The Reality of AI Adoption in Traditional Industry

Kevin Coleman discusses practical perspectives on implementing AI across a $1.2 billion logistics operation. Rather than focusing on hype, he addresses real challenges traditional industries face: data quality inconsistencies, security concerns, and avoiding technology adoption without scalability plans.

About the Guest

Kevin Coleman is the CEO of CJ Logistics America, overseeing the Americas division with $1.2 billion in revenue and over 3,500 employees. The parent company generates $9+ billion in total revenue globally.

Key Topics Discussed

Data Quality Foundation

Kevin emphasizes the "garbage in, garbage out" principle—the importance of quality data over sophisticated models. Consistent data structures must be established before implementing AI solutions effectively.

Technology Evaluation Framework

CJ Logistics uses Centers of Excellence (COE) to assess and scale technologies across their network. The focus is on multi-site scalability rather than isolated implementations, evaluating additional use cases beyond initial deployment.

AI Integration in Operations

The conversation covers how CJ Logistics shifted from a safety focus to broader operational intelligence, including productivity optimization through warehouse waste reduction analysis and emerging quality and compliance applications.

Organizational Change Management

Kevin discusses embedding AI tools into leader standard work rather than overwhelming teams with data. Training and development must be tied to specific processes, balancing innovation with security constraints.

Multi-Client Facility Challenges

Operating multi-client facilities means managing multiple WMS systems, complex labor and volume management, and exploring shared services model potential with AI enhancement.

Key Quotes

"Supply chain or logistics has a ton of data. It's limited on their information and it's harder to get to the intelligence piece. Where I think AI plays a huge role is that bridge to getting intelligence at a much quicker pace and a much greater scale."

"It's not always the technology's fault. It's not always the people. It's just the intersection of people, process, and technology—and how do you use that together to get the most out of everything?"

"How do you set boundaries but not stifle creativity?"

"People talk about cost savings. I like to talk about waste reduction—travel's really just waste. How do you continue to drive that waste out of the supply chain?"

"My advice would be—look at the tech for the use case you're trying to solve. Can you scale with that across multiple buildings? And does that provider have the opportunity to work with you on other use cases? It's too easy to chase the shiny object."

"How do we not throw so much data at our leaders? What you need to provide is linkage to leader standard work. So what am I going to come in to do for the day? It needs to be very digestible in how I go and lead my day."

Notable Statistics

• CJ Logistics America: $1.2 billion in Americas division revenue
• Parent company: $9+ billion total revenue
• Employee base: 3,500+ workers
• Safety metric: 1.60 (described as "phenomenal" for the industry)
• OneTrack deployed across 40+ CJ Logistics sites

Key Takeaways

1. Start with data quality - "Garbage in, garbage out" drives more than ever—quality data structures must be established before implementing AI
2. Scale thoughtfully - Use Centers of Excellence to evaluate technologies before broad rollout; if it can't scale across multiple buildings, it doesn't work for a network
3. Embed AI into leader standard work - Don't just give people data; tie it to what they actually do each day so adoption is natural
4. Balance innovation and security - Use closed AI environments (e.g., Gemini over open ChatGPT) and route AI adoption through tech steering committees
5. Think core, adjacent, transformational - Start with a core use case (safety), expand to adjacencies (productivity, quality), then pursue transformational applications (shared services + AI)