I’ve been working in technology for years, and throughout my career, I’ve watched several major shifts reshape how businesses operate, from personal computers and the internet to cloud computing and remote work. Each wave arrived with predictions, excitement and more than a little hype.

AI feels familiar to these other shifts in many ways, but the speed is different. Most new technologies unfold over years; AI capabilities seem to advance every week. The pace makes it tempting to focus on bold predictions about what work will look like in five or 10 years I’m more interested in a simpler question: What can we do with this technology today? At DigiKey, that question starts with complexity.

We offer more than 17 million products sold through our e-commerce platform. Every day, our teams navigate an enormous amount of information, from product specifications and pricing structures to trade compliance requirements, supplier data and customer questions. The challenge is rarely a lack of expertise, but rather that highly skilled people often spend far too much time gathering information and too little time applying what they know. That's exactly where AI can make the biggest difference.

One metaphor I’ve used with our teams is the idea of a “digital twin intern.” Imagine an intern who is eager to help, willing to take on repetitive work and capable of sorting through large amounts of information in a matter of seconds. They can complete meaningful work, but they still need guidance and context. And they still need someone with experience to review their work and make the final call.

That’s how I think about AI agents. They help by giving experts a different starting point. Instead of spending hours tracking down data across multiple systems, reconciling information or working through routine admin tasks, employees can begin with much of that groundwork already completed. Time once consumed by process can be redirected toward judgment, problem-solving and customer engagement.

It’s important to focus on what people can do when AI technology removes some of the obstacles in their way. Sometimes, we refer to the work AI can take on as low-value work. When farmed out to the aforementioned digital twin intern or AI agent, experienced employees are freed up to focus on solving harder problems, serving customers and driving innovation.

When people ask about our AI strategy, they often assume we’re focused on future possibilities and how we’ll transform the distribution space with AI. In reality, we’re really focused on solving very practical business problems that free up people’s time today.

Over the last year, we've followed a 70-20-10 investment model: Roughly 70% of our AI effort is focused on scaling operational capabilities. Another 20% is focused on enhancing customer experience. The remaining 10% is reserved for experimentation and longer-term opportunities.

This resource investment plan reflects what organizations are learning as AI matures—that the greatest value lies in removing friction from everyday work.

We are applying AI agents in areas like accounting, customer operations and trade compliance. These are all environments where employees spend significant amounts of time collecting information, moving between systems and navigating complex business rules.

In many cases, the work isn't difficult; it’s just a high volume surrounded by complexity. Consider trade compliance, where determining the impact of tariffs across millions of products involves a web of relationships between manufacturers, manufacturing locations, regulations and customer requirements. The number of possible scenarios grows quickly. AI is particularly well-suited for this type of environment because it can sift through large amounts of information and surface the details that matter most.

What Our First Successful Agent Taught us

Earlier this year, we put an AI agent into production within our accounting organization. The agent helps process customer remittance requests by reviewing emails, identifying what action is being requested, gathering relevant information and presenting recommendations to employees for approval. Today, that system automates approximately two-thirds of our remittance requests.

The productivity gains are meaningful, and we’ve gleaned many other valuable insights and lessons because of the process. Building that agent required us to establish the infrastructure, governance, security controls and operational processes necessary to run these systems responsibly. We learned where the technology worked well, where it struggled and where human oversight remained essential.

Those lessons and insights are now reusable; the opportunity is no longer limited to a single use case. In fact, we have something like 70+ use cases at some level of adoption throughout our organization right now. Once you understand the pattern, an entirely different set of possibilities opens up.

You can begin looking across the organization and asking the same question repeatedly: What work requires people to jump between multiple systems, gather information and make decisions based on established business rules? Those are often strong candidates for AI-assisted workflows. Getting that first operational agent into production was the hard part. Now we can focus on scaling what we've learned.

The Hardest Part Has Nothing to do with Technology

AI technology continues to improve at an extraordinary pace. What seemed unfathomable a year ago may now feel routine with a successful agent. The bigger challenge now is helping people adapt to a different way of working.

Previously, much of our time was consumed by administrative work. As those administrative tasks become increasingly automated, a different question has emerged: What should people do with the time they gain?

That question is harder to answer than it sounds. It’s tempting to think of freed-up time as slack in the system, but that’s not how it actually plays out. People who used to spend their mornings tracking down data are now expected to spend that time thinking harder about the problem in front of them, such as catching something a report missed, helping a supplier find a solution, or talking to a customer instead of just processing their request.

This shift in how we spend our time isn’t unique to DigiKey. Every organization experimenting with AI will face similar shifts. The hard conversations aren’t about the technology anymore—they’re about culture, leadership and what “good work” means once the busywork is gone. Organizations that figure that out will pull ahead of the ones that just bought the tools.

What I'm Watching Closely

If I had to identify one trend that deserves more attention, it would be the rapid evolution of open-source AI models. When most people think about AI, they think about a handful of big names. Those platforms are important, but specialized models also continue to emerge that perform exceptionally well in targeted areas, often at significantly lower costs.

That matters because AI is increasingly becoming an economics discussion as much as a technology discussion.

Selecting the right model is like selecting the right tool. A general-purpose model may be ideal for one task, while a smaller, more focused model may deliver comparable results for another task at a fraction of the cost.

As organizations scale AI adoption, those decisions become even more important. The companies that create the most value from AI will be the ones that understand which model fits which problem and can make those decisions consistently.

My View of the Future

Whenever I'm asked where AI will be five or 10 years from now, I'm reminded of similar conversations surrounding personal computers, the internet and cloud computing. At some point, the technology itself fades into the background.

Nobody introduces themselves by saying they use the internet for work. Nobody describes a laptop as an innovative strategy. Those tools simply became part of how business gets done. I believe AI is moving in the same direction.

The future is unlikely to belong to organizations that simply have access to AI. Access is increasingly becoming universal. The differentiator will be knowing how to apply these tools thoughtfully, securely and responsibly. It will be understanding where automation creates value and where human judgment remains indispensable.

The old rules still apply. Organizations must think carefully about governance, security, spending and accountability. AI doesn’t eliminate those responsibilities. If anything, it makes them more important.

For all the discussion about artificial intelligence, I remain convinced that the most important variable is still people. AI can process information and help navigate complexity. However, human expertise and the judgment of seasoned employees will always matter most.