How AI is Transforming Construction Estimating 

Podcast Transcript

Steven Rossi: 

Hello there and welcome to the Construction Revolution podcast. My name is Steven Rossi and here on the show we explore the latest trends, technologies, people, and organizations that are revolutionizing or disrupting the construction industry and are changing what the industry will look like tomorrow. Today we’re thrilled to welcome back Melvin Newman to the show. Melvin is the co-founder, president, and CTO of PataBid, with over 15 years of experience in construction estimating and project management. Since co-founding PataBid in 2018, Melvin has been revolutionizing the industry with cutting edge software and digital tools. PataBid’s advanced technologies, including AI and machine learning, transformed the bidding process for contractors. Their recent supplier integrations with City Electric Supply and Gaskin, along with automating pricing features, are revolutionizing the industry. If you’re a contractor or tradesperson looking to stay ahead of the curve, you won’t want to miss this episode. Let’s dive in. Hi Melvin, welcome back to the show. How are you doing today? 

Melvin Newman: 

Doing well. It has been a hot week here on every front, so doing really well. 

Steven Rossi: 

Yeah, that’s good. It’s finally warm here today. We’ve had so much rain, nonstop rain, it seems like. It’s nice to see the sun. 

Melvin Newman: 

We were at 32 degrees Celsius yesterday here. For any American viewers that is, I don’t know, like 80s, 90s, it’s smoking hot. 

Steven Rossi: 

Yeah, we’re something like that today. Yesterday we were closer to 20 and rain and thunderstorms. Lots of fun. Glad things are going well. Since we were last talked, you’ve been busy by all accounts and both yourself and PataBid have been evolving and growing it seems like at a ridiculously fast pace, so I’m excited to hear more about it. 

Melvin Newman: 

Yes, it has been an absolutely blockbuster year for us here. I mentioned it earlier off-screen here, but yeah, we’ve grown 300% since January in just our estimating platform. It has been wild, and yeah, it’s been just continuous development adding into the platform as our clients have needs. That’s what we do. 

Steven Rossi: 

Yeah, absolutely. That’s awesome to hear and congratulations on the growth and the year’s only halfway done, so I’m excited to hear where the year ends up. 

Melvin Newman: 

Yeah, it’ll be interesting to see what the fall brings. I mean, in the electrical industry, especially in most of the construction trades, the summer is when nobody thinks about home office stuff. Nobody wants to really get too excited about software at this time of year, they want to be out building things. But it’s been fascinating. It has not slowed down much. We’re getting requests from quite honestly all over the world. We signed on our first people in The Bahamas, our first people in Puerto Rico last week, and I’m pushing for onsite training so we go and help them onsite, but it’s tougher sell. But yes, it has been a really, really interesting year so far. 

Steven Rossi: 

Yeah, no, that’s great. I mean, I work on our events team here and I always say there’s construction season over the summer and then there’s event season and that’s fall and spring. Not too many events going on over the summer. 

Melvin Newman: 

Yeah, that’s when people need to be building while they’ve got the weather. Especially here in Canada, nobody wants to even talk during the summer about anything other than get the building closed in, get the heat on for winter. That’s all we care about. 

Steven Rossi: 

Yeah, for sure. All right, so yeah, well let’s dive in. So I’m curious to know if you can just tell us a little bit about what you guys have been up to at PataBid since we last spoke and the updates as a company and to the product as well. 

Melvin Newman: 

Yeah, so I mean we’ve expanded our team at the company. We now have several people in sales and customer success and those side of things, and that’s been absolutely awesome for the team to start to get the clients rolling, get all of that going. On the platform itself, we’ve expanded the AI so that things like tables, like tabular data on your drawings, so equipment schedules, lighting schedules, all of that can actually now be automatically extracted from the drawing by simply drawing a square around it and then the AI pulls that out, rips apart the table, identifies what you need in it, and builds preliminary assemblies for you. So that’s been a huge one. We implemented multi-monitor support, which is something typically unheard of in web-based platforms and it turns out it’s because it’s incredibly hard to do, but we went and did it and so it came out just a few months ago. 

Basically the ability to split the browser up and throw windows on different screens and see more of your drawings as you’re doing the takeoff. It also has me super excited because what nobody saw was the updates we had to do to the backend to get that working and it was a massive overhaul of the server infrastructure, but it is enabling a foundation for us going forward to do some absolutely wild collaborative capabilities, so real time chat, internal, external, all this stuff but built right into the estimate. So we’re looking at kind of a future with that. 

And then the other big one is our integrations with suppliers. So that has been a game changer for the contractors; the ability to pull in their real time pricing. And the first one out of the gate with us on this has been City Electric and it’s been an absolutely amazing response on how that all worked for the contractor. And we didn’t even really envision this, but for each of the branches at the supplier, their contractors don’t need to send them 100+ line item Excel sheets to price now, it’s just instantaneous. So I had a conversation a few weeks ago with one of the branch managers and he was just, he’s like, I’m trying to get every one of my contractors on here because he’s like, it saves my team so much time using this integration. And so that’s been just some huge feedback and we’re looking at integrating with others, but that’s been absolutely a massive game changer. 

So those would kind of be the top three or so. There’s been a ton of little updates throughout. New ability to report on your statistics over the year, which has been kind of cool. You click on a folder full of estimates and it tells you everything about those estimates like how you won them, all of that kind of stuff. On the other half of our platform that monitors public procurement, we just released an update this week where the AI now actually pre-reads a bunch of the documentation. And I got to give credit to our summer intern. She put together, basically handed her an impossible task and we just don’t tell people they’re impossible, and now it goes through and in basically real time, like the moment an opportunity comes out, the AI is able to grab the documents on that, pre-read them so that you get notices of liquidated damages, notice of schedules in a morning email for everything that we can get the documents on. And so now when you’re trying to find that kind of perfect sweet spot work to go take a look at, it’s doing a ton of the pre-work for you. 

So we’re taking our AI into just a whole bunch of different aspects of it. And what’s really unique about our stuff is it’s in-house developed. This is not run on ChatGPT or anything of that nature. Again, when you don’t know something’s impossible, you go and figure out how to make it possible. So we’ve been training our own AIs for a very long time to do this sort of thing and it’s really awesome to see what the capabilities are now. 

Steven Rossi: 

Yeah, for sure. No, all of that sounds really exciting. I mean the multi, as a web-based platform, it makes sense. Not really something most people would think of is the multi-monitor thing. 

Melvin Newman: 

It was interesting because browsers, for various good reasons because honestly some parts of humanity are not that great, they have to put all sorts of barriers up so that no two windows can actually talk to each other. There’s all sorts of security, that sort of stuff. But we actually found some really unique standards compliant ways of not bypassing them, but working within that construct to enable two separate windows to actually talk to each other and it’s really slick when it’s going. And it also allows now when two estimators are in the same project working together, they update in closer to real time, that sort of thing. And then, yeah, in the future it’s going to enable a next level of collaboration on this. 

Steven Rossi: 

Yeah, that’s really exciting. And then, yeah, one of the other things you mentioned are sort of those integrations and how I think what’s great is that your suppliers are getting value out of it as well as your end users. So I’m wondering if you can dive into a little bit more on how those work. 

Melvin Newman: 

So a key understanding is the problem, and this is where anybody can really solve any problem, I’ve discovered, the core problem is actually identifying the problem to solve and being able to properly formulate what the actual problem is. And in this case, we knew what the problem was from the contractor’s perspective because we had actually been in the trenches doing this. We didn’t actually fully appreciate that the suppliers have effectively the same problem. And the core issue here is, in these complex trades, so your mechanical, your plumbing, your electrical, these kind of core complex trades in construction, when you do a takeoff, electrical is actually the most complicated, then you get the plumbing guys, which are kind of the next most complicated, and then you have the sheet metal and up from there. But when you’re in these situations, an electrical estimate is going to have hundreds of line items of widgets that they have to install and they’re all different material types. 

You’ve got your cable, you’ve got your conduits, you’ve got your end connectors, your breakers, your panels, your devices, like all of these just crazy numbers of items and you have to price all of them. And the price of course, especially on wire, and in the last couple of years, PVC conduit, these sort of things is highly volatile. It is very difficult to keep on top of those things and they’re changing quite literally week to week. So contractors on any sort of major job would export these lists, these bills of materials and send them out to the suppliers to price. And these bills of materials would range anywhere from 20 to 100 line items and the suppliers would hate it because oftentimes the contractor is doing their takeoff, so for them to generate that bill of materials takes several weeks, and then they send it out to the supplier and the supplier they tell them now has two days to put together the price on all of this. And then the contractor has to try to match that all back in and there’s just a huge number of points of failure in there. 

Somebody literally types the wrong number into something and suddenly in the best case scenario you don’t win the job and you just look a little silly because you’ve got $50,000 where it should have been 5,000. Worst case scenario, you type 5,000 where it should be 50,000 and now you’re in a much more difficult situation. So the direct integration allows the contractors to tie in to these suppliers at their account level. So it lets them know what their exact price is for these materials. And for the day-to-day commodity materials, it’s instantaneous. So it’s whatever that day’s price is from that supplier. And in our cases here, especially with some of these suppliers like City Electric, it’s real time. You can call the manager and negotiate a better price. If you’re buying 10,000 meters of wire, you can negotiate for a better price for that, and the moment that they save that price, you can pull it in. So it’s effectively real time communication with the supplier. 

And then when we deployed this, we didn’t realize that the suppliers also hated this whole function of construction because it takes them many, many hours of one or more of their people to flush out all the pricing for all of these little fiddly bits. And everybody would rather be working on the big packages, your lighting, your distribution, all of those pieces that are going to be custom to the job, take way more intellectual effort if you will to sort out and design out, and are vastly more valuable on the job. But on a million dollar job, 200,000 of that, 250,000, might be all of these little fiddly bits. 

So you have to get them in very accurately. And on jobs where you’re within 5 to 8% of the next guy, getting that price dialed in as accurate as possible is your edge, because everybody’s going to get roughly the same price on the lighting package, roughly the same price on distribution. So you have to make sure you’re dialed in on all these commodities. And so yeah, it just is a massive time saver and it runs 24/7, 365. 

So when you’re that small owner-operator and you’re trying to close that job for your client who’s building some new massive house somewhere, if you’re in Vancouver, it’s practically a mall that you live in, but if you’re bidding on those jobs, you’re putting that pricing together at 10 o’clock at night, usually between the hours of 10 and midnight if we’re honest, and no supplier is going to get you your pricing at that hour. So with this integration, you just hit the button, hit next, and it pulls your contractor price into the platform for you. And that’s been for our kind of small to mid-size guys who are in those situations, it has been a lifesaver. So yeah, that’s kind of the big one on that front. 

Steven Rossi: 

Yeah, no, awesome. Yeah, that’s really interesting. I think it’s great that you’re able to provide value for everyone and I mean that’s the best part of new software and AI is that we can let people do those more sort of complex and also interesting tasks to them and free up some time. 

Melvin Newman: 

And the relationship tasks. No AI is ever going to call your supplier and negotiate a better price for you. Like at the end of the day, AI has a really powerful place in what we do day to day, but we still have to keep in mind when we’re building, it’s humans that we’re building for. We’re building for our future living spaces. So it’s humans on both ends of this equation. So using the AI where it applies to get rid of that grunt work is absolutely huge so that we as humans can hopefully find new and more innovative ways to actually interact with each other, not quite on the cell phone all day. 

Steven Rossi: 

Yeah, for sure. Yeah, that’s a great point. So then, I mean speaking of the PataBid’s AI, and I think you have different sort of AI personas across your different products, how do those help streamline processes for contractors? 

Melvin Newman: 

Yes, this drives my wife and kids nuts because they have these virtual brothers and sisters. Our three AIs are Doug, Diego, and Lucia. And Doug was the very first one, runs our tenders platform. Diego was the unfortunate middle child that is often forgotten and handles some of the bits and pieces on the estimating side. And then Lucia is the AI that really backstops the estimating side. And it was really kind of funny, Lucia was loosely named after my daughter and then turns out that Lucia is also the patron saint of the blind. And so it was sort of, huh, this is estimating. Yup, the blind leading the blind. So it worked out on several levels. But they really are there in order to take that grunt work. They’re very, very good at the things that humans are not the best at. They’re very precise, they’re very accurate at detailing, identifying, and a huge aspect of them is that they are very fast. 

And now where some people get a little bit scared about AI is they equate faster with being smarter. That is actually not a valid kind of connection. Just because AI is faster does not mean that it’s any smarter. I mean, you can look at a Formula One car, it’s vastly faster than the car that I drive, but it’s utterly useless on the roads that I drive on. So it’s one of those things where faster is not always better. It can be very much better when applied in the correct situations. 

And AIs are also very, very good at really those kind of menial things that when we do them, we get bored. AIs don’t get bored. So when you tell an AI to read a thousand-page spec document, it’s going to do that with a cold calculating efficiency and be able to extract things that are on page 900 that are hazardous to us that by the time we reach page 900 in the two weeks we have to read this, we’re brain-dead and we are looking for any excuse to do anything else. And so that’s really where these AIs are able to help backstop the people that are involved in this. 

And AIs like Doug and how it interacts with the tender platform. So there are thousands of construction opportunities that come out across North America every week. There are hundreds of them a day. There are thousands of total opportunities per day that come out. It’s absolutely crazy. Sifting through all of that is a monstrous pain for people. So this is somewhere where AI can step in and it’s going to, again, coldly, efficiently calculate all of those data points out and then bring to you what is most valuable. And now with the ability to pre-read the documents in there, there is that capability to identify better opportunities much more quickly. 

And often what happens in a lot of companies is there’s one person off in a corner whose sole job is to do this and basically they either miss things or it’s just not a fulfilling way to go. So with these kinds of tools, they’re able to take a much more intellectual approach to it because the AI identifies those kind of problem points faster. And then again, on the estimating side of things, it’s all about speeding up the quantification and then doing that kind of analytics so that extracting the tables, extracting all that kind of stuff, and then backfeeding all of that into the pricing and all of those pieces. That’s really where the AI step in to become incredibly powerful. 

Steven Rossi: 

Yeah, for sure. That’s great. I guess I’m curious from a business point of view, why name them three different things? Why not just call it one thing across the company? 

Melvin Newman: 

Because they were literally created at three different time points and they were an evolution of… It’s an excellent question. Each one is actually fundamentally extremely different from the other. This concept of generalized AI is a curious one to me. I think that a generalized AI is potentially achievable, it’s just going to not be overly useful. And even if you dive into the structure of that AI, it’s going to have different parts of it that are domain specific. Like our own minds are actually domain specific. You have the brainstem and all of the auditory processing parts, then you get into the vision processing pieces and the various functions of the mind that are actually, it’s one mind, but it’s got domain specialized areas in it. I think we’re going to notice that with AI. So yes, when we talk to people, we refer to it as our AI, but these are really actually three functional units. 

They actually do talk to each other in a number of points. Doug and Lucia especially communicate with each other now, but they are actually domain specific. They’re trained on specific pieces of data to do those things. And this is actually one of the core weaknesses of ChatGPT. It’s trained on everything. So if you ask it domain specific knowledge, it tends to fall down spectacularly. And the same goes for Google’s AI there that they released a few weeks ago that was very strongly advocating that cats lived on the moon and Neil Armstrong found them and it was hallucinating extremely badly around that. For us, you can’t have those problems when you’re dealing with estimates and that sort of thing, so that’s actually why we don’t use something like ChatGPT that’s been trained on everything because that’s utterly useless for very domain specific use cases. So ours are literally trained on the domain data. So that’s kind of why we have three. They are the specialized ones in each of those areas. 

Steven Rossi: 

Yeah, no, that’s great. That makes sense and I think that, yeah, the brain analogy is a good one. That’s not one I’ve heard before, but it makes a lot of sense. 

Melvin Newman: 

I got to be straight up honest, when we were sitting down and looking at this, I looked at it and I’m like, okay, we actually need an NLP, natural language processing AI, and we need a vision one. There’s no way that you can train one thing on the state of the art to do both right now. We need to actually create the domain specifics and then make them talk to each other. And that became the structure for our overall architecture on this platform. 

Steven Rossi: 

Right, nice. So I guess we’ll continue the conversation on AI and developing all of these and sort of getting the data, training it, and implementing it. I’m wondering if you’ve come across any issues like regulatory or compliance or also just adoption from users and hesitancy on that front with all of the new privacy things that people being hesitant as we become aware that our data may not be the most secure if we plug all of it into ChatGPT and all of these places. 

Melvin Newman: 

Okay, so that’s a huge, huge one. Yeah, 100%. So yeah, and again, I mean just simply things like Microsoft OneDrive, you have to be extremely careful now. Photoshop is another massive one. They changed their Ts and Cs a few months ago that basically said anything that you do in our cloud services, which they’ve now moved Photoshop actually over to, they claim ownership on all your work. I’m like, how is that actually a valid thing in today’s day and age? I mean, they’re literally going to drive people to the open source solutions like GIMP and Inkscape, not because those solutions are necessarily better, they’re literally the only way where artists can maintain control over their own works. And that’s kind of just on the art side. And so in our specific situation, technology has always outstripped the law. And so we have to be very, very careful when we’re on the leading edge, the one thing you never want to do is generate new law. You never want to be the guy that innovates to the point where they now have to make laws because of you. That is a massive issue. 

For the vast majority, like anything that we’ve trained our AIs on, it’s been domain specific publicly available data. So things like tender postings, we don’t touch anything in the private sector for all of this. We don’t have access to it, we don’t try to gain access to it. And so I personally had a chat with the privacy commissioner here in Alberta and one of the things that kind of came out, or this was a couple of years back actually, this was before ChatGPT, but we were training Doug on all of this data. And I was at an event with the privacy commissioner and I sat down with her and I’m like, hey, I just want to throw something at you. Here’s what we’re doing. And I’m like, I can’t find any particularly legal reason why we shouldn’t be doing it this way and here’s what our end goals are. 

And so she looked at it and she was like, okay. She’s like, you’re right, there’s no law against what you’re doing. You’re using publicly available information. It is, you’re not generating stuff from this, you’re analyzing it. And that’s actually a huge core difference. But she’s like, you do have to be extremely careful because you are consolidating a lot of data, and even though it is not technically private, it is not technically copyrighted per se, you don’t want to become a problem. And so that went into a lot of the design cases around how we implement security and isolation and all of that kind stuff within the platform so that it is very isolated and restricted. 

And again, one of the core issues with you’re allowed to consume things in life, like we as humans, we’re allowed to go and look at a picture in the Louvre that would be copyrighted. We’re allowed to do that. The problem comes in is when you try to copy that. So if you went and looked at the Mona Lisa and then went home and created replicas of the Mona Lisa and tried to sell that, that’s where things get far more murky. And if you were making direct copies of the Mona Lisa, the law is very clear that that sort of thing is… Actually I think the Mona Lisa is now public domain, but that’s a whole separate thing. Let’s say that you’re trying to copy something that is not public domain. 

Like the law is very cut and dry on that. You’re not allowed to do that. If it’s still under copyright, you’re not allowed to copy it. At what point have you consumed enough copyrighted material that you can make something that looks copyrighted, but isn’t, but that has all been generated very clearly from copyrighted. If a human does it, we are analog creatures, so we can say we were inspired by all of these things and that drove us to make this and that’s a valid counter argument. When you train AI on copyrighted material, it’s a digital thing. You’re actually adjusting formulas in there to create something out. It’s not actually creating a new piece per se, it’s literally blending all the other pieces together mixed with a little bit of your input to randomize it and add some nonlinearity, et cetera. The law is not cut and dried on that yet. And OpenAI, Microsoft, and Google are at the forefront of starting to cause those laws to have to be evaluated. 

And it is so fascinating because I’m not sure what the end game is here. We’re going to run into problems and I’ve seen this like artists looking at just no longer generating material on the internet and not posting it there. They’re going to possibly go back to just hard copy art, which is tough enough to distribute. And those sort of things because of what has happened with ChatGPT and what it’s opened up. It is definitely a Pandora’s box and in some cases gold dust is coming out and in other cases it’s spraying manure everywhere and it really seems to be day to day and the use case specific what’s happening. So like I know I’m throwing out a lot here, does that answer your question there, Steve? It’s a fascinating question. 

Steven Rossi: 

Yeah, no, for sure. I think that’s great insight. I mean, it’s definitely that you guys, you obviously know a lot about the topic and you’re aware of it and making sure that for your specific AI, you don’t run into those issues and I think that that’s great. 

Melvin Newman: 

We’re trying really hard not to. That’s the key thing. And I mean, yeah, a lot of this is evolving as society evolves with this. We’ve never been in a position as a human species where we have these tools to do this sort of thing. And so society is evolving with that, and yeah, we’re trying to do the good citizen approach as best as we can in all of this and see where it goes. So yeah, we’re keeping a lot of eyes on what is happening out there, what are the ramifications, those sort of things. If the user has a specific issue, we ask them to, you know, you send it to us. If you need something tweaked here, you send it to us and we’ll work on that with you, make sure that we all understand what we’re going to be doing and what it means. So that’s just a core part of it. 

And of course if you train on everything on the internet, it’s shocking how much bad stuff there is on the internet. We all know. I’m almost wondering if there’s going to be, like somebody some day is going to create something like a mesh net. Something that runs over top, not necessarily a dark web, but literally something functionally different where they can control the access to it so that you can have a safe area for copyrighted material or for artists to express what they’re doing, that sort of thing, where AIs are not allowed to approach it and carefully monitor. I don’t have a clue how you do that at this point. You’re heading into like Arthur C. Clark territory when he designed satellites before humankind had even made it into space yet, so it’ll be really, really interesting to see how we as a society start to address these problems. 

Steven Rossi: 

Yeah, for sure. I mean NFTs and that’s one way, but yeah, I think that’s not necessarily the way. 

Melvin Newman: 

But NFTs don’t prevent access to the AI. Like we need a fundamentally different network that is somehow gated off to that sort of thing. I’ve literally got a buddy of mine that is looking at setting up a mesh network just for fun and he asked me to join in and if I’d be interested in joining on this. So in our town, we’re actually looking at building a completely isolated from the web mesh network that runs over low powered radios and you can literally hook these things up to solar and stick them to a light pole, and nobody’s going to even notice them there, and they extend your mesh network. It’s a really unique little way of doing this, but it got me to think of something like some sort of overlay network on our planet that would be for humans only, if you will. Something of that nature. And it would have to run on some sort of different protocol that would be the antithesis of how AIs work and yet still compatible with humans. So we’re heading into just wild conjecture here, but it’s fascinating to think where it could go. 

Steven Rossi: 

Yeah, for sure that could be really interesting. When you were on last time, you mentioned some hesitancy when you were going to people with the new technology and how different it was. So I’m wondering if you noticed that change or how you’ve changed strategies in sort of approaching people and if that’s led to a different result? 

Melvin Newman: 

Yes to both of those questions. So I mean, one huge thing that ChatGPT did for us was it opened everyone up to AI and now everybody became a lot more comfortable with this concept and moving into that. So our clients have been far more receptive to it. Now, the flip side in what we did that kind of changed our direction with the AI bit was we tweaked the tooling so that the human is in far more control of what the AI does. So the human tells the AI what they want it to do, then the AI presents the results back to the human, and then the human can adjust those results however they want to their specific domain. And that has made it a far more comfortable topic. It’s not just, hey, look, the AI has gone and done your estimate, because honestly, you never want that to happen. 

There’s no way the AI, unless you literally built that AI in your company and trained it on just how your company operates, there’s no way you’re going to get an AI that’s domain specific enough to give you a valid estimate for you. They can do probably within 10% accuracy in a lot of things, but that can sink your company if it’s the wrong way. And so we wanted to make sure with our clients, and we worked with them to ensure that they are in control of the entire process and the AI is simply directed. And that became a major step up. It became much more comfortable for our clients to interact with it and work within. And they also then tend to learn what the AI sees and the AI doesn’t see things like humans do. I mean the AI will never in our existence go and install a plug in the wall, so it doesn’t have a clue what that is. So it doesn’t have any life experience to fall back on. 

So our clients are working much more collaboratively with it. And it’s been fascinating in the last, I’d say six to eight months, we can loosely watch what’s going on with the AI and the client interactions because we monitor the servers that the AI is running on and we monitor them for CPU usage. So it’s been fascinating to see how much more use, and you can just simply tell by the CPU’s usage ramping up during business hours and then ramping down at night and see what’s going on and how many more people are engaging with it now. And that’s been really, really quite cool to see. It’s become a much more useful tool to our clients and there’s a lot less hesitancy towards it. And the feedback that we’ve been getting has been absolutely awesome. 

I was out actually at an event last week here in Calgary. It’s Calgary Stampede season now, the big draw for everyone. I got invited out to an event and I was actually walking through the event grounds and one of our clients was there and he came running over and we shook hands and the sponsor who had brought me in, she was looking, she’s like, oh, how do you guys know each other? And he’s like, oh, I use PataBid. And so we got chatting and it came out that he had just recently submitted a bid on the biggest project that they had ever submitted in his company. And he’s like, the only way I was able to do that was because of the software. He’s like, we were able to tackle this project simply because of the AI assisted takeoff, the integrations with the suppliers, everything has come together to just enable this capability. And he had hit it out of the park on a brand new, it’s literally, if the project goes ahead, it’s going to be a brand new era for his company. It’s wild. So that was really cool to see. I’m not going to lie, that was awesome. 

Steven Rossi: 

Yeah, absolutely. I mean, that’s why you’re doing this at the end of the day, so that’s great to hear that all the work you’re doing is paying off. 

Melvin Newman: 

Exactly, 100%. And I mean, that is why we exist. It is to basically enable the industry, enable these people to take that step up. He had come from pen and paper. And he had started his own company, he was a younger guy, he had got his master’s certificate as an electrician and then decided to do the wild thing of jumping off the cliff and trying to build the wings on the way down and actually managing to pull it off. And to be part of his success story was just, honestly, it makes my day. If we can be part of the success story of our clients, that is straight up our goal. 

Steven Rossi: 

Yeah, absolutely, that’s great. So to close things off, I’m wondering if you can tell us about any new features or things that you guys are working on at PataBid that may be coming either soon or down the pipeline? 

Melvin Newman: 

Absolutely. We are working now on implementing the ability to write your whole proposal within the platform and just have your numbers effectively feed into your outgoing document. It goes back to that automating the communication aspect. And so we have found, again, because we live in a golden era of technology, we are able to basically have, I’m not sure if you’ve ever heard of this, LibreOffice? We could use Microsoft Word, but the licensing agreements basically would destroy my soul. So we didn’t go that route, but we’re able to integrate like full Word/Excel capabilities right into the UI and stream that out to the user in some really, really cool ways so that they can do full detailed publishable documents for their submissions. So with all of their certificates attached, with all of this detail, attach the drawings that they’ve marked up or drafted out in the platform, plug all of this in and basically just hit that send button. 

So again, they’re able to take out these incredibly time-consuming pieces and automate them. And through that we’ll be able to actually tie in some of the AI so that it will be able to, in this one instance I really want to build the SQL to Lucia, and this AI would be one that actually works with each individual company. It becomes domain specific in that company so that when it looks at a job that that company has done, and after about 15, 20 jobs that they do in the platform, it’s able to start to identify where their key risks are and start to suggest out of a pool of risk analysis, here’s the things that you should take exception to on this project. 

And not to write that in for them, but to just help them get that second glance when they’re that 20 minutes to closing and all hell is breaking loose and they need just a cold set of eyes that does not feel that tension to help them identify these risks. And so that’s where we’re headed. The key part is that proposal authoring tool that’s going to be hopefully coming up before the end of this year. And more integrations with suppliers, that’s kind of the next big one. But yeah, there you go, Steve. Any other questions? 

Steven Rossi: 

No, that’s super exciting. I think they last just if anyone’s interested in sort of learning more about PataBid and getting started, how would they do that and where should they go? 

Melvin Newman: 

Https://PataBid.com. Google us on PataBid. Alternatively, you can go to our YouTube channels. We are regularly uploading tutorial videos, that sort of stuff on the platform there. So if you search PataBid on YouTube, you’ll come across us. If you need riveting, insomnia curing content like our estimating tutorial videos, they might be a cure for your insomnia, but they’re full of really good information and that’s probably the best way. Or you can just email me also, melvin@patabid.com. 

Steven Rossi: 

Okay, awesome. Thank you so much for your time and joining us again and I’m excited to see how PataBid keeps growing. 

Melvin Newman: 

Awesome. Thank you very much, Steve. It’s been a pleasure.