Infrastructure Technology Podcast goes in-depth on transit innovations and AI for safer roads

In the fourth episode of the ITP, Brandon Lewis interviews Stacey Matlen, senior vice president of innovation at the partnership for New York City. They discuss how the lab navigates the "choreographed chaos" of NYC transit, the innovations driving infrastructure and transportation resilience and the challenges of modernizing transit systems. Later, Mark Pittman, founder of Blyncsy and director of transportation AI for Bentley Systems, shares insights with Gavin Jenkins on leveraging AI to improve roadway maintenance and streamline traffic operations. 

Below is a transcript:

GJ: Welcome to the Infrastructure Technology Podcast. I'm Gavin Jenkins, senior managing editor of Roads and Bridges, and with me today, as always, we have Brandon Lewis, the associate editor of Mass Transit and Harlee Hewitt, our jack of all trades, associate editor of Construction Equipment and Roads and Bridges. How are you two doing today?

BL: ITP podcast episode numero quatro is here.

GJ: That is right.

HH: Feeling great on that note. Feeling good.

GJ: Brandon, I got a question to ask you off the top. You are a fan of professional wrestling, are you not?

BL: I am. Yes, I am.

GJ: Tell me a little bit about your love for professional wrestling.

BL: Well, it's a very strange love. I started because my brother and my dad originally started watching it. For those of you that are fans, you know that every Friday night, SmackDown would come on from about 8 to 10:00 p.m. Eastern time and one night on a Friday night, I was just randomly up when I was little, could not go back to sleep. Came out in the living room and they were watching it, and from that point forward, I have fell in love.

GJ: Who's your favorite wrestler?

BL: Oh, I have a lot, but my favorite of all time is Edge, if you've ever heard of him, but I mean, wrestling is so strange because it kind of goes through these waves, right? And it's like you're either a wrestling fan or you're not. You either get it or you don't. It could be the greatest feeling in the world, or it can be something that you never want to watch again. It's a weird love-hate relationship.

GJ: Harlee, what is your relationship? You grew up in Oklahoma. You come from a musical background with you and your dad. What is your relationship with professional wrestling?

HH: Professional wrestling…I have virtually no relationship.

GJ: You know what it is though?

HH: I know what it is. That's the very basis of what I know though.

GJ: So my relationship to it is I grew up in the 80s, I’m 44 years old. So, I fell in love with Hulk Hogan in the mid- 80s, but then I grew out of it after one or two years. But I do have lots of pictures of me wearing Hulkster outfits. So we have the full spectrum of loves it, loved it for a little bit and has no relationship it. So, what is your thought then, Harlee, of when you see a group of men in spandex, tights in a ring, all pretending to hit each other?

HH: I think different strokes for different folks because I have loved other sports, but that is just not one. I can't understand the general appeal. I just can't. But I do respect, I respect other people's wishes to watch and embed themselves in that because people are very passionate.

GJ: Yeah, it is a great tightrope you just walked. I congratulate you.

HH: It's a very diplomatic answer.

GJ: Yeah. So the reason why I bring this up is because when I think of professional wrestling, I think of hype men like our Brandon Lewis

HH: Oh, I see what you were doing.

GJ: Well, hold on. Also, I think of the Royal Rumble, and I think of just dudes just running into the ring, and there's chaos. However, it's choreographed. It is choreographed chaos, which leads me to our first interview today with Stacey Matlen, who works with the New York Transit Tech Lab in New York City. And anyone who's ever been to New York City knows that transit and New York City is chaotic. However, things do tend to be slightly on time. You get where you're going. And so it is a choreographed chaos, to at least me, the outsider who is from Pittsburgh and not used to it. And Stacy here will get us more into that and how the Transit Tech Lab is working with the testing processes. And Brandon, why don't you tell us just a little bit more about Stacy as we go into this?

BL: Well, Gavin, let me tell you something brother! Stacy Matlen, she is the senior vice president for the Innovation and Partnership in New York City, and she talks about the birth of the Transit Tech Lab, why it was created and obviously helping all of the transit agencies in New York, whether it's the MTA, whether it's the Port Authority of New York and New Jersey. She talks about how they are using projects that deal with technology to better advance their transportation systems, and just a little bit of inside baseball for this interview today, she talks about the fact that in 2018 for transportation in New York specifically, it was the "Summer of Hell", and that's why the Transit Tech Lab was born.

GJ: Oh wow. That is awesome. Our hype man never disappoints. Never disappoints. That is awesome. Alright, well, let's get into this. Brandon Lewis' interview with Stacey Matlen.

BL: And welcome back to the Infrastructure Technology Podcast. I am Brandon Lewis, associate editor of Mass Transit magazine, and I am here today with Stacy Matlen, senior vice president, Innovation Partnership for New York City. Stacy, welcome to the ITP.

SM: Thank you so much for having me.

BL: Thank you. So we are here today to talk about your role as the partnership for NYC and how transit agency intersect with this role. So can you talk to me a little bit about your role?

SM: Sure. So I oversee our public sector innovation programs here at the Partnership for New York City, and for background, the partnership for New York City is a nonprofit that is a civically led organization representing the business leaders in New York. Our members are comprised of the CEOs of the 300 largest organizations in the city with a mission to make New York the global hub of innovation, commerce and upward mobility. So the reason we've been involved in public transit since our inception is because we recognize business leaders as well, recognized that public transit is the lifeblood of New York City's economy, and in order to have a thriving economy, we must have a thriving public transit system. We more deeply got involved in the public transit space back in 2018 when at the time it was the “Summer of Hell”, and it was really a crisis with the subway systems where the governor of New York at the time called a state of emergency because there was such a bad crisis around the subways, derailments and a lot of bad issues that came up. So we worked with the MTA at the time to develop a partnership to understand how we can leverage the business community, their expertise and really understand how we can help modernize MTA. So one of the outputs of that convening was a program called the Transit Tech Lab, which is an annual process. It's copied off our innovation model that we created in the private sector with the big banks called the FinTech Innovation Lab, and it's been proven to the private sector, so we transported it to the public sector with the MTA, and we developed this model to help quickly identify new technologies that could solve pressing challenges with the MTA.

BL: You're not only working with the MTA, but you're working with other transit agencies as well, correct?

SM: Yes. In the first year, we worked just with the MTA to demonstrate this technology and the process, and then in following years we’ve now have been joined by all of the regional New York agencies and that includes the Port Authority of New York and New Jersey, New York City Department of Transportation and New Jersey Transit.

BL: So the Transit Tech Lab in May of this year selected 19 companies to collaborate with the new technologies, and these technologies are part of the Customer Experience, Resilience and  Curb Activity Challenges. Can you go through each of these challenges and talk a little bit specifically about each one?

SM: Yes, would be happy to. I'm first going to take a step back to just share a little bit more context around how the program operates because we don't identify and launch these challenges in a vacuum. Part of the reason our model works is because we have really deep conversations with both operational staff and agency leadership to identify their top challenges on an annual basis. So for the 2024 challenges, we spoke with 92 agency staff from all four regional agencies to identify their challenges and then we requested the CEO and presidents of these organizations to prioritize those challenges and select the ones that they would like to go to market to recruit companies from around the world to apply. So we have a four phase process where it's identifying those challenges, going out to market to recruiting companies to apply and the third step is the evaluation process, where we have pitch presentation and product demonstrations, and the fourth stage is a testing process. So going back to this 2024 challenges, after we met with the regional agencies, their staff and their leaders, they identified three focus areas for 2024 priorities, and that is, as you mentioned, Customer Experience, with the question, "How can we improve customer's experience to better communicate service changes, reduce delays and augment safety and cleanliness initiatives?" The second one is Resilience, so, "How can we build a more resilient and adaptive transit system?" And the third one, in partnership with New York City DOT specifically to help them identify technologies for their Curb Management Action Plan was the Curb Activity Challenge, which asks, "How can we maximize the city's curb space to serve the multiple and varied needs of New Yorkers?" So those were all launched as an open challenge competition in January. From there, we recruited multiple companies to apply, and we had 150 applications from around the world who submitted applications. And then from there, we had agency staff review those applications, participate in product pitch presentations and product demonstrations, and from there, 18 companies move forward with an eight-week proof of concept, which is what we call a minimally viable test, to really understand if the technology is viable and valuable in the context of the system. So just a couple of outcomes that came through those tests. The technologies that tested were everything from environmental sensors for the resilience challenge that could identify when there's extreme heating and flooding to notify staff ahead of time. There's sensors on bridges to help with proactive maintenance on bridges. There was a lamppost that could be researched to identify and understand how lampposts could be retrofitted with EV charging capabilities in a cost efficient manner. So those are just some of the examples and ideas for the Resilience, really, both from a climate resilience perspective, and organizational resilience perspective of how are we making our transit agencies as strong as possible, and especially in lieu of climate change, and make it more adaptive and responsive.

BL: So once these ideas finish the eight weeks of testing, what's then the process of implementing them?

SM: That's a great question. So our testing phase is broken into two separate phases. The first is an eight-week proof of concept, and then there's a decision point on whether or not the technology needs to be further tested to be adopted or if there's additional learnings. So this summer, we had final presentations where those 18 companies came to New York to demonstrate what they were able to accomplish over the eight weeks, and from there, we had agencies select companies to move forward with a yearlong pilot, and that's the second phase of the testing process. So through that second phase or in the process now of working through which companies will win a year-long pilot contract for further testing, and we should have an announcement hopefully by the time this podcast is live.

BL: Now, when you take a look at these technologies in particular, how do they scale? Can you share some specific success stories demonstrating how that technology can result in positive impacts for actual riders?

SM: Yes, so we have this two phase testing process because especially when it comes to new technologies, it's very difficult for anybody, let alone government, to have their pulse on how these technologies can scale and what the value is. So we have this small eight-week proof of concept, followed by a year-long pilot, and then from there, it's up to the agencies to procure the solution if it makes sense for them or go out to bid with a competitive RFP. So an example, a couple of success stories and examples of how this technology is scaled, especially for riders is the first year of the Transit Tech Lab, one of the challenges asked the question, "How can we speed up buses?" And originally the thought was, "Oh, we just need cameras in to identify cars parked in bus only lanes." And that was the thinking. However, the challenge opened up a wide variety of different types of solutions that could help speed up buses. A couple of examples include a predictive bus maintenance software that can identify when buses are about to break down 48 hours before the check engine light turns on and so, if you don't have failing buses out on the road, then buses will be faster, and the company who makes that technology made a big announcement that shared they have been commercially scaled to 25 percent of the New York City Transit bus fleet. So they went through our process. They did a demonstration and then they've now been commercially scaled.

BL: Is there certain technology that you're looking for, whether it's for buses, for rail, new applications, mobile apps, is it just sort of a whatever they come up with type of deal?

SM: Yeah, it really depends. So we prioritize the challenges, and we are a challenge-based process, which is really valuable because the agencies aren't necessarily looking for a specific solution because there's a variety of solutions that could solve that challenge. We're really opening up the field to entrepreneurs to come and tell the agencies how they think they can solve their challenges, and from there, the agencies will decide. So it can be anything from software to hardware to, it really runs the gamut of what the types of technologies could look like.

BL: Now you mentioned entrepreneurs. Are those entrepreneurs, I guess the question would be, do they specifically have to have a transit background?

SM: No. Actually a lot of the companies that have been successful, we have an example called Runwise. They are a New York-based company that monitors heating controls and buildings, and they were able to demonstrate with both the MTA and the Port Authority that by installing sensors and boiler rooms and throughout the building and just by controlling boilers, by turning them off when the building's not needed to be heated, they can reduce overheating and building by up to 40 percent in certain cases. And so they've scaled their technology, so that's an example of a company that had never previously worked with a transit agency, but they were solving a critical transit agency challenge by saving operational costs and reducing carbon emissions in their facilities.

BL: Well, speaking of working with transit agencies, what sort of technology, has there been a common theme of this specifically works in the transit agency environment, even if it's not necessarily for a transit agency?

SM: I think it's not the type of technology that is what works and doesn't, but it’s the ethos of what the company is able to do and their responsiveness to work in partnership with the transit agencies, and in order for the testing process to really work, and where we found success is, when the agency themselves embrace the technology, they make sure that the project manager is embedded with the agency, and they can really identify the challenges that need to be addressed by that technology, and it's supported by executive leadership. So it's a full process to have this change management to try something new. It doesn't matter what the technology is, it's the trial and error and making sure that the technology solves a real problem, and it can be fully adopted at both the operational level and the leadership level are really important.

BL: Now, as we know, technology is always constantly evolving. Is there a scenario where all this technology gets tested, it gets greenlit, it goes to potentially be implemented in a transit agency and then something changes in the technology space, and it may have to get retested?

SM: Yes, that always happens with technology tests, where it's a constant trial and error and a constant understanding of what needs to happen, what needs to change, and that's the value of this iterative testing process that we've created because there is no, "Oh, there's a technology solution, it solved everything". It's always a constant tinkering, always a constant of iteration, and that sort of mindset is really important for transit agencies to have as they go into the 21st century and continue to digitize their assets and tools.

BL: You mentioned the Transit Tech Lab is about six years old now. How has the lab grown from its time and where do you think it could still potentially grow in the future?

SM: Every year it grows. In 2024, we engaged over 100 agency staff and leaders throughout the process. We had a record number of companies that were selected for proof of concepts. We had 18 companies this year compared to 15 companies the previous year, and I think in the first year we had four companies. So every year it grows and not only has it grown within the transit agency field, and we've continued to get more technology tests and get more agency staff involved, but this model has sparked innovation to other sectors. So last year we launched a program with New York City Department of Environmental Protection to basically copy this model and demonstrate how our process can help the city's water and wastewater utility adapt new technologies and climate change. And then this year in July, we also announced a third program working with New York City Department of Buildings to understand how this process can help DOB, which is the country's largest municipal construction regulator, identify technologies to improve their internal processes.

BL: Now, in terms of, you had said that even though you guys are based in New York, the Transit Tech Lab is working with companies worldwide. Is there a difference in the type of technology that's being developed or maybe used for in U.S. North American transit markets compared to maybe international?

SM: I am not sure if I've seen that. We have conversations with folks from London and other countries around the world to understand what the technologies they use there, and every agency has their own way of working. Every agency has their own challenges and their own technologies that they use, and we do our best to keep tabs on what other technologies are being tested, so we can recruit them to come to New York and work with our agencies here.

BL: Now, you talked about the model that the Transit Tech Lab has used, and obviously, expanding that model to other businesses in New York, but if another agency in another state decided to use this model, what can other transit agencies learn from the Transit Tech Lab model?

SM: Yeah, I think there's three critical pieces that could be scaled across any transit agency. I think the first piece is that we have a reverse procurement model. So instead of being incredibly prescriptive and saying, "Yes, I know I need 21 buses that do X, Y and Z", opening up that process to a challenge-based model where you're just sharing broadly, you're defining your challenges and then you're sharing them broadly and recruiting companies to apply. The second piece is around strong project management and strong goals and KPIs that are set from the beginning. My background is I used to work in government. I worked with the city of Detroit's Office of Mobility Innovation and worked with the city's transit agency, Detroit Department of Transportation, and when there are so many things and priorities to be done, it can be very difficult to manage all the projects and have those strong goals and KPIs and strong project management background. And so that leads me to my third part. There's a strong third party nonprofit that can help manage it or help track those and hold everybody accountable is very helpful, and so to the extent that you can have a partner do that, that's great, but if you're not able to have a third party nonprofit, do this on your behalf to have somebody who's trained in project management and to really focus on the goals and KPIs and the follow-ups and making sure those don't fall through the cracks because government's doing a lot of different things all at once, and it can be very overwhelming, and in order for technology to be adopted, you do need that sort of really focus on it.

BL: We covered a lot of things today. We covered a lot of different angles. Is there anything that you think that we did not cover?

SM: Check out our report. It's on transitinnovation.org. You can see our report that we have tangible results from all of these 18 proof of concepts, so if you're a transit agency that's looking to solve critical Climate Resilience, Customer Experience or Curb Activity Challenges, you can check out the results and look at the companies, and we're happy to make introductions to any of those companies. Also be on the lookout because in January 2025, we are launching a new set of challenges, and so, if there are any startups or entrepreneurs listening, we would love for you to check out. Also, go to transitinnovation.org to check out our open challenges, and we would love to have you apply.

BL: Stacey Matlen, senior vice president, Innovation Partnerships for New York City. Stacey, thank you for joining me today on the Infrastructure Technology Podcast.

SM: Thank you, Brandon.

GJ: Welcome back. That was Brandon Lewis' interview with Stacy Matlen. Harlee, what did you think of that interview?

HH: Yeah, so I thought it was really interesting how the Transit Tech Lab approaches their testing process for all of these different agencies, and as Brandon mentioned in the beginning, there is a lot of them in New York, and I could imagine they all require very specific sets of things, so it makes sense why they have such a long process, but I especially thought it was interesting how that almost directly ties into our next guest and everything he talks about with the company he founded. Blyncsy, and how it could easily be a company that the Transit Tech Lab works with. So Gavin, if you want to go ahead and introduce Mark.

GJ: Oh yeah, absolutely. So Mark Pittman is the founder and CEO of Blyncsy, and in 2023, Bentley Systems acquired Blyncsy, and he is now also the director of transportation AI at Bentley Systems, and let me tell you a little bit more about Mark. He founded this company while sitting at a traffic light in 2014, and he is just a passionate guy. He's passionate about improving the world through technology and innovation. Prior to this position, he worked in political campaigns, law firms, startups and even one Fortune 500 company, and he's just a really energetic, positive guy, and he knows what he's talking about. And Blyncsy is an AI power, real-time roadway, condition assessments and asset inventory and that is a mouthful, but it's something that every DOT wants and needs to work with, and it's Bentley, so it's excellent technology. So without further ado, here's my interview with Mark Pittman. 

GJ: Well, Mark Pittman, welcome to ITP. Thank you so much for joining us today. 

MP: Thrilled to be here. Thanks for having me. 

GJ: Alright, let's start out with Blyncsy. This is a company that you founded in 2015. What is the history of the company? How did it get started?

MP: So I was in law school back in 2014, and I got stuck at a traffic light one day, and frankly, I didn't really want to be a lawyer. The practice of law is pretty harsh, and I got stuck in traffic light, and I thought to myself, "How does traffic signal timing work? How do we manage our infrastructure?" And for the first time in my life, I questioned the physical infrastructure around me that honestly had taken for granted my entire life. I called the folks at Utah DOT and I said, "Hey, can you tell me about what's going on"? And they invited me to their offices, and I got hooked. And so we set out to build a big data intelligence company that was venture backed and wanted to make an impact in the transportation world.

GJ: UDOT were just like, come on down?

MP: Totally. They were so open. I also talked to Salt Lake City. Everyone invited me in. They like to say people used to call with complaints, not with questions. So come on down, let's talk about it. We're excited to tell you about the every day.

GJ: Those DOTs are really friendly, especially that one out in Utah. That's awesome. Okay, so that's what inspired your passion for technology?

MP: That's exactly right. So I worked previously, I got my MBA as well at the University of Utah while I was getting my law degree. I did my internship at IBM and was focused on big super computing IBM Watson at the time. I got into the entrepreneurial program at the University of Utah, started working on university technologies and all these things combined. I have background in political campaigns. I'm really interested in the physical infrastructure, worked in supercomputing, worked on startups, all those things came together into this perfect little business that we call Blyncsy. And we were off for the races.

GJ: Why is it called Blyncsy?
MP: So we were looking for a name right, and.com is really hard to come by anymore. And so a friend of mine said, "Let's just throw everything into a word scrambler." He is like, "What do you want to do?" And I was like, "How about we synchronize the lights?" And so he is like ‘blink, light, sink, let's throw it into a word scrambler and see what comes out’. Blyncsy came out, there's a 'Y' in there. It's a couple 'Y's in there which throw people off, and I like to remind everyone that sometimes 'Y' is also a vowel. And we had a web domain and a trademark available for Blyncsy and now it's been around too long to change.

GJ: That's great. That's great. Okay, and let's dive into why you're here now. So Blyncsy just released a roadmap that details pavement marking reflectivity and a lot of it was done with artificial intelligence. This technology collects street level imagery for over 800,000 vehicles nationwide, and it assesses the condition of assets in the images in as little as 60 seconds of a vehicle passing. Tell us how this AI technology works and what is the benefit of this map?

MP: Yeah, absolutely. So every day, delivery drivers are out driving the nation’s roads. Semi-truck, last mile delivery vehicles that probably that vehicle shows up at your front door too often, and they've all got cameras on them, right?! And their cameras are in these vehicles for their own purposes. Driver safety monitoring, insurance purposes, federal motor carriers compliance or you put these cameras in to help understand the world, and these guys are paying a really high cellular bill, so we went to them and said, "Hey, can we buy that imagery from you, so we can subsidize the cellular connectivity you've got and use this imagery to help the state DOTs in our cities and communities use that data to understand what's wrong on their roads and how to fix it to make the roads safer"? We see this as this really unique symbiotic relationship, right, where vehicles are effectively talking to infrastructure through our technology, the cars on the roads, the users of the system are reporting to us issues and we're telling the DOTs how to tackle that. It creates this beautiful symbiotic system that's really this vision of card infrastructure technology that the industry has been talking about for decades.

GJ: That's really fascinating. And the map, it's a public map, so anyone can access it?

MP: Totally. So we really wanted to democratize the technology. We have a big belief that we should open source and make data available whenever and wherever possible, so it's also a core part of our business model. We don't charge user licenses. We don't charge seats. We provide a license to the entire agency, and so we did the same thing here. We took a public map that we collected from our fleet of vehicles across the entire country. We collected nighttime imagery, we applied our nighttime correlated retro reflectivity analysis to this and we said, "Which of the paint lines out there on the roads that we've collected imagery on would pass or fail the federal guidelines if they were to take effect today?" And the idea here is to provide a really inexpensive, scalable tool to everyone. Now, this is a massive shift in the way things are done previously. For the first time, the federal government, the FHWA, through the MUTC, the Manual of Uniform Traffic Control devices, has set a new standard that starts in 2026, and they've said, ‘Hey, all your paint lines have to reach a minimum visibility level". That means we as humans have to see enough light from those paint lines at night. That's called retro reflectivity, and we said, "Alright, so you're going to send someone out with a retro reflectometer handheld device that costs about $30,000, and they're going to scan every paint line in America?" Well, that's not going to work, right? Think about the cost, think about the burden, think about the risks. And so we said we think we've got a better solution. So, using crowdsource dash game inventory technology, we can understand the exposure on the camera lens. We know how much light is coming in. We can see how much reflection there is off the paint lines from the headlights of the vehicles out there and use that to identify how visible the paint lines are at night, how reflective they are. We can correlate that to the scores that are out there and determine a minimum pass fail. So we published on our website a simple map, red and green line strings to tell you if a certain segment would pass or fail those rules if they were take effect today to demonstrate what this would look like at scale across the entire country.

GJ: I want to ask a question before we move on. What is the relationship with Blyncsy and Bentley so our listeners know and understand that?

MP: Yeah, so just over a year ago probably, Bentley Systems acquired Blyncsy. We operate as a business unit of Bentley systems today, and I have a privilege to continue to lead that business unit, as well as lead initiatives at Bentley as director of Transportation AI.

GJ: Okay. Alright, so I imagine that this project really helps out some other Bentley's products and their missions in general of helping America reshape infrastructure and roads.

MP: Totally. I mean, look at Bentley. Our mission is to accelerate transportation, and we really focus on that from a data perspective and a software perspective. We just celebrated Bentley's 40-year anniversary this week and really the crux of what the Bentley Brothers built 40 years ago and what continues today is an amalgamation of over 80 different product lines in all the areas of critical infrastructure. And really our focus historically had been on the design and build part how design roads, bridges and infrastructure, and now with Blyncsy, we're moving into the operate and maintain area of the business. If you look at a DOT's budget or city's budget, right? They got to build the roads, that's the construction. They got to manage the roads, that's traffic signal operations. And then, they have to maintain the roads, that's pothole filling at paint lines. With Blyncsy, we have a complete lifecycle that fits really naturally into Bentley's architecture and their vision of a transportation system, and so we're proud to be an AI accelerating part of this product strategy.

GJ: Okay. Let's dive back into artificial intelligence then. Is the AI used for the public map project, something transportation agencies can leverage themselves? And if so, what would be the benefits to those agencies of implementing AI?

MP: Yeah, absolutely. So a couple of really cool things, right? When Bentley acquired Blyncsy, we were six employees, and we were delivering applications across the entire country. So just to give you a sense of scale of how technology leverages humans, we are a great representation of that. We are literally operating from coast to coast across the entire country, delivering nationwide data sets a year ago with six employees. The only way that's possible is because the scalability and the power of artificial intelligence, and we're looking to bring that same technology to our end users. It's harder than ever before to retain and recruit employees. At the DOT level, we have more roadway miles, everything's more expensive and we have fewer dollars. How do we help the DOT stretch their budgets to cover more things? We use technologies that augment them. This is a perfect example of that, just like voice dictation makes it easier than you having to hand write something on paper. We want to apply this technology to help scale these applications. So AI plays a really crucial role here. We are specifically focused on what's called narrow AI. So the specific application using computer vision technology, where we analyze every single pixel for tens of millions of images every day, identify issues and problems on the DOT's, roadways and make sense of them in near real time.

GJ: You just touched upon an interesting fact. You said that DOTs are having trouble hiring and retaining employees, and as you're talking about all this, I'm thinking, "Man, send this guy into some high schools, have them talk to some teenagers that love playing with AI and they'll be entering our industry." We need to get you on the road.

MP: Totally. Well, I spend a lot of time on the road actually, having these conversations nationally because part of our job is to educate our end users to tell the DOT's what's even out there. There's so much to do day in, day out. That's hard for them to even get a sense of what's going on, and there's still much noise in this space, right? We are presenting ourselves as the, call it the chatGPT of the transportation industry, but there's six new AI vendors every day. I mean, the news is constantly telling us about all these new things going on. It's hard to get a sense of what's going on. It's hard to have confidence in knowing what's going on, and this is a really unique application of the Blyncsy Bentley partnership that comes to the table. Fourty years of entrenched relationships. Our core products designed build and maintain the roadway infrastructure across the entire country. The majority of the interstate highway network was designed using Bentley software. They trust us, they know us and we get to come to the table as the trusted AI partner on top of that.

GJ: Wow. Yeah, couldn't agree more on that about how important all this is. I didn't know that about the percentage of roadways that were used designing Bentley. That's pretty fascinating. Are there currently any roadblocks to the adoption of AI that you're seeing agencies run into?

MP: Yeah, I mean most of it is your traditional technology adoption cycle. It's understanding, it's confidence, it's fear. We have a lot of customers that come to us and say, "I don't even understand what this is. Can you break it down?" So I oftentimes give presentations just trying to explain the concept of AI, and this AI goes back into the early 1900s, right? The concept and framework of this. Since we've had computers, we've used machine vision and really it's all part of the same application. Machine learning, AI, are fairly interoperable at these stages. Today, many companies are operating at the bleeding edge of generative AI that's actually creating works more like a human would, but we're still far away from what's called general AI, the supercomputers that you picture in SY-FI movies. We're still a ways away from that, right? Today, we're focused on tactical execution options, and we are really spend a lot of time explaining that this is in the black box. It's not AI making some random decision for you that you have no control and insight in. We walk through the specific applications, we detect a piece of guardrail, we inspect how much damage is on that guardrail. We tell you the likely severity of that damage based on the extent of it, and we report up to the DOT, "Hey, there's a piece of the guardrail here that's severely damaged. You might want to take a look at it". So, we're using this to augment humans. You don't have to get in the truck, go out in the field, make it really easy to integrate to their lifecycle, into their existing tools.

GJ: That would be really helpful for some DOTs. That's really cool. Are there any technologies though, that involve AI that you're working with in your roles at Blyncsy and Bentley Systems that you're excited about in the future? 

MP: Yeah, well hopefully everyone will tune into Bentley's year end infrastructure event happening in just a few weeks. At the beginning of November, I believe November 8th and 9th, which will be in Vancouver and also live streamed. We'll be announcing some really big new data partnerships that we will take things to the next level. So excited to double click on that when the time comes. But until then, we're really focused on is using cloud computing to view this and leverage these things at scale, so we're exploring the use of synthetic imagery and generative AI to train models. So in the United States, there are over 1,200 MUTC signs under the federal library, so the feds have said there's 1,200 different signs you could put up on our roads. In addition to that, almost every state has their own supplemental signs. Some states as many as 400 signs, so leverage 400 times 50, plus the 1,200 on top of it, and then take that problem internationally. We've now kicked off international sales globally and now apply that problem to the whole world. How will I ever teach a machine learning algorithm? what a sign looks like in every country in the world? On my own, right? it's nearly impossible to do. So what we're doing is leveraging generative AI technologies and synthetic sign training applications to scale that technology, understanding what does a stop sign look like in eight different countries to teach a computer model what looks like a stop sign in the U.S. might not look like a stop sign in Saudi Arabia, for example.
GJ: That's really cool. That is really fascinating as well. One thing I've noticed from my travels is the big green highway sign pretty universal.

MP: There are a lot of signs that you see all over the place, right? And generally, this is what's really cool about this technology. Imagine that you are an alien, and you're dropped on anywhere in urban environment in the world. If you stand at an intersection long enough, you through context can understand what's happening. You'll see, okay, when people pull up to these big white lines and the light's red, that means they stop. When it turns green, they go. When there's this big stop sign with the word stop on it, it's big and red and has a distinct shape, people come to a stop and then they continue. But when there's a triangle, they slow down and continue. This is learning by context, and we are exploring that as the next future in generation of Blyncsy technologies. This is the same problem autonomous vehicles have, right? Elon Musk and others have talked a lot about the problem with self-driving is really difficulty, and the difficulty is not putting in boxes around stop signs and saying, "This is a stop sign." The problem is extrapolating this to the whole world. So context-based learning is something that AI enables that we as humans do is completely unique to how we think, and now we're looking at exploring ways to allow to computers to learn through context.

GJ: That's really cool. That is really cool. I wish that—forget aliens—I wish my dog could sit at an intersection and figure things out. Tries to run into traffic while we're walking all the time. So we'll get you out of here on this question: What do you see as the future and potential of AI in the transportation and infrastructure space?

MP: Yeah, so much of the leverage future is going to be in the design space and then the operational space, so we're very focused on maintenance generally, but you've got to think about the lifecycle of a DOT. Generally 30 percent of all DOT funds are spent on construction, 10 percent on operations, and 65 percent or so on maintenance, right? We're really focused on the maintenance category, but the construction part is huge, so how are we leveraging and finding new ways to digitize and automate using AI design technologies? How can you drag and drop bridges, right? Think sim city style for roads and bridges and sewers and transmission lines in the middle of electrification of everything. So that's what's going to happen very quickly. We're going to see technologies arise from Bentley Systems and probably from others specifically looking at those applications. The other thing we're going to start to see is we're going to see the power of workforce augmentation. You're going to see one person really punching above their weight class in the DOT space because of the tools that they've been enabled for them. That's changing quickly. It used to be the DOTs kind of got everything last, and now they're at the forefront of these technology revolutions. They have some, oftentimes the largest budget of any state agency, the most responsibility and it's the most tangible thing that you as a consumer or as a taxpayer will ever touch. So it's critical that they have all services within the agency, within an estate agency have access to resources like AI.

GJ: Awesome. Very cool. Well, thank you for your time. We appreciate you joining us here on the ITP, and we hope that you can come back and join us again soon and talk more about this. This is a subject that is only going to continue to evolve, and I imagine you'll be working with more and more DOT's across the country.

MP: Yeah. Appreciate it. Thanks for the time.

GJ: Alright, well thank you.

HH: Alright, welcome back. So if you're hearing my voice, that means you made it to the end of the episode. Thank you very much, and I'm going to give you an opportunity to hear from Gavin and Brandon here in a second about their thoughts on this episode, but I thought that if I'm offering my two cents, this was a very interesting episode. It was really, really fascinating to hear from Stacy from the Transit Tech Lab of NYC, which like we said in the beginning, is such a organized chaos, so there's probably a lot to gain there for agencies across the country and how they're handling that multidisciplinary work. And then obviously from Mark Pittman and all of his insights and how he got started and how he is basically just trying to solve a problem that all of us face every day, which is dealing with traffic. And he is, as Gavin mentioned, very passionate about that, so it was great to hear from them. Gavin, Brandon, give me your thoughts on this episode.

BL: Yeah, I thought it was very interesting how much he talked about data because I think no matter what industry now, transportation, non-transportation, I think we all just want data. We want everything at our fingertips, and I think he talks about the ability of cars on the road, reporting issues and then basically telling the state DOT's how to fix 'em and obviously open source and making that data available open, because we all know that obviously the less crashes on the road, the better, the safer we all are. But I think the more data you have in anything, no matter what business it is, the better decision making that you can make. And so it's not just collecting data, it's getting that data and then how to use the data that is generated by all these technological elements that we may not even know how they are getting this data or where this data comes from. But using it to better society is what our land is going to look like in the next years to come. 

GJ: Yeah, absolutely, and I think with my interview with Mark and Blyncsy and Bentley, it's all about the AI aspect and what artificial intelligence, and this is breakthrough technology that Blyncsy created and now it's being expanded at Bentley, but it's breakthrough technology and it's going to help DOT's across the country perform maintenance on our roadway networks. And Harlee, did you notice I say across the country?

HH: I did, and I was just going to tell you that after we finished, but since we're bringing up here, just for some context for folks, every time I write news, almost every time, I somehow managed to include the phrase "across the country", earning me the title, "Harlee Across The Country Hewitt." 

GJ: It's now seeped into my head. 

HH: So you're going to start doing it now.

BL: See, I am known as "Across the U.S." at Mass Transit because we not only cover U.S., but we cover North America as well, Canada. Every time we're dealing with a transit agency or anything like that, I always make sure to say across the U.S., across the country, we do have that Canadian aspect to it.

HH: Okay, well you have utility for it, I guess. Me not so much. We cover the U.S.

GJ: That is a little inside baseball of how we edit our content, our written work, but in terms of editing our audio here today, we have to thank Ryan Curtiss, who is our fellow co-producer of this show. He edits the audio and none of this is possible without Endeavor Business Media, our parent company. So, if you listed this far, we thank you, and also, we want you to check out Endeavor Business Media, a great company with other great brands, but the two greatest brands, of course, Roads and Bridges, and Mass Transit. Thank you for listening, and we will be back next episode, and we'll be talking about the state of Michigan. See you next time.