From the DMV Line to the Governor's Office: The Pine Lakers on Building AI for Real Government Problems
Season 3, Episode 36 of Kinwise Conversations · Hit play or read the transcript
From the DMV Line to the Governor's Office: The Pine Lakers on Building AI for Real Government Problems
Episode Summary
What happens when a student's frustrating trip to the DMV turns into a meeting with the governor's office? In this fourth episode of Kinwise's special series on the NC AI Solve-a-Thon, host Lydia Kumar sits down with the Pine Lakers: Aankan Das, Gillian Keith, Prakyath Shankar, and Cooper Brown, juniors and seniors at Pine Lake Preparatory School in Mooresville, North Carolina. Second-place finishers in North Carolina's first-ever AI Solve-a-Thon, the team built an AI-powered app designed to make DMV visits faster, clearer, and less frustrating. What started as what one student called "a little resume booster" has since taken them to the Lake Norman Chamber of Commerce, Good Morning Lake Norman, and the governor's office, where they're now in ongoing conversations about how their ideas could shape a statewide initiative. This conversation is about what happens when students are given a real, messy, bureaucratic problem and the tools to actually take a swing at it.
Key Takeaways for Superintendents, K-12 Leaders & Mission-Driven Educators
A lived frustration is a legitimate design brief. Prakyath's four-hour DMV visit, made worse by a wasted trip for missing paperwork, became the spark for the entire project. The team didn't need a contrived prompt; they needed a bad afternoon and a teacher who said, "let's do something about that."
Vibe coding has a low floor and a high ceiling. The Pine Lakers started with Playlab, a beginner-friendly platform built for the Solve-a-Thon, then leveled up to Base44 as their ambitions grew. Base44 even recommends which AI model fits which task, building model literacy directly into the tool.
Working with a government agency requires a different kind of patience. As the only Solve-a-Thon team tackling a state agency, the Pine Lakers learned that "shipping" doesn't always mean a public launch. Instead, they're feeding ideas, code segments, and feedback directly into the DMV's own modernization efforts.
Banning AI doesn't make students safer, it makes cheating more appealing. Gillian's framing was direct: when AI is treated as the enemy, students are pushed toward using it the wrong way. Pine Lake Prep's leveled system, rating appropriate AI use from 1 to 4 depending on the assignment, gives students a framework instead of a blanket "no."
One teacher can change four trajectories. Every student on this team credited their computer science teacher, Coral Riley, with opening doors, from the Solve-a-Thon to the AIM Conference to DECA. Her approach: give students independence, a creative starting point, and consistent check-ins on how they're doing as people, not just students.
These students are already thinking like policy advisors. From AI hallucination in their own chatbot to the future of work to data privacy at a government agency, the Pine Lakers brought a level of nuance to this conversation that wouldn't be out of place in a leadership meeting.
Where Did the Idea Come From?
Lydia: How did you all come to the DMV as the problem you wanted to focus on?
Prakyath: A few weeks before this project, I went to the DMV to get my limited license. I forgot some documents, so they sent me back, and I had to go again. I had to wait more than four hours, which was really inefficient. I didn't like that, so that sparked my interest in wanting to solve that specific problem.
Lydia: When I got my license, I also waited at the DMV for about four hours, and it felt like a torture chamber. Did you bring this problem to the group, and did everyone else get excited too?
Cooper: Yeah. We had a few ideas, and through brainstorming we landed on Prakyath's idea because it resonated with all of us. Between getting our learner's permits and now working on after-nines, all of us have been in and out of the DMV recently.
"I had to wait more than four hours, which was kind of inefficient. I didn't like that. So that just sparked my interest in wanting to solve that specific DMV problem." — Prakyath Shankar, junior, Pine Lake Preparatory School
Lydia: You first did this for the NC AI Solve-a-Thon. Was there a moment where it shifted from a competition into something that felt more real?
Gillian: Honestly, Ms. Riley brought up the idea of doing the Solve-a-Thon and we were all like, "yeah, sure, a little resume booster." We went into the preliminary round during Thanksgiving break, presented over Zoom, and then got called back about a week later for the actual North Carolina state competition. That's when it started feeling real, like this wasn't just a Zoom call for a participation award. There was money involved, and we could see how it could grow.
We ended up getting second place, and Ms. Riley started getting emails about larger opportunities and people who wanted to meet with us about our idea. That's when we started realizing this could actually be feasible in the real world, especially with North Carolina's growing population and our DMV being ranked 50th out of all states.
What the App Actually Does
Lydia: What specific problems were you trying to solve, and how does your tool address them?
Cooper: The main problem was inefficiency. We built in a map system showing live wait times and average wait times at different locations. Instead of going in person or checking a website that's hard to read, you can use our app to get prepared before you go, so you know exactly what documents you need.
A big part of the DMV's inefficiency comes down to human error. People show up without the right documents or without an appointment, and that creates long lines. There are usually only two or three people checking customers in for around 300 people a day. So we proposed setting up about six kiosks for every 20 customers, letting people circulate faster, whether they have an appointment or not.
Lydia: So you're picturing a kiosk where you check in instead of waiting for your number to be called, and if you're missing something like a Social Security card, you find out right away.
Gillian: Exactly. And ideally, the app version is something you'd use at home first, when you're booking your appointment and reviewing the checklist of what to bring. Once you arrive, there would be booths acting as receptionists for check-in. That frees up staff to move into higher-impact roles, like working as driver's license examiners in the back end.
"It allows for the lines to expand. It allows for people who don't come in with appointments, who make a mistake and don't bring something, to get through it quicker and repair the mistake they made." — Cooper Brown, junior, Pine Lake Preparatory School
Choosing the Right AI Tools for the Job
Lydia: Where does AI actually come into this process? Was it mostly about vibe coding, or were there other parts that relied on AI?
Cooper: The whole integration and build process relied mostly on vibe coding in Base44, where we used Claude 4.6. But we also used ChatGPT and Gemini for research, and to cross-check the information we were using.
Lydia: You're the only team I've talked to who mentioned using Playlab. Why did you start there, and what's it like to use?
Gillian: This Solve-a-Thon competition was sponsored by Playlab, so before jumping into other platforms, we started with something small and basic. Playlab is where you vibe code: you enter prompts, say what you want, and it builds it for you.
Lydia: How did you decide which AI tool to use for which task?
Cooper: In Base44, it actually tells you which LLM is suitable for a specific task. For example, Claude is good for high data output and more accurate information, while other models are better for more creative output. You describe your app, and it suggests an LLM based on what you're trying to create.
Lydia: That's something even adults in the workforce are still figuring out, not just when to use AI, but which AI to use for which task. How did you find Base44 in the first place?
Gillian: Ms. Riley actually gave us the idea to look into Base44. We took the opportunity, looked through it, and realized it was a good next step for our DMV app.
Solving for Hallucinations in a Government Chatbot
Lydia: Walk me through the chatbot you built into the app.
Cooper: We have different sections covering all the types of documents people need. We also integrated an AI chatbot, so users can ask things like, "what do I do if I'm missing this document?" It breaks down the information, gives accurate links, and tells you what to do in those scenarios.
Gillian: The chatbot scans the official NCDMV website and looks through all the resources there to find accurate information to answer user questions.
Lydia: Have you run into any challenges with hallucinations, or had to verify the chatbot's accuracy?
Cooper: Yes. At first, it would answer questions completely outside the DMV, like questions about DMVs in other countries. So I made sure to prompt it carefully so it would only answer DMV-related questions.
Lydia: That makes sense. Without clear direction, AI can end up somewhere you don't want it to go.
Taking a Government Project from Concept to Reality
Lydia: Is your app live right now? Can people try it out, or is that complicated since you're working with a government agency?
Cooper: It's usable right now. We're trying to get it implemented with the DMV and figure out how to move forward. We're still adding ideas, working toward less human interaction so it's quicker and reduces wait times.
It's also important to recognize that North Carolina's DMV runs differently than other states. While our specific app might not roll out exactly as we built it, working with the government has been really cool. They're asking for coding segments and advice on what we'd like to see. By sharing that with the governor's office and other groups, they've given us suggestions on what we could change, and we've given them feedback on how to make the DMV more efficient.
Lydia: What features have people been most excited about?
Cooper: One is a "New to NC" button. If you just moved here from another state, it walks you through exactly what documentation and processes you need. We also built in ADA accessibility features like voice-to-text, text-to-voice, adjustable background colors, and multiple languages, plus a live map that shows nearby DMV locations based on your zip code.
"They're asking for coding segments. They're asking for advice on what we'd like to see in there. By showing that to the governor's office, they're giving us suggestions of things we could change, and we're giving them feedback to make the DMV more efficient." — Cooper Brown, junior, Pine Lake Preparatory School
What AI in Schools Should Actually Look Like
Lydia: A lot of people are worried about AI in schools right now. From your perspective, what are the opportunities, and what should people be cautious about?
Gillian: It's all about how you frame AI in classrooms. If it's framed as a big enemy, something you should never use for help, that simply makes it more appealing to use unethically, to cheat on assignments and take the easy way out.
I like how our school is starting to frame it: they level AI use one through four depending on the assignment. That allows research and creative idea generation with AI, while still making sure we're critically thinking in certain subjects. I don't think AI should be used in math to do the work for you, but if you need a hint on the next step or want to understand why you got something wrong, AI is a tool teachers should allow.
There are real ethical challenges, like AI getting things wrong or misusing personal data. But AI is amazing, and it opens so many doors. I think it will really change the world, and the earlier students learn to use it, the more ethically they'll be able to use it in the future.
Prakyath: I'd add that it depends on how it's framed. I might use it for study guides or study help, and that's good usage. But there are unethical uses too, like using it just to get answers for homework you don't want to do.
I visited a middle school classroom recently and saw an AI platform with built-in guardrails for a specific age range. Instead of just giving students answers, it would guide them through the problem and even help them improve their prompts, so they get better both at the subject and at prompting AI.
"For teachers to block off the opportunity simply makes it more appealing to use in unethical ways, to cheat on assignments, to take the easy way out." — Gillian Keith, senior, Pine Lake Preparatory School
How Students See the Future of AI and Work
Lydia: Gillian, you said AI is going to change the world. Do you see those changes more positively or negatively? Do you have predictions about your college or work experience?
Cooper: I think it definitely will change things. We don't know exactly how it will affect specific jobs, but I think for sure it will change how people work, and how some jobs might be replaced or done differently. That's scary for some people and exciting for others.
Lydia: Even in your app, those kiosks could replace receptionists, but that doesn't necessarily mean fewer DMV workers. It might just mean different jobs and faster processes.
Gillian: Right now there's a lot of concern about AI and jobs, the environment, and the ethics of how data is collected and used. I've seen a lot of progress on fixing those things, and I think it'll only get better. When cars were introduced during the Industrial Revolution, horse wagons didn't just disappear overnight, the two existed simultaneously while cars slowly got better. AI is already doing cool things in medicine, like research into Alzheimer's treatments, and in other fields too.
Prakyath: And in addition to medical, there's military use too, like AI drones. I see a lot of positives, but right now we're in a neutral zone. As Cooper and Gillian said, it could take away some jobs, but it could also make fields like medical, military, and manufacturing more efficient. That might mean a transition of jobs, and also some loss of jobs.
Aankan: I think that summed it up well. AI is the kind of topic where if you take it too far in one direction, it can be a bad thing, and if you take it too far the other direction, it can be a good thing. It depends on how you use it. Prakyath's right that it can take away some jobs, but I think it will create just as many, because as AI grows, there will be people who need to regulate it and work with it. If a job like a receptionist is replaced, that creates new jobs for people who need to work with AI. So those jobs get shifted, not just eliminated.
From a Cybersecurity Class to the Governor's Office
Lydia: You all have pretty nuanced perspectives, talking about data risks, the environment, new types of jobs, AI regulation, and military uses. Were you having these conversations before the Solve-a-Thon, or did that experience push you to think more deeply?
Gillian: I took a cybersecurity class my junior year, and Cooper and Prakyath took it as sophomores. Aankan hasn't had the chance yet. That class opened up the first real exploration of AI for us. As freshmen, we were seeing ChatGPT for the first time, and honestly I was worried about it taking away my ability to think and tempting me to cheat. But in that cybersecurity class, we started talking about what AI can do and exploring prompting. Prakyath and I even went to an AIM convention and taught teachers how to teach their students to prompt AI.
The Solve-a-Thon extended that reach even further, because we were working with the DMV, a state agency, and that meant thinking about documents being processed and stored by AI, the possibility of that data being hacked, and other ethical concerns. It pushed us further into the government side of AI.
Lydia: Aankan, you didn't take the cybersecurity class. How did you get involved?
Aankan: At our school, you choose either computer science or engineering as a pathway. I stuck with engineering for four years and only took computer science for three years, which is how I missed the cybersecurity class. But AI has been a relevant topic throughout high school regardless. My freshman year I took Computer Science Principles with Ms. Riley, sophomore year I focused on engineering, and junior year I took another computer science class with her. That's what got me thinking about AI more, and being in that class is what gave me the chance to join this team for the Solve-a-Thon.
What Makes Ms. Riley Different
Lydia: Every one of you has talked about Ms. Riley's impact on your learning and opportunities. Not every student has a teacher like that. What does she do that other teachers could learn from?
Cooper: Something pretty unique about Ms. Riley is that she makes sure we're working independently. She wants us to think on our own and be independent, while also giving us access to other opportunities. A couple of days ago she ran an in-class prompt-a-thon where students worked in groups or individually to find and solve their own problems. She does both group projects and individual thinking projects, and that helps students a lot.
Prakyath: I'd add that she makes assignments more creative by letting us choose our own path. If she wants us to build a chatbot, she won't tell us what it has to be about. It could be anything from Pokémon to school. She gives guidelines and a rubric, but lets us figure out the specifics ourselves, while always having a clear sense of what she wants us to demonstrate.
Gillian: I've had Ms. Riley since the summer before my freshman year. All the teachers at Pine Lake care about their students, but Ms. Riley is especially in tune with how we're doing mentally, because that affects how we're doing academically. She does small things like check-ins, noticing the good and bad parts of our day, and she's always available to help, whether it's with another class or something totally unrelated to school. She's opened so many doors for all of us. She found the original opportunity for the Solve-a-Thon and other opportunities like the AIM Conference. She's really involved in how we're doing as people, not just our grades.
Aankan: Ms. Riley does a great job of taking what students are good at and finding outside opportunities for them. Beyond the Solve-a-Thon, she's connected students to things like the MIT Solve Challenge, the Congressional App Challenge, and DECA. She runs business and computer science classes, and if she sees a student is good at something, she finds them a place to pursue it.
"I hope everyone has a Ms. Riley out there. She's opened so many doors alone for all of us." — Gillian Keith, senior, Pine Lake Preparatory School
What's Next for the Pine Lakers and Their App
Lydia: What do you hope happens with your DMV app in the next year or two, and what do you hope for yourselves after high school?
Aankan: For the project, I want to see more collaboration with the governor's office. We've started building that connection over the past few months. In April, we met with some of the leaders of the new DMV initiative, and I want to see that collaboration continue, especially over the summer when we'll have more time to focus on it.
Gillian: Even if our entire app doesn't get pushed out exactly as we built it, I'd love to see pieces of it integrated into what the state is forming on its own. We've researched this for so long, and so many people agree the current system is inefficient. Whether it's our project or something the governor's office builds, I just want to see the DMV become a more respected agency in North Carolina. Everyone jokes about how long the lines are, but if that's really true, we should be working on it.
Cooper: Same here. I'd like to keep working on it, making improvements, and showing it to the DMV. I just want to see the DMV become more efficient, because it's better for everyone.
Prakyath: I'd like to work with the DMV to get even small pieces of our ideas implemented, to support the workers and people currently doing this job and make things more efficient overall. We've gotten so much support for this idea, and I want to push it further than what's been tried before.
On the career side, Cooper and Prakyath both said they want to study AI and cybersecurity in college. Gillian is headed to UNC Chapel Hill for computer science, with plans to go into patent law focused on AI and technology inventions. Aankan wants to go into engineering, with a focus on combining engineering and AI in his future work.
Guest Bios
Coral Riley is a Computer Science teacher at Pine Lake Preparatory School in Mooresville, North Carolina, where she teaches both Computer Science and Business. She coached the Pine Lakers through the entire NC AI Solve-a-Thon process and has a gift for matching students with outside opportunities, from the Solve-a-Thon to the Congressional App Challenge to DECA.
Aankan Das is a junior at Pine Lake Preparatory School pursuing the engineering pathway. He joined the Pine Lakers through Ms. Riley's computer science class and brought an engineering perspective to a team with strong cybersecurity and computer science backgrounds.
Gillian Keith is a senior at Pine Lake Preparatory School heading to UNC Chapel Hill for computer science and pre-law. She hopes to become a patent lawyer working at the intersection of technology and intellectual property.
Prakyath Shankar is a junior at Pine Lake Preparatory School whose own frustrating DMV experience, a four-hour wait after arriving with the wrong documents, sparked the entire project. He plans to pursue AI and cybersecurity in college.
Cooper Brown is a junior at Pine Lake Preparatory School heading into his senior year. He plans to study AI and cybersecurity, with an eye toward combining that interest with his engineering background.
Resources Mentioned & Related Concepts
The Pine Lakers' DMV App: An AI-powered app built to help North Carolina residents navigate DMV services more efficiently, including live wait times, a document checklist by service type, a "New to NC" onboarding guide, ADA accessibility features, and an AI chatbot trained on the official NCDMV website.
Playlab: The AI-assisted vibe coding platform the team used to build their initial Solve-a-Thon prototype. Users describe what they want to create and the platform generates it in code, an accessible entry point for students new to development.
Base44: The advanced vibe coding platform the team used to continue developing their DMV app after the competition. Base44 recommends specific AI models based on what you're building, offering built-in guidance on which LLM fits which task.
NC AI Solve-a-Thon Series: All five episodes at kinwise.ai/nc-ai-solve-a-thon
Iredell Free News Coverage: "Pine Lake Prep Team Finishes Second in Statewide AI Competition," local news coverage of the team's second-place finish
