Duke student Jasper Svenson started exploring AI agents in the fall of 2025, focusing primarily on their real-world business applications. And with a background in restaurant-focused investment banking—as well as parents who spent their careers in the restaurant industry—he kept coming back to one space: the drive-thru.
“After some conversations, I fell onto the drive-thru, which is a spot that I think is going to be very automated over time,” Svenson said.
With that in mind, he brought on his classmate Carlos Penzini, a math and computer science major, as a technical co-founder, and the two quickly built a minimum viable product focused on voice AI and automating the order-taking process.
Early conversations with restaurants were strong, and the product was functional, but within a few months, they ran into a problem: the space was crowded, with larger companies offering similar products and little differentiation.
“We took a step back,” Svenson said. “Eventually, we found what we’re currently doing now, which is the personalization opportunity in the drive thru.”
That opportunity is at the heart of the duo’s budding startup, Serv Technologies, which provides a data platform enabling drive-thru restaurants to recognize returning customers and personalize orders in real time.
Pivoting to drive-thru personalization
In the restaurant business, personalization and recognition are massive tools to improve customer retention. Currently, especially at fast food and fast casual restaurants, the drive-thru has no memory; every interaction starts from zero, even for repeat customers.
“We know that personalization works really well in the restaurant setting,” Svenson said. “Companies that integrate personalization or some sort of loyalty program, they’ve seen a massive boost in revenue from those efforts.”
With its personalization concept, Serv is aiming to change that by giving restaurants the ability to recognize customers and tailor interactions in real time. The company has built a proprietary hardware and software system that uses a camera setup to identify returning vehicles and link them to past order behavior, integrating directly with restaurant POS and API systems.
The product has two versions. The first, and the one they’re currently leaning into, is fully anonymized. It uses license plate recognition to link a car to past order history, but the plate is immediately hashed into a random ID—meaning no names or identifiable data are stored.
“What we’re able to do is connect your order history to your car before you get to the window,” Penzini said. “So we’re practically doing everything the restaurant does after they get a customer’s credit card number, but we’re doing it earlier so they can have a great interaction with the customer before they even reach the window.”
The second version of the product integrates with a restaurant’s existing loyalty program. In that case, users opt in, enter their license plate, and connect their identity to their order history, allowing for a more personalized experience tied directly to the customer.
Solving for what?
The concept—personalization and recognition in the drive thru—is aiming to drive a few major outcomes, Svenson and Penzini said. One of these outcomes is the ability for more strategic and effective upselling. Based on previous order history, and the recognition of that order history through Serv’s tech, the restaurant can use past order data to predict what someone is likely to add and surface relevant suggestions. The other major outcome is recognition, which ultimately builds loyalty.
”If a customer came the last three days, [and the restaurant recognized that] and wanted to give [the customer] a free lemonade, that goes a long way compared to treating the customer the exact same as the customer in front of them,” Svenson said.
Pilots and the near future of Serv
Serv Technologies has rapidly grown into itself. At the beginning of this year, the team was building out the new product after pivoting from voice AI. Now, they are on their second version and have already moved into deployment, with pilot programs underway in three large North American quick service restaurants to test performance in real drive-thru environments.
QUICK BITS
Startup: Serv Technologies
Co-Founders: Jasper Svenson, Carlos Penzini
Founded: 2025
Team size: 9
Location: Durham
Website: servtech.co
Funding: Bootstrapped
“When we got our first yes to pilot, that was the ‘woah’ moment,” Penzini said. “We know we have something very valuable, but it’s exciting to know that these bigger operators are willing to take the chance on it.”
In the short term, the team is focused on ensuring these pilots and restaurant integrations run smoothly, while continuing to iterate on the product.
Longer term, the team sees an opportunity to layer its system into existing ordering infrastructure rather than rebuild it. Instead of developing its own voice AI, Serv is exploring integrations with established voice providers, allowing its personalization and recognition technology to sit on top of those systems.
The goal is to complement, plugging into the ordering layer while enhancing it with customer-specific data and real-time insights.
