Fraser Patterson of Skillit. A Data-First Job Platform Aims to Solves Construction’s Labor Shortage

We talk to Fraser Patterson, CEO and founder of Skillit, about how digitizing skilled trade workers would affect hiring speed and quality, why construction jobs are resistant to AI displacement, and what new data reveals about women entering the trades and pay equity. Practical AI supports recruiters and workers without removing humans from high-stakes decisions.

•Diagnosing labor shortage as an access and data problem
• Founder’s journey from journeyman carpenter to tech builder
• Fuilding data-rich worker profiles and semantic search
• Privacy-preserving eligibility checks via co‑pilot outreach
• Scale and demand for data center and energy projects
• Worker-friendly onboarding designed for deskless realities
• Rigorous screening and trade assessments as skill signals
• Women in trades data, faster ramp, smaller pay gap
• Why construction resists automation and offshoring
• AI for scheduling and sourcing with humans in the loop
• Focus on US now with future global expansion

Transcript

Welcome And Labor Shortage Reality

Roopinder 0:19

Hello and welcome to FODES, the Future of Design and Engineering Software Podcast. My name is Roopinder Tara. On the show, we will have guests that will discuss tools and technology that engineers will find interesting and useful.

Fraser 0:33

Hi . Nice to meet you.

Roopinder 0:34

Nice to meet you. I had not heard of Skillit, I admit, but that was a very interesting subject because you're talking about introducing more women to the skilled labor force. I'm an editor, I architecture, engineering, and construction, and I go to a lot of conferences, and all I hear about is the labor shortage in this business, right? That's all people complain about. A couple of times there have been some women on a panel, and gotta say, they all feel like they're the only ones. So when I heard your that you're studying this problem and maybe even have a solution for it, I was like, I want to hear more.

The Core Problem: Undigitized Workforce

Fraser 1:07

Cool. There's a lot of tailwinds, there's a lot of things happening at once here that I think are enabling this kind of seismic shift in the labor force. We had a slightly different take on what the problem is, if you will, a more correct diagnosis, which is there's definitely a labor shortage, but the problem has been no one's ever digitized this workforce before. And therefore it's really an access problem.

Roopinder 1:28

You mean by digitizing the workforce?

Fraser 1:31

So if I well, if I said to you, go find out about me, you could look up any myriad sources online, including my LinkedIn. You'd learn about me professionally, my past work history, maybe my skills. This is a historically offline demographic and workforce that nobody's ever built the appropriate technology for for them to put their digital professional self out there, which makes them very hard to find and really hard to connect with. And that that's I think correctly diagnosing that problem is where the data starts to flow, and then you get a better understanding of where things are, who's entering, why are they entering the trades, you know, the velocity of you know, you just uh it it opens up this entire kind of universe, if you will, of correctly understanding the labor market and uh the trends and so on and so forth.

Roopinder 2:16

Tell me a little bit more about yourself. What's your background and how did you come across this idea?

Fraser 2:20

So I'm originally Scottish. I'm not sure if you can you can tell. That's where the accent's from.

Roopinder 2:25

Now I can tell.

Building A Data-Rich Talent Engine

Scale, Customers, And Growth Metrics

Fraser 2:27

Save you the cognitive tax of trying to figure it out yourself. Thank you. I basically grew up in the trades. I was a journeyman carpenter for many years. I then became a GC, so of a construction company builder and owner for many years. When I moved to the US, I had a bit of a background in mathematics and academia. I lectured for a while. So I've essentially been in these two worlds most of my life. When I came to New York, I built a vertically integrated construction company, very data-driven, and I was hiring skilled trade workers directly. And what I realized was that they would often not have any relevant information that made finding screening and connecting with them available, and realized that it's a huge blocker because so much of our industries and economy and infrastructure is contingent upon these highly skilled workers being discovered and routed to the right opportunities. I mean, just look at our data center growth. And all of that, I think the bottleneck in solving housing, solving for energy, solving for infrastructure, both in America and globally, is the correct discovery and routing of these skilled workers across all trades, all skill levels and experience levels to where they're needed most. And that's impossible to do. There's no digitization of their skills, experience, preferences, et cetera. So that's really what we've spent the last few years doing is how do you build that engine to bring workers onto our platform, help them build these really data-rich profiles? How do you build the talent engine that you can actually find them? And then how do you connect them when one group's in the office and the other is often in the field or driving? You know, so these are multiple problems to solve, unlock this labor force.

Roopinder 4:08

I assume that you've found and vetted a lot of these skilled workers and you have a database, and this is something that it potential employers could access. Is that how it works?

Fraser 4:18

We're four years old. We've grown the labor supply 7x year on year on average. We're at about 190,000 workers now. We plan to be at a million next year. That would give us about 20% of the labor force. And our customers are the large national GCs and self-performed contractors. So those that are actually doing the hiring the trades directly and responsible for the delivery of data center energy infrastructure projects. We're very lucky to have a lot of the large tier one contractors as our customers. We're growing very quickly throughout the US.

Roopinder 4:53

You mentioned a need for skilled labor for data centers. I heard one time that a data center, a big one, can employ something like 13,000 electricians. Is that that can't be correct? That's there's too many zeros there.

Privacy, Eligibility, And Co‑Pilot Outreach

Fraser 5:07

We typically see large-scale data centers requiring multi-hundred or small one thousandths of each trade of electricians, 700 drywallers, 600 equipment operators. I mean, you know, even those are just those numbers are are really something. I mean, and they're all mobilizing on day one. I mean, you've got hundreds of skilled professionals showing up on the same day to start a massive construction project. It's kind of unprecedented to Rome.

Roopinder 5:35

You hear a lot in the news about documented workers, right? So do you take extra care to make sure your workers are documented?

Fraser 5:44

We have a pretty rigorous onboarding process. We've made it easy for workers to create a profile. It's predominantly using voice, and they can talk to our team as well, but they can build it asynchronously themselves. We capture all the relevant information. One of the things we've been mindful of is to protect the privacy of workers who, for numerous reasons, may not wish to reveal the fact that they're eligible, say, to work in the United States or not. And so what we do is we don't have that information on the profile. But when a customer reaches out, repender we've built a co-pilot is designed for the first time to tap a button on a worker's profile, it reaches out to the worker by text message, which is a nice low pressure way to start that interaction. The recruiter can put questions into this flow, which might be can you lift 50 pounds? Are you eligible to work in the US? Can you pass a drug test? And the worker's free to respond to those. That way we protect the privacy of the worker, but we also capture the needs of the recruiters trying to hire.

Roopinder 6:41

And putting myself in the position of a construction worker. Now, okay. So I need I need a job, or maybe not. Maybe I've got more jobs than I can handle. How do you try to find me and put you put me in your database? What is the recruitment outreach? Like you said you had 20% of the labor force. That's quite a bit.

Fraser 6:59

We're currently at about 1.5%. We're at 190,000 right now. There's about six and a half million workers. We're growing fast enough that we'll have about a million workers in the next six months. That's 20%, which is a milestone.

Roopinder 7:11

I feel like if I'm a construction worker, I'm already in pretty good demand. How do you entice me to come and go and deer debate?

Fraser 7:17

Yeah, you are in demand. That's sure. The subtle difference here is if I'm a craft worker, say I'm a journeyman carpenter or an electrician, quality of the companies and the brands they have, it's not the same as if you work as a knowledge worker and your Google and your Netflix. And you're often expected to continue filling out forms all the time and applying for different roles and travel across town, do an assessment only to find out that you're going to get offered the bottom of the pay band that you were promised. And it's just a very kind of painful process. And what we do is we basically make it possible for you to really easily create a single profile. So bear in mind, fans are a premium in construction, right? Like I might literally be driving or I'm at work or it's cold, or in my case, my fingers are quite literally bigger as a result of doing manual labor for well over a decade. So typing into a screen is quite a deranging experience. So no one's really ever thought of these before, I don't think, as deeply as we have.

Roopinder 8:16

Well, everyone talks about fat fingers, but it's a real thing in construction.

Fraser 8:20

No, it's a real thing. It absolutely calluses, your muscles get a bit bigger, but it just becomes more difficult to be typing onto a small screen. So we've made that easy for them. And then you can tell when you join Skillit that it's for craft. It's for these skilled trade and craft professionals. And they give us the information. It's in a very thoughtful way. And then we can show them these are the best companies for you. This is their ranking. This is the kind of work they're doing. And then they're coming to you. So we are a sourcing platform. We've made it easy for the recruiter to source you, find exactly who they're looking for. So it's quite a revolutionary experience for a construction worker where you're used to this kind of ambiguous, painful, labor-intensive process, and now just make one profile out there, and then opportunities are coming to you, qualified ones.

Women In The Trades: What The Data Shows

Roopinder 9:11

You have that experience, you can relate to them and your staff can relate to them. Having been a craft worker myself lends me an empathy for the pain they go through. And to design something unique for them, I really don't think they've ever been afforded the correct investment. Most of our software is built for people sat behind desktops typing away on keyboards. And we've kind of just been building tools for each other, and then you've got this massive demographic. It's not just construction, deskless industries, which is 80% of all employment. I don't believe we've given enough attention to how you solve their unique requirements when your mobile all when your office is your phone. It's a different employment experience, and it comes with a lot of challenges. And I think the stack for that demographic and that hiring experience is badly broken and arguably has never really been designed from first principle. Tell me more about your vetting process. Is it hard to become part of this Skillit database?

Why Construction Resists Automation

Practical Uses Of AI With Humans In Loop

Competition, Focus, And Future Expansion

Closing And Listener Invitation

Fraser 10:05

The onboarding process itself, if you're doing it asynchronously on your own, takes maybe 20 to 30 minutes. Depends on what information you have available already. It's pretty robust. We typically see 30% of the workers that are completing profiles actually go live. We do have a phone screen with them. We developed a little rubric around does your profile match what our team can determine? And then as you're live on the platform, you're getting customer feedback as well regarding if you're interviewed, if you're hired. We also have proprietary assessments that we developed that. So when I was hiring carpenters myself, I kind of stumbled into this hack, if you will, which was to ask point student later questions. So one of them would be, what were the true dimensions of a two by four? A lot of people think, well, that's probably a trick question, because it's called a two by four. So it's probably two by four. And others will say, like a few would say, well, it's one and a half, three and a half. And I'd say, What why is that? And they'd say, well, because it comes off the mill. So the only way you would know that is if you're a journeyman and you've maybe installed or purchased timber for framing or whatever. And so that is really where Skillit kind of began, was realizing that that data problem could be solved and that it would be valuable. So we have proprietary assessments now across about 40 trades on the platform built by us and our customers. And they're a really good signal for, you know, they use knowledge as a leading indicator of skills. So they're a really good signal if you know your stuff or not. I think eventually we'll get into background checks, the drug test piece that helps our customers, but for now that's really our focus, the data piece. I think this starts with the data itself. How many women are actually in the trades? The US Secretary of Commerce, a while back, her name is Gina Raymondo. They launched this million women in construction initiative. And the goal was for to double the number of women in the trades over the next decade from five to ten percent. It turns out there's a much bigger problem, which is the numbers underpinning that are just the wrong assumptions. And there's far fewer women in the trades than we when we'd originally thought there was. There really there's only 3.4% according to our data. And our data comes from the worker. It doesn't come from a government agency or an employer. But what's really interesting is, and this is why we're very bullish, there's a number of reasons to be bullish about women entering the trades. Despite having less experience than men, they score significantly higher on trade assessments. So the women who are actually in the trades, they seem to be, at least intellectually, understanding the truth a little bit better on average than men. So they have about four and a half years less experience, but they score 27% higher on skilled trade assessments. The other is the broader gender wage gap of about 16.3%. I mean, it really depends which stat you look at, but it's usually high double digits. That's just 5.3% in construction, according to our data. The hurdle's bigger, but the potential is enormous, right? Like there's clearly the aptitude to excel in those roles and they're ramping up faster than their male counterparts. And then once they break into these higher skilled roles, they're actually getting a more equitable future. That's an interesting kind of insight that no one's really talked about because no one's had the data. The other thing that's really interesting is why? Why are skilled trades more equitable from a pay gap standpoint? And what we've come to believe is the bias has nowhere to hide. So physical work has got a much tighter feedback loop and clearer productivity signals, right? Like you either install drywall or you don't. That makes it difficult to rationalize on equal pay for essentially equal output. Knowledge work, I would argue by contrast, is very abstract. It's often measured through perception, meetings, presentations, visibility reviews, and it creates space for that bias to really creep in. And I think that's why the trades are different. And I so I feel like that explains why there's less of a gap. And then I think what's coming next is there is this massive tidal wave of a resurgence in what I would call AI like proof work in point. Construction's never had a foundational technology to solve the problem of finding workers or augmenting like think of robotics. Like think of retail. Mostly it's in a physically static space. So you can use robots a lot and they work quite well. Amazon's a good example of that. Software engineering. You're quite literally building code on an internet connected device. So you can outsource anything in the world. Nursing. About 30% of the job is at so AI LLMs are starting to be used to fill that gap and give nurses more time to patient. You can't do any of that in construction. It's hyper-local work, cannot be outsourced. Robots are nowhere near dexterous enough or economically viable enough to support labor. And no admin work on a construction site. So what good's an LLM? So no one's built a foundational technology to solve that problem. So I think this is one of these rare instances where AI what we're building. I'm not trying to pitch Skillit here. I'm just, I think that what I'm really trying to get at is we're starting to see more and more people move away from knowledge work into roles that they think are much safer because they can't be outsourced, they can't be automated, and they can't be LLM'd away. The correct diagnosis of this problem is that nobody's built this foundational technology to help them connect to the work. So if you take these two things in parallel, you could end up with a lot more people in construction, let alone just women. But I think you'll end up with all tides, rising tide, raises all boats, et cetera. I think we're going to end up with that share growth as the barriers to access drop. And you're also just going to see a lot more people flock to construction. By the way, we've been saying this for a little while, Roopinder, but I just read today the CEO of one of our customers is saying this exact point, how these are becoming future-proof careers. We're even seeing it in the data, another piece of data that we have. And it's not gender specific. We don't have that data yet, but there is a growing number of workers joining Skillit that already have a college degree. So that means they went into the market, probably because they were told college was the path to success. They come out, they get a bit of a punch in the mouse from the labor market because maybe we didn't need as many histors or that we created. And they're now trying to find uh safe pavement. They're trying to find a future-proof pathway, debt-free pathway into the middle class. And I think construction is unequivocally one of the best paths to do that.

Roopinder 16:19

Where are you located, Fraser? Oh, New York. Okay. We are in the Bay Area and we see hundreds of thousands of tech workers being of being laid off here, hundreds of thousands. It's like a, I wouldn't say a cataclysmic event, but it's a big event, how many of there are. And it seems to me they could, but they don't have the hands yet, but they could just go into construction. Construction needs people. I'm sure they can be trained to understand electrical code or carpentry or anything, but they have soft hands. So a little bit of time would take care of that. I'm kind of joking because it's not that easily transferable. You were absolutely right about permanence of these kind of jobs. They can't be outsourced because everybody here is worried about how they're going to be replaced by AI. What I wanted to show you was this ad campaign or this banner that was slung on a building and it was said on it. AI finish this building. Also employment agency, and they were making that same point that you'll never replace skilled workers with AI. You just you simply can't. It's like I thought, well, this is genius. This was a couple of years ago this actually happened.

Fraser 17:22

No, no, no. We were we were around before this, but you know, in seriousness, robotics is making progress, but the reality of automating away the whether when something's technically feasible, it's not enough. It needs to be economically viable, it needs to be culturally adopted. I mean, but just look at just look at where we are with autonomous vehicles.

Roopinder 17:44

Right.

Fraser 17:45

We're supposed to be fully autonomous. We're not even at one percent penetration. I think fully bipedal hominid, dexterous, economically viable, skilled construction workers are like decades out, and humans will be working alongside them, doing the more highly skilled, doing the you know the safety conscious from where I sit, I hear a lot about hominids, and it's really all hype.

Roopinder 18:06

The simplest task takes so long. I'm talking about simple tasks like fetching a water bottle from a fridge, and I'm not talking about even hammering nails into drywall or something like that. They can't do anything worthwhile. So I think not in our lifetime. Do you find use for AI at all? Is it are you using AI to, for example, to vet workers?

Fraser 18:25

When workers onboard to Skillit, they have the option when they're actually onboarding. We have a co-pilot that is able to help them onboard quicker when our customers are searching, they're just prompting, like you do chat GPT, and it will go into the data pull out, their semantic search. And then we also have a co-pilot, which we invented, which is automated outreach, scheduling, rescheduling, call manager. So because it's a lot of work reaching out to workers and trying to schedule calls, and they're out in the field and maybe message at midnight, and then they respond and they never hear from you. So we solved that problem intentionally using AI. That was actually our first use case for AI on the podcast.

Roopinder 19:02

Got it. So a recruiter, somebody who wants workers could use natural language to find someone, and a person who's actually looking for work could use natural language. They wouldn't be talking to a LinkedIn recruiter or something, they'd be talking to AI, which would be probably programmed to be more empathetic or more understanding of their position.

Fraser 19:22

Kind of trending in that direction, but we believe that you really want the recruiter talking to the worker directly. AI should be in service of construction recruiters being able to talk with workers. Because that's what workers want too. They want to know about your culture and what your company's like to work for. The other thing, as well, is I would just say, generally speaking, it would be hard to justify removing the human in the loop for high risk industries. Because the cost of financial and cost of life potentially of hiring the wrong person and not being accountable, I think is a real problem. In my view. That's just not where construction is headed. Yeah. Like some other industries where You can have an AI interview you, just don't see that in construction.

Roopinder 20:03

What are your biggest competitors? I take it it's not indeed, is it LinkedIn or are there already construction-specific job boards out there that you're competing with?

Fraser 20:12

So there is status quo, agencies, this plethora of job boards. The challenge that all of those have is they don't solve for the problem of being a construction company and needing to source talent, highly skilled talent at scale. In any region, across any trade, union, open shop, per diem, traveling worker, no one's built the infrastructure to enable that. And therefore, with an agency, it's very expensive. You're paying a lot of labor to find you that talent. We've already done that work, automated that work of finding workers and helping them create these profiles and automating that process. The other kind of horizontal job boards indeed to suffer the problem of any solution that's across multiple industries. There's only so deep it can go into solving any one industry's problem. Construction is quite unique. You know, there's a lot of best challenges in sourcing talent that these are not designed for. So we're radically more economical and we're also laser focused on solving one industry. And then in the future, once we've built the hiring infrastructure for scaling, for sourcing and screening construction, we want to go downstream more. We want to help the workers train and upskill, essentially improve the quality of the workforce. And then zooming out, big vision is to scale the world's craft. We can imagine ourselves in other industries, automotive facilities, energy manufacturing. Once you build that infrastructure and the means by which you can improve the quality of the become the career engine for the worker, we can move into other industries. And I think that's how I think that's what we're going to find is one of our biggest problems. We talk about becoming multi-planetary, we've got energy, data center. How are you building that if the bottleneck is we have insufficient access to the world's skilled humans, craft workers?

Roopinder 21:59

Are you favoring the US market now?

Fraser 22:01

We're focused on the US for now. We want to make sure that we're doing right by our customers, and that means maximizing the amount of workers on the platform and making the product as robust as possible so they can find and hire the best talent. But we are frequently asked when will we be in Canada, when will we be in Europe, when will we be in Japan? We have customers that are international corporations that are looking to us to be able to expand internationally. So that's definitely in the future of Skillit. For now, we're just going to focus on the US. I mean, it would be beautiful for me. It would be full circle for me.

Roopinder 22:39

So it'll probably be carrying you on your on their shoulders when you come back home. A successful entrepreneur from America. Very nice to meet you, Fraser. Well done. I think you've already built a nice company when and you have a you're solving a critical need. Good for you. Thank you. All right. I'll leave you to your real work. Thanks very much. Bye-bye. Thank you for listening to FoDES, the Future of Design and Engineering Software show, brought to you by ENGtechnica. I hope you have learned of a new application or technology that will help you with your job. If you have an application you think would be of interest to other engineers, please let me know by emailing me at roopinder at engtechnica.com or message me on LinkedIn.