Automation is a popular topic in boardrooms today. Yet factory floors still rely on human hands for many delicate tasks. After 15 years of scaling software platforms and spending time at a massive employment agency, Alex Kartsel saw this disconnect firsthand. He noticed that while companies pushed toward full automation, simple tasks like inserting a car bulb remained stubbornly manual. This observation ultimately led him to join the early-stage company Pixoft as Co-Founder and Chief Operating Officer (COO), stepping away from established businesses to help build a new model for industrial robotics.
Identifying the Gap in the Market
Before his current role, Kartsel spent over two years at EWL. The agency placed thousands of blue-collar workers into warehouses and factories across Central and Eastern Europe. These workers were filling the exact physical gaps that machines simply could not handle. That hands-on exposure to manufacturing bottlenecks shifted his perspective on where the robotics industry was falling short.
“You have humanoid robots that can run a marathon, but they cannot pick up an egg because they lack that sense of touch and manipulation,” he explains. The core problem was never simply mechanical hardware. Training a robot on an active assembly line requires halting production and installing numerous cameras to capture data. Factory managers naturally resist stopping their lines for experimental technology. This makes the traditional training process expensive and difficult to scale. For example, producing car headlamps often involves a fully automated line right up until the pre-assembly stage. Inserting and rotating a single small bulb remains highly complex for a machine to execute safely. Originally, Kartsel entered the picture as an angel investor, but conversations with the technical team prompted him to take an active leadership role.
The Reality of Building Again
Starting over comes with a specific set of hurdles. Today, his primary focus revolves around securing funding for research and development. The robotics sector is currently experiencing a massive surge in new entrants. In the Bay Area alone, hundreds of similar companies launched this year, making the competition for venture capital incredibly intense. “You feel like you’re doing the same thing on a daily basis with some result, but even if it’s a five percent conversion, that means you need a hundred calls to get five yeses,” Kartsel notes. Enterprises also require significant confidence before adopting unproven software. Selling an unfinished product means leaning heavily on pilot programs and demonstrations to earn client trust. Selling into the industrial space requires immense patience. The sales cycles are notoriously long, and software providers rarely get a second chance to impress plant managers. You cannot approach a major facility with an incomplete tool and promise to fix it later. This dynamic places enormous pressure on the founders to get the underlying technology right the first time.
Leaning on Past Lessons
Decades of scaling previous ventures taught Kartsel a valuable lesson about early-stage companies. Technology alone rarely closes a deal with industrial clients. Success hinges on assembling the right mix of people and pointing them at a very specific commercial problem. When approaching investors, leadership must prove they have the right personnel to execute the vision. To that end, the company includes individuals with decades of practical automation experience. These are professionals who have spent years integrating machines into factory settings. Their presence shifts the conversation from artificial intelligence (AI) theory to practical, ground-level solutions. They understand exactly where factories are losing money and time. “We are not just an AI company focusing on technology; we actually focus on real problems on the ground,” Kartsel says. “A lot of companies focus on the product only, but we want to solve issues for our customers.” By targeting things like faster implementation times and lower training costs, the team aims to deliver clear financial value.
Looking ahead, the broader challenge for the industry involves teaching machines to handle highly precise tasks. Up until now, these delicate movements have been off-limits for automated systems. Pixoft is testing its simulation models on printed circuit boards, a process that still relies heavily on manual labor. Developing skills in a virtual setting bypasses the need for disruptive physical testing. “Based on synthetic data, we were able to train our systems for the PCB market to perform a through-hole insertion task where the hole is less than one millimeter,” Kartsel explains. If robots can master these tiny movements in a virtual environment, they can eventually be deployed to handle similarly complex tasks in the physical world. This breakthrough would shift how modern electronics are assembled.
This approach is not limited to electronics. The strategy involves adapting these simulated learning models for automotive manufacturing, food production, and textiles. The goal is to provide a horizontal tool that multiple sectors can utilize. For Kartsel, bringing his operational background to this evolving space offers a chance to change how factories function over the next few years.
Follow Alex Kartsel on LinkedIn for more insights on industrial robotics, scaling early-stage companies, and AI-driven automation.
