Schneider Electric Ventures Panel “AI-Native Startups Changing the Game”; Post-Event Q&A with Founder and CEO Daniel First

Axion was privileged to participate in the Schneider Electric Ventures panel, “AI-Native Startups Changing the Game,” a conversation moderated by Amit Chaturvedy, Global Head & Managing Partner at SE Ventures,  with panelists Abhishek Singh, CEO/Co-Founder of AiDash and Sami Shalabi, CTO/Co-Founder of Maven AGI, and our Founder and CEO, Daniel First. 

The panel covered several aspects related to enterprise AI adoption, including Axion’s approach to solving real-world manufacturing quality problems, the importance of integrating engineering expertise into AI systems, and the unique value of human-in-the-loop observability. The conversation also highlighted how Axion is partnering with Schneider Electric Ventures to support stakeholders responsible for quality, digital transformation, and advanced services for industrial and data-center customers.

Here is a recap of the discussion:

The first question from Amit was “What problem is Axion solving and how did you identify the gap in the market?”

Daniel:
“At Axion, we focus on a simple but painful problem: manufacturers are often the last to know when their products are failing in the real world. When you release a new airplane, a medical device, or even something as simple as an ice cream machine, there aren’t just one or two issues. There can be hundreds or thousands of “death-by-a-thousand-cuts” quality problems that customers experience. Inside these companies, enormous data sets—warranty claims, technician notes, fault codes, IoT readings, one-star reviews—are scattered across dozens of systems.

The gap I saw working with industry leaders was the absence of an “Issue 360” view. While there are BI dashboards, data lakes, and AI prototypes, nobody had an always-on command center that could answer: “Here are the top issues impacting your customers this week, here’s why they’re happening, and here’s who needs to fix them and how. ” The goal is to move manufacturers from being surprised by recalls and field failures to a world where every product is constantly learning and getting better based on real customer experience.”

"What problem is Axion solving? Manufacturers are often the last to know when products fail in the real world—Axion brings together scattered data sources with AI observability, so issues are detected and resolved before recalls and failures, leading to safer, smarter products for everyone.”

The next question focused on how to be successful with Enterprise AI. Specifically “What has been the biggest roadblock you’ve run into with customers, and what makes your approach different from legacy platforms?” 

Daniel:
“Before Axion, many of our customers had tried ”Enterprise AI” the traditional way: big consulting engagements that didn’t know how to bake an understanding of the problem into the software itself. That’s why most AI pilots never make it to production. Instead, we saw three persistent gaps:

• The knowledge gap: Models didn’t capture an engineer’s expertise. For example, the AI might say, “You have a wire harness issue,” but an expert knows there are four different harness types that fail in specific ways. If their insight can’t be embedded, engineers stop trusting it.

• The process gap: Even if the AI finds a real issue, it’s often unclear who owns it, what meeting it belongs in, or how it turns into a design change or supplier action. Great insights die in someone’s inbox.

• The action gap: Legacy platforms stop at dashboards. They don’t help manage the lifecycle from “we detected an issue” to “we validated it, fixed it, and saw the impact.”

Our approach is different in two key ways: 

1. Human-in-the-loop by design, combined with proprietary manufacturing intelligence. We study how aerospace and medical-device engineers actually scope and solve issues, and encode those steps in the platform. Experts can correct the AI, and that knowledge is persistently reused. 

2. Axion facilitates teams to drive action through embedded workflows. For example, Axion tracks which groups actually pick up and close issues, surfacing bottlenecks - “engineering is fast, but this supplier gate is always slow”—to lay the foundation for process improvement that goes hand-in-hand with quality improvement.”

Enterprise AI success means human-in-the-loop bridging manufacturing intelligence, process, and action gaps. 

Amit and Daniel went on to discuss Axion’s partnership with Schneider Electric and how the two companies work together.

Daniel:
“With Schneider Electric, we’re partnering with teams at the intersection of critical infrastructure and AI-driven operations—the people thinking about keeping power, cooling, and industrial systems reliable as complexity explodes.

We work with stakeholders responsible for quality, digital transformation, and advanced services for industrial and data-center customers. Our joint objective is tightly aligned: detect emerging issues in critical systems early, understand them quickly, and orchestrate coordinated response across engineering, service, and suppliers.

Schneider brings deep domain expertise in energy management, automation, and large-scale deployments. Axion brings the AI “Issue 360” layer to unify fragmented data—service logs, alarms, sensor data, field reports—and turn that into a prioritized list of issues and actionable recommendations.

Together, we’re focused on:

• Improving uptime and reliability for critical assets,

• Reducing warranty and service costs,

• Creating field-to-design feedback loops for next-generation products.

Axion and Schneider Electric are working to boost asset reliability, lower service costs, and enhance next-generation product design through actionable field insights

Amit and Daniel then moved on to the topic of the future of AI in Manufacturing focusing on the question: “How will AI transform manufacturing in the future? What downstream effects do you expect?” 

Daniel:
“AI’s impact on manufacturing is often misdiagnosed. The conversation is about robots and cost reduction, but the true transformation is that every product will continuously learn from its customers.

Today, most manufacturers only focus on quality issues for the top 5–10% of SKUs, because it’s a heavy manual process to gather and analyze the data. AI agents flip the equation. By automatically reading warranty claims, technician notes, telematics, service calls, and reviews, you can detect issues at the earliest warning signal , understand which customer segments they matter to, and then propose fixes to design, software, manufacturing, or service. This drives competitive advantage shifts from “who was first to market” to “who learns fastest from the field. 

"AI is not about robots and cost reduction - AI will transform manufacturing by enabling products to learn from customers, driving smarter, safer innovation.”

The closing topic of the panel was “From an AI-native founder, what advice would you share with those adopting AI?”

Daniel:
“Start with a real, painful problem—not with AI itself. In our world, that’s “We’re spending billions on quality and recalls, and we don’t even know what’s driving it.” If you can’t state the problem in plain English and tie it to P&L or safety, you’re not ready for an AI project.

Second, design for humans and processes from Day One. Most failed AI pilots didn’t flop because the model was bad—they failed because experts couldn’t inject their knowledge, nobody knew who was supposed to act, and the project wasn’t wired into real workflows.

Leaders in this era will be those who treat AI as a way to learn from customers faster and turn that learning into better products and better organizations.”

“Adopt AI by solving real business pain, and build solutions around people and processes first”

That recaps our discussion. It was a great two days learning about innovation across the industry, hearing from customers, and spending time with the broader Schneider Electric team.

Interested in learning more? See how Axion connects disparate data and surfaces insights that help manufacturing teams fix product issues at the first signal, and sign up here to explore a custom demo on your own data.