plotter scanners, but at the end of the day, the
13:23 same mentality is, I mean, you could buy one of those desktop brother scanners and be doing the same type of thing. It wouldn't be at the speed we're doing it, but you're taking a document,
13:32 scanning it into this digital form. And then from there, that could be the end of calling it digitized, you know? But I think, and John correct me if I'm wrong, but I think you and I look at
13:44 something digitized as like, hey, we've got search capabilities. got, you know, and now it's, well, once it's reached its final resting place in these, you know, systems, well, can I ask it
13:56 questions? Can I, you know, have it create a form for me based off of, you know, how all of these other documents were, you know, originally created, you know, and so that's it. I think it's
14:08 a changing term, you know, 10 years ago is just getting it into from paper to showing up on your screen, you know, but now there's so much more, you know, even people that were just scanning in
14:20 like invoices, they want to be able to search these, you know, different components of these documents and not just an invoice number. They want to see, okay, how many invoices did we have
14:30 under10, 000 this year? You know, let's apply that to every invoice that's in the system, you know, and that's more than just, you know, OCRing a document, you know. Yeah, for sure. I mean,
14:43 even just thinking about it now, like, I think probably everyone's expectations of digitized means is different too, because like I can go on my phone right now in search of term. It's going to
14:51 search through my photos also and my screenshots and find those kind of things where it's like, again, that that wasn't the case even two or three years ago. But now I mean, and I'm sure I'm sure
15:01 Android probably has it too. Google's even probably further ahead than Apple, because Apple's AI is awful, but I digress.
15:11 But
15:13 so one thing, and I think maybe you and I talk about it, Scott, but I mean, obviously we want everyone to hear about this. But like, where does this process fit, say, in like the emergency and
15:23 acquisitions workflows? I mean, like, I mean, it could be after, you know, an acquisition that happens in like, again, I'm sure some of these bankers boxes have changed hands four or five
15:33 times and never been opened. Just, you know, just got, you know, got passed around, but then also like data rooms. I mean, I'm just thinking there's a lot of time. Yeah, I've got an
15:42 interesting, operator that we were working with that one day a semi truck showed up and had like 800 bankers boxes in it and they were like oh we didn't know this was arriving today and it was from a
15:54 past acquisition that you know the data was being transferred and it all showed up and a lot of it was really important data. So you unload about 800 bankers boxes into this client's warehouse and
16:06 you're walking around in there and you see the different field's these came out of you see the different you know operators that had these before it ended up with this you know and you go in and let's
16:18 say we've got four different styles of boxes you know you've got one
16:27 that everything is organized by legal description you go in the other one and it's all just by land files you go into another one it's all well files you know just these variations that are really
16:34 crazy even in between companies and so you know the first kind of objective is going in and. organizing and indexing these things so you know what is there. That's kind of the first thing. You pop
16:47 the top of these things and start looking through is there any organization to it at all? A lot of times that answer is no.
16:57 Yeah, what was the original question? I think just like where does this kind of - Oh, yeah. Is there some - In your MA. In positions and things like that, yeah. So a lot of times we're catching
17:07 these on the backside So it'll come in, after an acquisition's been made, all these boxes have been dumped and they're like, oh my God, we didn't realize all of this stuff is digital or wasn't
17:17 digital, you know? And so then we go in, we kind of inventory it. We realize it's all still super unorganized. Then we go in and we start the digitization process. We think it's a great value ad
17:29 that we're bringing to these people by digitizing it and helping them understand what's in these boxes, organizing it, finding similar homes all these documents to live in, the environments for
17:39 them to live in. This is a very manual process again. What is needed is something like collide or document management system that can be deployed on these to bring that to the next level. We are
17:51 not that technology. We are not that person. We like deploying into these other areas and we are the labor force behind this. But what we've also been trying to push is companies selling, people
18:04 that are selling their assets and things like that and being able to digitize them and hand those off as, you know, we think that that's another value add, you know, hey, not only are we selling
18:14 this to you, but all of our data for this is already digital. Here you go, take it from here, you know. And so, pre-sale, we would like to be a part of that as well. And we're seeing some
18:24 traction there. It's a, you know, mixed bag on both sides. Some people don't care about it. Some people want it digital when it's going out the door. Some people say leave it up to those guys.
18:36 It's their problem, you know but I think there's a lot of value. in having the stuff digital, whether it's yours going out the door, you're receiving it. Like John said, you don't ever know
18:46 what's in there. Is there something of value? Is 80 of it not valuable, possibly, but you don't know what value that 20 in there is going to have, you know? Yeah, and I guess it could be an
18:58 exposure to you. I mean, where like, if you've got a digital and then, now, I mean,
19:06 I should know more about this, but like say for like a data room when you're trying to sell, like, is, they choose what they put in there to make available, right? Or I mean, or can that
19:15 company that's wanting to buy a request up in this? 'Cause there could be information in there that the purchaser could find interesting that maybe the company that's selling doesn't even know about
19:25 themselves or know about the asset in that. So just kind of curious, your thoughts there, John Slots.
19:34 Yeah, John, you wanna answer that one? Well, I mean, I think the, the overarching thing there. is, and Bobby, we kind of saw this when we were working together, right? Like the, I don't
19:44 remember, I think it was time or Forbes, whoever put out the article about data is the new oil, right? Like five, seven years ago, something like that. And I absolutely believe that even in the
19:57 oil and gas business. And so when you look at non-digitized files, it's like, to me, that is valuable data that is not being included in the price of the asset because there's a cost associated
20:13 with getting it to a usable form. And so it's like, if you already had it in the usable form, that obviously makes it a lot easier for your buyer to use the data to evaluate it, so on and so forth.
20:24 And so I think that's kind of the big aspect of this is just like, when you're making hundreds of millions of or billion dollar decisions, Like, of course you're gonna want as much data as you can
20:37 possibly have to. use that data to make an informed decision. And so I think that's really the big thing about this, is like, you know, if you're selling, in my mind, it makes it more valuable
20:48 because you have everything there. There's no, you're de-risking it by having the data, right? Like that's ultimately what it boils down to. And of course, as an industry, the first thing
20:58 everybody wants to think about is the risk. And so when you're making these big decisions, I think, you know, you're not gonna be scolded by management for, you know, having, using all of the
21:11 data available to inform your decision. Right. But I don't have any - John, on the, on the land side, you know, every land man that I talk to that is, you know, interested in digitization,
21:26 interested in AI, those types of things, you know, their big request is, you know, can I have AI run title for me? Can I have it, you know, chain this out, even to a point where like, I've
21:37 got to go and check a few. you know, but can it build my run sheet? Can it, you know, find the ownership chain, you know, those types of things? That's obviously the big ask in, on the land
21:48 side. I think we're heading that direction. I think things are getting close. I've been seeing a lot of really cool things coming out of that. On the operator side, what are you seeing as some of
21:60 the like most important types of documents that are extracting data from? Or maybe just kind of expand on what is their wish list when they're coming to these things. Is it things like accounting?
22:14 Is it things like, what is that? Yeah, no, that's a good question. So the main stuff we're seeing is one, logs, right? Just out of the gate, that's great information that the stuff that we've
22:30 been working on. There's logs in there from the '70s to '80s, the '90s and it's like there is not a digital version of that. that image and that PDF is the digital version of it. And so it's like,
22:41 you're dealing with huge assets. And so having those logs to understand what the hell you're working in and around is obviously incredibly important. The isopax and the maps along with that are
22:53 another one that right there with it.
22:57 Another big one is just like the well history, right? Like a lot of production engineers get a lot of flack, but one of the big things that we're really aiming to solve with collide is giving the
23:08 operator the ability to come in and say, Give me a history of this well, and it will go through all of the files that are associated with that well and give a timeline by date of all the activity
23:19 that has happened on these wells, because something happens, production falls off, so then the production engineer is tasked with fixing it, and so he's like, Well, shit, I just inherited this
23:29 well, or, We just bought this acquisition. I have no idea what has happened to this well. What do I do? So most of them spend, you know. days to weeks, digging through well view or whatever,
23:40 trying to find all the different files, reading through everything, trying to put together their own timeline. And it's like, well, if you've already got the data, then you just need something
23:49 to go through and organize it, structure it and put it together. And that's one of the things that we're really focused on with with collide. And we think language models are one of the keys to
23:59 doing that. And so, you know, people
24:03 don't really make the correlation a lot that language models are just ML models for like, what we, the reason machine learning models came about is because we had so much machine numeric, you know,
24:15 IOT sensor data that we as humans couldn't process it. So we needed a machine to do it. It's the exact same thing with paper documents or digital documents. It's like, we've got all these paper
24:25 documents and it would take a human days to weeks to read through the every single word of them. And he probably only needs 10 to 20 of the information that he's reading or they're reading through,
24:36 whereas. language model can come in and just extract exactly what you're looking for when you ask for it on demand. And so those are, in my experience, a lot of the big ones, right? Just like,
24:47 what is in my well bore? Like, I don't even know, because we don't have the original survey, because, you know, the well is drilled in the 80s, and it's in a filing cabinet somewhere. I think
24:57 of things like training manuals also as being something that like, you know, one, these people can reference back to some of the things in the training manuals, but it also expedites that process
25:08 of onboarding someone, training them, giving them the knowledge that they need. You know, it's funny when you were talking about, you know, and maybe a new engineer coming on, trying to fix
25:16 this problem that he inherited. In my head, I don't know why it went to this, but I was like, oh, this will just make it easier to fire people too. You know, you can get them off and then have
25:26 someone else come in and they can learn really, really quickly, you know. So, it's conflict resolution getting to that, you know, well, we don't want to fire this guy because because he knows
25:33 everything. Now this irreplaceable person is replaceable. I know it's maybe a grim outlook on it, but that's where my head went to some degree is like these irreplaceable people that are maybe
25:48 bottlenecking things can be replaced by someone that just needs to inherit that knowledge and not just the shitty well. Yeah, no, the big thing there to me is it democratizes data, right? Whether
25:60 it's just paper data or text data or even once we've got a text to SQL service that we use in house for, you know, Julie and the growth and marketing team so that they don't have to ask me to write
26:14 SQL queries anytime they're looking for specific data about collide users, they can just query it. And so it's like imagine that in the hands of accounting or legal or, you know, your
26:26 non-engineering related teams but giving them the ability to go get the one quick answer continue to do their job instead of having to put in a ticket or go hunt down the engineer to have them find it.
26:40 There's just so much inefficiency around that stuff.
26:44 I think in the near term, AI is not replacing as many people as it is just empowering and getting rid of the things they already hate doing, like regulatory filings or generating the same report or
26:58 procedure or policy or whatever. And then, yeah, in the long term, it's going to be fascinating to kind of see how much human guidance ends up being required on certain things versus others. But
27:12 there's still lots of engineering and things that we need humans for in the field. But the language models, in my opinion, in my experience, at least, just really help streamline getting your job
27:26 done, finding the answer,
27:29 completing the form, the, the procedure, the policy, whatever it may be, um, you know, they're not perfect, but getting you 80 of the way they're in, you know, a matter of minutes is much,
27:42 much faster than hours and hours and hours or days. Exactly.
27:48 Um, I guess same question, Bobby, what is, what do you feel like is the greatest value to being able to search data the way we're able to search it today? Yeah, I mean, I think it takes those
28:02 disciplines that are more text or language based, like whether it's all the land, like anything but like all the legal documents that are just so massive, you know, them or even legal, like
28:15 general counsel, just being able to, you know, search through that way more efficiently and way easier. I mean, we did a project at Greece and Mel and end up using like some OCR, you know, into
28:28 Elasticsearch and searching for certain
28:31 And it worked, but I mean, it took some time, but I feel like now it, again, I think we can weigh over simplified, but can be as simple as, you know, scanning those things in if they're, if
28:40 they're already digitized, and then feeding it to a language model and getting out. You know, that answer pretty damn quickly. At least in a way that you expect
28:53 so I think, like John was saying, it's like big on institutionalizing knowledge. So, you know, things have already been done like again, just because someone did it. Five years ago doesn't mean
29:04 no one else should know about it. I mean, we have so many things where people. Yeah, all the year views and this and that are people done the work and it's but it's saying a PowerPoint or a PDF or
29:14 somewhere else and it just goes to die and no one ever sees it again But, but then someone else has to find the same someone else has to find the same problem again, redo the work, even though it's
29:25 already been done so I think I'm limiting that double work or triple work whatever is a huge deal. which is like Colin's basis of why, that's what I understand why collide is what collide is, is
29:37 because there was this knowledge being lost. One or two guys know it and problems are being solved, but then it goes on and it is dissolved somewhere else shortly down the line and retaining that
29:50 data in a time
29:53 when maybe the workforce is slimming down in this sector and being able to continue to bring that information and knowledge to people and retaining it for your business. And, you know, in certain
30:03 scenarios, hopefully, you know, really helping grow the business by efficiencies, just workforce efficiencies, but also like just, you know, what do we know about these things we're drilling?
30:17 What do we know about these, you know,
30:21 things that are happening and accounting? What do we, you know, I just see so much changing, you know, so quickly and it's benefiting so many people right now. Well, there's a lot of
30:31 relationships too. Like the Bobby, you all mentioned the more tech space data, right? Like for example, gas contracts or takeaway contracts are written on a lease basis. So within those
30:45 contracts, the only thing that ties the contract back to any kind of asset is a, oh, I forgot the name of the table, but there's a table in the contracts, which normally have the section township
30:57 range. And that's it Right, and so then it's up to the operator to go find what wells that they operate or have interest in that are in that section township range. And that's a very manual process
31:10 that they have to do on every well. It's like, hey, it's called the dedication's table. If we can identify the dedication's table in every single contract, I can go through and I can extract that
31:20 information from the contract, then I can go look up all the wells that are within that bounding box and give it to you in a matter of seconds. And so it's like those are the types of things that
31:31 like, you know, are really making, it sounds simple, but that literally takes hours to days for these, you know, one of the most simple things that we're seeing, you know, there's going to be
31:43 a huge asset on the land side is taking a handwritten physical index and turning that into a table Excel sheet, you know, with grant or grantee type of document data file, you know, something that
31:56 takes a landman And if you're talking about a 50 page index with 50 entries on every page, you know, you're taking a day or two just to sit there and type that stuff out. So how does that increase
32:07 our efficiencies? We can spend more time running reports and chaining the title or reviewing what AI has said the title is, you know, and that's not my business, but that's, you know, the guys
32:18 we're working for. You know, that's the efficiency is that we see possible and it's something very simple, Just like what you said, it's just tabling a
32:28 piece of. paper. Yeah. Yeah. Well, even, you know, another piece that comes out of this is there. It's amazing. Like, if you can read the handwriting from a scanned document and the PDF,
32:40 just looking at the digital version of it, there's a very good chance that you're able to extract that data now. And that wasn't the case. I mean, if I was talking to you in December of last year,
32:47 I would have been on this, you know, high horse of how it'll never be done. It'll never, you know, I can learn new
32:58 things. And I'm seeing new things happening, you know, and I'm seeing it happening very quickly, you know, and that's just not a thing anymore. Where it's getting hung up is where me, you and
33:09 Bobby would be getting hung up. Where is that an M? Is that an N? What is that, you know, it's degraded writing, you know, and it's like, well, you would expire. I'm actually glad that you
33:21 have an error there and aren't able to decipher what that is because even upon, you know, human review, there is question. you know, yeah, no, that's a big, a big element to this is just
33:32 because it's scanned doesn't mean it's worth a shit. That's another element of this. And then well, I mean, it's an important thing, right? Like we talk about it all the time junk in equals junk
33:41 out. But you know, if you're, if you're trying to index and embed, you know, a PDF that is just the extracted text of that is just junk. One of them,
33:56 one of, I don't even remember what type of document it is, but it extracted this sentence and it said butter and pain. And I was like, butter and pain. And I was like, that is, I'm going to
34:07 make a t-shirt out of that for that client because it's just like, that's not even remotely close. Like, I couldn't tell you what it says, but it doesn't say butter and pain on this well, well
34:17 summary report.
34:19 And so it's just stuff like that. So, you know, just because it's digitized doesn't make it perfect for for these tools, but a large majority of it.
34:29 uh, you know, kind of pass the, the, the check test. And so, you know, you're getting hand-written notes about exactly what they did. And the craziest part is the language models now
34:39 understand a lot of the wheelfield jargon. So you put in, you know, it, it extracts poo and then you get an answer and it, it says poo. And then explains pulling out a pole and you're like, huh,
34:50 that's pretty awesome. Right. Like you can really start seeing how, how, even outside of a technical person's hands, they're able to get valuable information out of it without having to have,
35:02 you know, all the, the SME in the world about the specific subject. It's pretty, it's pretty awesome. Yeah. We do some work with a company or manufacturing company and we help them digitize some
35:12 of their, um, kind of warehouse, uh, like blueprints and schematics of what they're manufacturing, but we also help them with their invoices And something really cool that I have seen just this
35:25 year is being able to take what is like a quick books invoice, but then was printed and someone along the line and manager went in and scratched out500 here and put a discount in, just handwritten.
35:42 It's now capturing and understanding that, hey, that was an adjustment made here. And it's an actual reduction in the cost. And it shows up like that when you're searching for it. Where was this?
35:55 If it's1, 000 and it had a500 handwritten discount, you search for invoices that are500. And that one's going to show up. It's intuitive and it's learning these things and it's reading and
36:05 catching, even if it's turned on a different direction than the invoice was. So it's really, really interesting stuff.
36:14 Yeah. Yeah, I think an interesting - I mean, I think on the similar note, because John and I were just on a call with a company that deals with open invoice and turns those into nice, pretty
36:26 rectangular tables but like as someone who lives in kind of a data engineering. world. Like I do think there's two sides of this here. So I can just throw up all this crap in a LOM and it's
36:37 searchable and index. But I mean, that's not super efficient always in and out of John's kind of working through some of that too, where how do we make some of these things faster or narrow the
36:44 context faster, but is just, you know, you can use LOMs, you know, for what we're saying, just slam it all in there and we can search it. But also we can use is, you know, kind of ended
36:54 engineering pipelines, where we can actually pull out relevant data out of various forms that are not standard, but then get that same data out and put it into a kind of rectangular format, if you
37:05 will, you know, more traditional data formats that makes it super easy to work with and way better. I mean, one thing I think of that we always struggle to capture in our land system, you know,
37:16 when I was like, GME was like the penalties for when they, uh,
37:24 not consent.
37:26 I mean, we should have those in a. table where it says the drilling penalty is 300 and the other week penalty is 150, but we didn't have it in a good format. I mean, you know, doing payout
37:36 calculations really hard. But if we could just pass all the JOA's in, we should be able to pull that data out and put it into a rectangle format way more easily.
37:45 And, you know, that's just one example, but where we could just take data out of these or drilling reports or whatever it is and go, you know, from here and put it into a tab of the format that's
37:53 super valuable that we can use in, so it might be eye tools and reports, you know, I think as a whole other side of this, that's, you know, they're the similar but different use cases. How much
38:05 help are we getting from people outside of oil and gas? As far as like other people that are working on these problems or? Yeah, yeah.
38:14 I mean, I think that part is definitely being utilized across. I mean, there's some, I think everyone's got a, a paper problem or a, you know, unstructured data problem. Um, so I think there
38:23 are a lot of things Maybe a better question is, do you think there's more people today? trying to get in and solve problems in oil and gas, and maybe this is an obvious yes, then there were 10
38:34 years ago. Because of like the advancements in technology and what we're seeing possible, or do you see people that weren't originally or had no business in oil and gas getting into oil and gas
38:44 because of these types of things or energy altogether, or is it about the same number of people trying to solve the problems, or same people in the industry trying to solve the problems? I
38:59 mean, you've got your typical incumbents of the big tech companies who are always trying to get, you know, deeper into the energy industry, the Googles and azures of the world, of course, but I
39:10 mean, I've seen a mixture of both, you know, some kind of coastal companies, but mostly it's, I would say most of it's homegrown. People seeing problems within what they're doing and saying we
39:23 need to solve this Yeah, it's, you know, it's an engineer who dealt deals with the same stupid problem every day and is like, Oh, now I can use
39:35 an AI-based IDE to write some code and solve my problem for myself. And then I did that, and then my friends who also have that problem, it just starts to grow like that. But
39:51 which is normally where a lot of our best stuff comes from is an engineer that got pissed off with a specific workflow. I
40:00 think combo curve. Not someone that just kind of generally understands like chaining title. It's a land man that's dealt with
40:11 terrible HVP title for years and years and years, and wants to try to provide a solution for that. So, no, it's, you know, it's, it's interesting. I, you know, yeah, interesting observation.
40:24 To that point, I would say probably in the last couple of years, it's probably become more homegrown because people are more enabled by AI on that side of it, like John's on the coding or the
40:34 development side. Maybe we have Brandon on for a few weeks a month or two ago. And again, I think he was someone who was hacking together some R code and stuff a few years ago and now has a fully
40:45 fledged platform. It's interesting. So many of these people started self-teaching them. You know, as COVID was like this, like, I'm going to teach myself how to solve these problems that I had
40:56 pre-COVID that, you know, I want to solve once this is all done. At least that's what I've seen is so many people I talked to is like, where did you learn to code? When did you learn to code?
41:04 It's like, if it's not COVID, you were like, you went to school for it and you always wanted to do it. You know, it's interesting though. You know, seeing these homegrown people trying to, you
41:13 know, and we see other tech companies reaching out to us, trying to help solve these problems with us, be it indexing or things like that. But it's like working with John directly is like, no,
41:22 you understand what an oil and gas lease is understand the data we're trying to. all from here, you know, all of those things. So yeah, I think it's a huge benefit to have people, like you said,
41:34 homegrown, kind of working on this, people that are in the industry, solving the industry's problems. I just wanted to touch on your point about, like not enough people are thinking about using
41:45 LLMs in their data pipelines, because it's a very valid thing. You don't have to pass the entire document into a vector index and set up a full RAG model to take advantage of LLMs. You can pass in
41:60 a document and ask it to extract one very specific thing, and whether it's from a drilling report or a post job track report or a contract or whatever, but those are important things. There's a lot
42:13 of the bounding data for our numerical traditionally structured data lives in paper documents And so it's like, you know, if you want to do a study on all your fracks and your. completions and
42:26 stuff. Yes, you could go to the, you know, the ASCII files and compile all the post job or, you know, the post job ASCII raw data and get, you know, put that into a database and try and
42:39 aggregate it and hope that everything works. Or as those post jobs come in, you could put them through a process that goes into them and extracts them in the maxes, the volumes, you know, all the
42:50 important shit that you're worried about. And then just stores the document and that's it, right? Like, there's going to be so many of those workflows in the future that it's going to make a lot
43:00 of our data a lot more functional and distributed data. For sure, it's interesting. It's like not so much anymore about just warehousing the data or having a data, you know, management software,
43:10 document management software. It's, you know, that's now like a byproduct of it. It's like, yeah, it's all being warehouse here, you know, but we're doing things with it. you know, it's not
43:22 just go into this file, find this document or find the hundred document you need. It's, yes, of course the documents live within here now, but this is what we do now. So that's an interesting
43:32 transition that I've seen from just the document management solutions to now AI application. All right, sorry, Bobs. No, you're good. So I was gonna ask Scott, I mean, so if you just always
43:45 had a passion about scanning documents, like we're the kid that, you know, just played around with a grandma and grandpa's Xerox machine and, you know, or I get into this. No, I
43:58 think I've always had a passion for finding better ways to do things, quicker ways to do things, and I also really love, I went to school for psychology and sociology, so I really enjoy
44:11 interacting with people. I love working with my crew, I love working with the management staff, I love all of that, and I love seeing, you know, unique parts that you know make each one of these.
44:21 individual but also like the group mentality of like how can we get everybody behind the same idea of we're trying to grow the company and be the best and even if you're out in these courthouses and
44:32 you see people leaving at 2 o'clock in the afternoon they've only been there since 10 am. you know how do we as a company provide this culture and this idea of like just don't look over the fence at
44:44 what they're doing over there we've got you know we've got our goals in front of us and so I've always I've always liked that I was working at a bar when a buddy of mine came in and said he was doing
44:54 land work and wanted to take me out to a courthouse one day and took me out and I saw the inefficiencies I saw the lack of responsibility I saw the you know and even just the workforce I mean it
45:05 doesn't take a college degree to go out and do this but it takes and I some of the best people I've seen in my side of the industry don't have college degrees they've got a really good you know
45:15 mentality if I want to go work hard and get something done and we've been able to kind of incentivize those people as best we can, you know, chiefly, I think that's money incentivizes people pretty
45:26 decently. We try to take care of our guys as best we can and provide long-term work for them. They're all contractors, but most of our guys have been with us. We've got one guy that hasn't missed
45:36 a day of work with us for eight years. I mean, it's, you know, and I think that's pretty rare in the land space. So, you know, I was tired of making30, 000 a year, working 50 hours a week at
45:49 a bar and things like that. And so, this gave me an opportunity. You know, I mean, I would run with any industry that I was in, but this one, I've kind of sunk my teeth into, and I've been
46:00 here for so long now. You know, I say that I've been in the industry for about 10 years now. And for me, that's a long time. And I've enjoyed the people I come in contact with, I enjoy the
46:10 technology we're seeing, and I've just fallen in love with it now. So, I never thought this would be where I'm at and never thought it'd be. in oil and gas or energy. I thought we'd maybe be
46:23 sitting here having a therapy session.
46:27 But yeah, that's what life threw at me and I couldn't be happier. I loved meeting people like you guys interacting. This is a dream come true for me, not just to kind of make you
46:38 guys feel too special, but this really is. I've sought out people in the industry like you guys that are forward thinking and tech forward specifically, but also cool guys to chat with. So this
46:52 has been something that I've been real thankful to get the exposure with you guys and get to know you guys as well. You guys are super, you know, for as niche of an industry as I'm in, you know,
47:03 you guys are right there with me with a lot of it. So there's a lot of stuff that you guys do that I don't, but you know, we're able to sit around and job out, you know, documents and paper,
47:13 which a lot of people think are boring, so yeah. It's still the need to know you guys and learning more about, you know, I love it. I, you know, I'm a huge supporter of our local chapter of the
47:25 AAPL here in Oklahoma. And it's just, you know, that continuing education, I know nothing about oil and gas. You know, everything I know is from sitting here talking to people, it's from going
47:35 to events, it's from opening up a book and reading. You know, I don't have petroleum background. My dad wasn't an oil and gas, anything like that. So it's just this education, which, you know,
47:44 digitizing a lot of things is helping that, you know, helping us learn more. So I appreciate it guys. Love doing stuff like this with you. Appreciate the time and the platform to talk about what
47:56 we're doing and, you know, what you guys are doing too, you know. For sure, man, I appreciate it. We also appreciate you. And, you know, that's probably my favorite part of this industry as
48:07 a whole is it's full of people who just like to solve problems or identify that there was a problem and just sunk their teeth into it.
48:17 Yeah, I relate a lot. My dad was a food salesman. You know, it's like - Yeah, my dad worked for Anheuser-Busch for 25 years. Yeah, so I had no experience at all. But that's another cool part
48:28 about the industry is if you come into it knowing nothing, but you're humble and you show that you care and you want to make an effort and learn, we will absolutely teach you all day long. It's -
48:38 Teach you, but also accept you. I mean, like, again, if you work hard and do the right thing, like, this industry will reward you. For sure. It also goes to show just how much low hanging -
48:52 like, there are so many problems that exist out there that we are still trying to solve in this year of 2025. Yeah. You know, we had a call earlier last week with a client where we just need to
49:06 pull some data from their production database and they had recently moved there. their instance up to the cloud version of their software. And so I was like, well, surely we can get API access to
49:21 that, right? No, no such thing as API access to the cloud based software that they have upgraded to. And so it's just it's just shit like that that, you know, that still exists in our industry
49:32 today, that, you know, it completely prevents, you know, technology advancement. And so there's just so much of that still that, you know, if you've got the mindset and you want to learn and
49:46 dig it and solve problems, there are so many problems still to be solved. Yeah, I'd be interested to schedule a podcast for this time next year and rewatch this one and see what we're doing then,
49:59 you know, maybe writing our hoverboards into work and, you know, all of that kind of fun stuff. So I hope so. Yeah. Well, again, guys, I really appreciate you guys having me on today. Um,
50:10 yeah, other than that, love to catch up with you guys soon. Probably in the next month so Hetch up there and and yeah anybody that needs documents taken care of wants to talk about paper give us a
50:22 give us a shout We're not done with you yet. We got to give you the speed around. Yeah, we were pivoting more than wrapping it up I thought the gauntlet
50:34 I had a good one where would
50:37 Where would you take someone visiting Oklahoma City for lunch and then dinner? First place we would go we'd hop in the car. We'd drive out to a little town called Ocarci, Oklahoma It is in Canadian
50:50 County You go out there and there is one stoplight If I remember correctly and at that one stoplight you hang a riot and there is a chicken place They're called isons chicken. It is
51:04 My you know electric chair meal if there was one thing left on this planet to eat and it was that I I went out there one day and there was some tornadic activity at knockdowns. some power poles and
51:15 we showed up and there's a crowd of 30, 40 people standing out front and they had kicked everybody out because they lost power. There was dude standing out there with these two go chickens, you
51:23 know, walking out and leaving with their leftovers and I went up with a100 bill trying to buy a chicken from someone. And every person looked me in the face and said like, my wife will kill me if I
51:34 don't come home with this. Like I'm sleeping on the couch if I don't come home with this bird and no one took me up on it. We had to drive out the next day. It was a buddy's birthday. So we had to
51:42 go back the next day. But yeah, that's where we would go for for lunch and dinner. There's a number of spots, but you know, I love a good burger. There's there's a place called Sun Cattleco in
51:56 downtown OKC. It's just like a real traditional. You walk in, they got pictures of cows up on the wall. You walk up to the counter. It's an onion burger, but yeah, super delicious, kind of a
52:05 newer place. But yeah, those are those are the two. If you're from Oklahoma, know what ishons is. Yes. Yeah. I. I did a frack up there for Devon. God, probably 10, 15, not 15, but over 10
52:18 years ago, and it was a Woodford frack up there, and I think we were staying in Elk City. Yeah. But anyway, that's what they wanted me to go get them for lunch, and so that's what I got them for
52:30 lunch that day. Yeah. I was like, in my little - I got little paxas full of chicken. Yes, I was in my little rental compact car driving dirt, you know, drifting around dirt roads, trying to
52:42 get there I've got some - It was worth it. It's a bar too, a real famous bar. And so I've got some after hours stories that probably aren't appropriate for the pod that we can discuss next time I
52:55 see in person. So
52:58 sometimes had there, sometimes had had ice and chicken, so.
53:03 Nice. Okay, so you got a psychology background. What's your, is there a book on psychology that you, you know, was really influential or anything that you really enjoy. I had a psychology
53:14 teacher in high school. That was just the coolest guy in the world, super understanding. I went to a pretty uptight Catholic school and the guy was just, I think he got me. I was a little bit of
53:27 a Helian back then
53:32 and he was like,
53:34 I understand these things and also helped me understand some of them myself. But there is a book that I read recently. It's taken me a year to get through it It's kind of good. There's like 50
53:45 different practices in the book. And so for me, I was taking like a week at it. I'm telling everyone I was taking. I'm a slow reader, but I was taking a week at a time to kind of reflect on each
53:53 chapter. It's a little book of stoicism. And just kind of helps, you know, reset your brain on, you know, things in the world that are happening and how you're reacting to them, how they're
54:04 making you feel. And you know, kind of the biggest overarching thing I got out of the entire book is
54:10 that, things in this world, whether they're good or bad or just happening, they're not happening to you. And that's what I have to constantly remind myself is like, you know, you wake up in the
54:18 morning, you step in dog shit, like the dog didn't shit there specifically to fuck with you that day. This is just what is happening. Life is happening. Things are happening, they're not
54:28 happening to you. And once I can accept that with a problem, and I don't do this well every day, but once I can sit there and really think, okay, this isn't happening to you, Scott Like, this
54:40 is just happening, it makes it easy to put that emotion aside, and then just start dealing with these problems, whether it's something at work, home, whatever that is, even just something
54:49 mentally challenging. It's like, you know, let's look at this, like it's not the world's out to get you, you know? So, good book, good read. I'm a big supporter of therapy, all in all, and,
54:60 you know, all those things. How about that? That's awesome, I'm big. Yeah, I appreciate your transparency just even there, but even like some, your transferring you've had on LinkedIn.
55:11 stuff like that helps more people than people realize. For sure, you know, I'm an open book when it comes to these things. You know, you've got to either share every part of your life or you
55:21 can't share anything in my opinion, you know, on these social platforms. You can't just take pictures of the margaritas and the stakes you're eating, you know, you got to tell people about the
55:30 times you were down and broken also, so. For sure, we appreciate that, man.
55:37 Big fan of the Stokes as well. I think I was in like junior high or high school when I heard, I don't even remember who's quote it was, but it was, you know, be like a young sapling in the wind.
55:48 It bows, but it doesn't break and it always kind of comes back up to the center. And I thought you were originally gonna say, Walden by Thoreau, 'cause that's another dense one. Yeah, I do have
55:60 a really cool Walden story. If we've got some time, my neighbor in college, we live in these townhomes that were just right next to each other, but there was a bush in between. And we would
56:11 always walk out our front door and walk right to his door, you know? In and out, he'd walk out right to my door. He was a grassed area, and by the end of the year that we lived there,
56:23 the grass that was there had died. It had died in between where we were walking, just sidewalk to sidewalk or door to door. You know, and that was like Walden's deal. It was like, you know, he
56:32 walked down to this pond every day and walked down to this pond every day. And, you know, after so long, the path became clear It was, you know, the grass died, everything opened up, and that
56:41 was, you know, what happened. And so we had a Walden experience in my life who was in college and walking to the neighbor's house, just, you know, trudging. And that guy's actually my best
56:50 friend to this day. That's awesome. Yeah, pretty cool. Big fan of throw in all the existentialists from back in the day. I never get to talk about any of that stuff, so I'll happily take that
57:02 this time to do so. Yeah, for sure. No, I love talking about that stuff. college was a fun time for me, you know, not just the hanging out and drinking and partying and stuff, but, you know,
57:14 the education, I've always been a big fan of learning. And that was one thing that like I enjoyed learning about was, you know, my own brain and everyone else's brains and how we all why we all
57:26 react and act the way we do. And, you know, all of this thing. All right, I got one more for you. What is your favorite outdoor activity in the state of Oklahoma? Favorite outdoor activity in
57:39 the state of where? Are you going to a lake? Are you going camping or are you going camping? Yeah. So I've got a school bus that I've been converting. Oh, that's awesome. Yeah. So I got it
57:50 about a year ago. I stopped drinking about two years ago now. And through that process, I was finding it hard to find things that I had hurt my brain a little bit. And was finding it hard to, you
58:03 know, find things that I was, you know, getting enjoyment or satisfaction or feeling like I was progressing in outside of the business. And so I was working with a coach at the time and we did
58:15 some kind of like deep work on finding what that was. And I came to the conclusion that I love camping, but with three young kids, it was hard to get them intense. And I've always loved kind of
58:25 the bus life. I watched a movie about five years ago with these people living in buses and it's always been attractive to me. And so I went out and got a bus about a year ago and I've been working
58:34 on it. Ripped all the seats out, took me way longer than it should have. I don't know anything about woodwork and tried to build a bed in the back, which is stable. I had some help with that.
58:44 But you know, it's a beautiful thing, I love it. And so camping is a huge thing. My wife and I always really, we met and kind of our deal was going to music festivals together and doing that and
58:55 that all slowed down when we had kids. So this is allowing us to get back out and do that kind of stuff. I also really love riding my bike diagnosed with epilepsy two years ago when I stopped.
59:05 drinking is kind of all one and the same and so my main mode of transportation is cycling but through that I got into road racing and things like that and so kind of two and one I'll throw the bike in
59:17 the bus take the bus to these races hang out and do stuff like that so there's a lot of I've gotten into the kind of the active community of biking and road races and things like that and so I like
59:29 doing that a lot around Oklahoma there's a good community here for that and there's a lot of outdoor stuff to do as much as people think Oklahoma's just a flat plane you know there's there's quite a
59:38 bit of you know we've got the watch of Tom mountains which is you know not like Colorado mountains but it's beautiful you don't feel like you're super slow down there super beautiful broken bow area
59:49 we've got a place here called Little Sierra it's like a desert out there people go and race side by sides and do all sorts of stuff out there and you know we've got some cool little gems around I
59:60 don't know if I'll live here forever, but I do love it for now. Nah, that's awesome. The watch towels are super slept on. Yeah.
1:00:09 As an Arkansas grad, it's, yeah. Oh yeah, I forget you went to Arkansas. You gonna pay at chill? Were you a Fayette chill guy? I wore a Fayette chill shirt yesterday, man. You dog, if I
1:00:20 didn't go to OU, I think that's where I should have ended up. No, that guy was in, the guy that started Fayette chill was in one of my entrepreneurship classes. Oh, really? Actually, yeah. So
1:00:32 it's a small world, but, yeah, people don't realize that the Ozarks or the watchtaws and stuff come into Oklahoma. Right there, yeah. I used to like flying into Tulsa just because I was, you
1:00:44 were basically right in on, and I was like, oh, I didn't even know this, but it's a, yeah, very slept on, beautiful part of the country, for sure. Yep. Repasset, Bobs, if you got one more.
1:00:59 Not really, I thought you were the last one, so I got to take the brain out on that. That works. Scott, where can, uh, where can people find you if they want to get in touch or they need, they
1:01:08 need to do docs. Linkedin is the best spot. Uh, last time I was on a pod, I rattled out my phone number, email, all of that kind of stuff. And I was sitting in the cab on the way home and I was
1:01:19 like, why, why did you give everyone your phone number? I mean, I used to DM me. I'll give you my phone number or look at some of my post, my numbers on there, but I don't need to give it right
1:01:28 now. Just find me on LinkedIn. Um, easy, easy to locate. So perfect Well, this has been great man. We're glad I'm sorry. Let's plug collide. I'm also on collide DM me there as well. Um, I,
1:01:39 I post almost as much as I do, I, I, I equal, I, I equal right now. So, um, yeah, find me through collide. That's actually where you need to find me. DM me there, ready to chat. Perfect
1:01:51 man. We appreciate you coming on, jump on collide DM Scott, get your data useful again. Um, and we appreciate it, man.
1:02:03 We will see y'all next time, appreciate it. Appreciate it, guys.