Inside TE Connectivity's AI Journey: Interview with Brontë Hamilton

As AI continues to reshape industries and organizations, the world seeks insights from pioneering companies that have boldly embraced AI to learn how to navigate its complexities. Recently, we had an incredible opportunity to sit down with four industry leaders at AMLD EPFL 2024 and talk about their perspectives, thoughts, and experiences on structuring and designing winning AI programs.

This time, we're bringing you a piece of our conversation with Brontë Hamilton from TE Connectivity. Brontë shares her experience, detailing how TE Connectivity began their AI journey, the challenges they encountered, and the key lessons learned along the way.

Bronte Hamilton

CAN YOU TELL US A LITTLE BIT ABOUT YOURSELF?

Brontë Hamilton: Originally from Australia, I began my career in the Australian Army. Later, I earned a triple major in Information Systems and National Threat Infrastructure Protection. Afterward, I pursued an MBA at Penn State, which led me to a career in finance with TE Connectivity. In my role, I tackled increasingly complex problems that helped shape my expertise. Now based in Zurich, I manage strategy, business development, marketing, pricing, M&A, and integration at TE Connectivity.

COULD YOU SHARE A BIT MORE ABOUT TE CONNECTIVITY?

Brontë Hamilton: We produce connectors, relays, sensors, and switches. We're an essential yet relatively unknown company, involved in everything from microwaves and cars to utilities, computers, and air conditioning—essentially anything involving data, power, or signal.

TE Connectivity manufactures over 250 billion connectors, switches, and sensors annually across numerous global locations. My focus is on industrial connectors, which we have over a million parts of, and actively sell about 400,000 of them, creating a significant opportunity for AI due to the complexity of our customer base and product set. For instance, our connectors in an electric vehicle might only account for a few hundred dollars of the total cost, but the vehicle can't function without them. It’s critical we deliver the right components to the right place at the right time.

SO HOW DID TE CONNECTIVITY'S AI JOURNEY ACTUALLY BEGIN?

Brontë Hamilton: TE Connectivity began its AI journey in 2015, driven by the need to refine processes and maintain high productivity in its plants, which produce millions of components daily. The initial application of AI was in machine vision for quality testing, a critical area since even minor component failures can have significant repercussions.
In 2018, we hired a PhD expert from Caterpillar to formalize our approach to analytics and AI. The goal was to leverage data more effectively. Over the next few years, we developed a hub-and-spoke model: a core team of data scientists supporting ten business units on specific projects, building models and deepening their expertise.

CAN YOU SHARE SOME SPECIFIC AREAS WHERE AI HAS MADE A BIG IMPACT?  

Brontë Hamilton: One key area of focus has been part-level forecasting within the sensors business unit. With a million parts to manage, predicting production needs is crucial to avoid costly downtime and inventory issues. By combining statistical methods with neural networks, the company improved forecasts, especially for historically difficult-to-predict parts.

As we continued our AI journey, we began to recognize the growing importance and impact of AI on business processes. Today, we are developing a more comprehensive AI governance model while continuing to use the hub-and-spoke approach. We also plan on establishing an AI center in Singapore and ensuring we are more consistent and structured in how we approach AI.

When it comes to our AI program, our main objective is to develop a scalable enterprise foundation. This is challenging due to the company’s structure, with ten business units operating like independent entities. Trying to get 10 different cats galloping in the same direction can be tricky, however, achieving alignment across these units is essential for scalability and cost-effectiveness. Therefore, we need a scalable foundation, with a robust governance structure, resource allocation, an AI hub in Singapore, and the right talent, such as data scientists.

AI initiatives must have a clear purpose and intent. We operate in a highly competitive and complex ecosystem involving the supply chain, customers, and products, and so demonstrating tangible benefits to both customers and shareholders is crucial. Each business unit prioritizes its applications, ensuring that the models developed are used effectively. The best way to ensure this is by allowing people to suggest the problems they want to solve, fostering ownership and practical application of AI solutions.

WHAT KEY LEARNINGS HAVE YOU GATHERED FROM YOUR AI JOURNEY?

Brontë Hamilton: One major takeaway from our AI journey is that it’s non-compressible. In R&D, there's a maxim: If a product takes five years to develop, you might as well wait for advancements that allow you to develop it in two, three years.

However, AI is not just about the tools or the models; your organization has to learn how to use them, which takes time. My advice is to start early, start anywhere, make mistakes, and try again. This iterative process helps your team frame business problems in a way that data scientists and AI/ML specialists can address. Building this relationship is especially challenging in traditional organizations with a long history of avoiding high-volume experimentation.

Secondly, data is crucial. We were fortunate because, about 20 years ago, someone advised us to start collecting data. Like dragons hoarding treasure, we collected all the data. So, now we have a data lake. While we didn't initially know how to use it, having the data was a crucial first step. Now, our journey involves assessing the quality and power of this data and identifying the right datasets for the problems we want to solve. When scanning for problems to solve with AI, the first thing I consider is the quality of the data. For instance, I am currently working on a pricing project with excellent data. This data is frequent, allows for market feedback within 0 to 12 months, and is repeatable. Its depth and cross-referencing capabilities make it ideal for AI applications.

Another key point is that every change breaks someone's workflow and that someone is probably going to be unhappy in some way. You have to deal with the change management aspects. Learning from the journey is essential, and it ties back to the non-compressibility aspect. As your organization learns to execute AI/ML, capturing these learnings in a structured way is crucial. How do you embed these insights in your team? How do you develop competencies? It's not just about technical expertise but also teaching decision-makers to leverage both human and machine learning. We need to capture these things.

For more insights into the AI journeys of leading companies, such as dsm-firmenich and Tetra Pak, check out the entire panel discussion here.