Your new engineering assistant runs on algorithms: AI reshapes parts procurement

July 17, 2025 by Aimee Kalnoskas

Collected at: https://www.eeworldonline.com/your-new-engineering-assistant-runs-on-algorithms-ai-reshapes-parts-procurement/

Machine learning is enabling smarter supply chains and frictionless component sourcing for design engineers.

The electronics distribution industry is undergoing a quiet revolution, one that’s fundamentally changing how engineers source components and how distributors predict demand. At the forefront of this transformation is DigiKey, where artificial intelligence isn’t just a buzzword — it’s becoming the backbone of operations that serve engineers worldwide.

Dave Doherty, president of the Thief River Falls-based distributor, oversees this AI-driven evolution at a company that ships globally from a relatively remote location in Minnesota. With over 70 different AI programs either deployed or in development, DigiKey represents a microcosm of how the entire distribution industry is leveraging machine learning to solve age-old challenges.

EEWorld sat down with Doherty recently to learn more about DigiKey’s AI transformation journey and his insights into the industry, AI tools, customer experience improvements, and trends.

From rearview mirror to front windshield

One significant breakthrough lies in demand forecasting, which has traditionally been distribution’s Achilles’ heel. “The ability to forecast, or the concept of forecasting, has always been a rear-view mirror,” Doherty explains. “Distributors [would ask] What have you been buying? And put a recency on whatever I’ve been selling, I’ll buy what I’ve sold.”

Today’s AI-powered approach represents a paradigm shift. DigiKey now aggregates data from multiple forward-looking sources: customer web searches, quote requests, bill-of-materials uploads, manufacturer homepage activity, and even Google keyword trends for datasheets. “Now we start building the forecast algorithm with more inputs on what’s going on the front wheel, not just what you bought,” Doherty notes.

This predictive capability is particularly crucial for DigiKey’s business model. Unlike broad-line distributors focused on high-volume production, DigiKey targets the “widest breadth” of engineering applications and low-to-mid-level production runs. “We want to support engineering everywhere,” Doherty says, making early demand signals essential for maintaining inventory breadth without excessive risk.

Streamlining the engineering experience

For design engineers, AI is quietly eliminating friction from the component selection process. DigiKey’s parametric search improvements exemplify this trend. Rather than manually filtering through voltage ranges that might span “one volt to 43 volts” versus “two to 52 volts,” engineers can now use natural language queries, such as “V out greater than or equal to 100 volts,” and automatically see relevant results.

“This will eventually lead to more of the Natural Language search,” Doherty predicts, envisioning queries like “looking for an A to D converter, a 14-bit in this level precision.” The goal mirrors Google’s evolution: “You notice that first block in Google — I rarely go beyond that.”

The company is also developing more innovative technical support systems that leverage collective knowledge rather than starting each inquiry from scratch. “Instead of having 80 techs answering the same question, the next call comes in and starts from scratch,” the AI system builds upon previous interactions, supplier FAQs, and accumulated expertise.

Internal operations revolution

Behind the scenes, AI is transforming DigiKey’s operational efficiency in ways that directly benefit customers. The cash payment application, previously a labor-intensive nightmare of matching customer payments to invoices, now operates with 90% automation for customer applications and nearly 70% accuracy for specific invoice matching.

Export compliance coding, historically handled by technicians making “best estimates” when supplier documentation was incomplete, has been fully automated for four years. “Got the hit ratio up, where they code these parts as good or better than a human does,” Doherty reports.

The company’s massive automated warehouse benefits from AI-driven load balancing that predicts staffing needs across receiving, picking, and packing operations. This predictive approach eliminates the costly half-hour transitions when workers move between stations.

Navigating geopolitical complexity

AI is also helping DigiKey address geopolitical challenges affecting the electronics supply chain. The company has implemented transparency tools that show tariff implications and offer filters to exclude tariff-eligible products. “We’ve got a filter now on the US website as well that says, don’t show me any tariff-eligible products,” Doherty explains.

More sophisticated is the evolving definition of “country of origin” for semiconductors. With industry discussions around “country of diffusion” — where the silicon wafer was processed rather than where final assembly occurred — AI helps manage these complex determinations that affect sourcing decisions.

Industry-wide transformation

Doherty sees this AI integration as part of a broader industry evolution toward system-level solutions rather than individual components. Development boards, starter kits, and integrated solutions are becoming the fastest path to customer acquisition because “customers really want to self-serve, but they want accuracy and they want timeliness and instant gratification.”

The transformation extends to discovering unexpected applications. When suppliers describe parts for specific end-uses, DigiKey’s broad web presence often reveals applications “outside of their target scope.” This serendipitous discovery process, enabled by AI-powered search and recommendation systems, helps both engineers and suppliers identify new market opportunities.

Looking forward

As edge AI applications proliferate — reminiscent of IoT’s emergence a decade ago — Doherty sees the fundamental building blocks remaining constant: sensors, data converters, microcontrollers, and processing power. “What do you need? You need to sense something. You convert it to a digital signal. You had a process, something to do with that signal, and then you needed to be able to go out to some sort of an actuator.”

The revolution isn’t in creating entirely new component categories but in making existing technologies more accessible and discoverable. “It’s quietly continuing to infiltrate in a good way, our lives and our interactions,” Doherty observes.

For electronic design engineers, this AI-driven transformation promises a future where component sourcing becomes as intuitive as consumer e-commerce, while maintaining the technical depth and reliability that complex designs demand.