Thesis

The idea stemmed from firsthand experience in industrial environments. We identified a major pain point: a lack of real-time visibility into production line operations.

In most factories, especially in the food and packaging industries, machinery is controlled by PLCs (programmable logic controllers). These systems are rarely connected to the cloud or centralized management tools. As a result, factory managers often discover issues hours after they occur – too late to respond efficiently.

Our goal was to build a system that could monitor these machines in real-time without requiring complex integrations. The solution would use AI to extract data directly from Human-Machine Interface (HMI) screens by capturing images with a smartphone or camera. These images would then be uploaded to the cloud for analysis, allowing the system to deliver real-time insights on machine peformance, including production rates, faults, and bottlenecks.

The key advantages of our proposed system were:

• Seamless Integration with existing equipment
• No need for PLC access
• Predictive capabilities for production and faults

Validation

Validation began after we shaped the concept through direct exposure in factories. We believed the problem was real and worth solving.

We met weekly to refine the idea and created a one-pager and pitch deck to share with potential clients. Our guiding principle was: “Let’s find out if someone is willing to pay before building anything.”

We held 15-20 discovery calls with potential customers – mostly factory and plant managers – to test the concept. Although we saw the global potential, we focused first on the Israeli market to secure early traction.

Our primary objective during validation was to assess customers’, not just whether they recognized the problem.

We also engaged with industry communities and Industry 4.0 accelerators to expand our reach and collect more feedback.

Why it didn’t progress 

Despite early interest, validation fell short. Five potential customers were open to trying the system in their factories, but none were willing to pay for it.

That was a turning point. We realized we lacked the deep commitment needed to push the idea forward. Without customer urgency or willingness to pay, and without strong personal connection to the problem or customer base, we couldn’t justify going “all in.”

Additionally, there was a misalignment in our motivations: one of us was more drawn to the underlying technology (LLM and computer vision), while the other was focused on building a business.

These factors combined to prevent us from reaching the conviction needed to raise funding or pursue the idea further.

Insights

1. Customer commitment matters more than problem recognition: People may acknowledge a problem, but what really matters is whether they’ll pay to solve it.

2. Passion for the specific customer is key: Without a strong connection to the people you’re serving, it’s hard to persevere through the challenges of building a company.

3. Start local: Validating with nearby customers is a practical first step, even if your long-term market is global.

4. Co-founders make a difference: After a previous solo venture, I realized how valuable it is to have a partner – for support, complementary skills, and shared capacity.

---