Supply Chain Data Readiness: Why Perfection Is the Wrong Benchmark for AI (with Brian Reed)
Supply Chain Data Readiness: Why Perfection Is the Wrong Benchmark for AI (with Brian Reed)
Episode Summary
Everyone says garbage in, garbage out. Almost no one acts on what that actually demands. In this episode, Derek Aranda sits down with Brian Reed, an executive sponsor at Oracle who has run supply chains, advised the C-suite, and built autonomous planning systems across 56 countries, to pull apart what "good data" really means on the road to AI. The conversation cuts against the reflex to hold AI to a standard of perfection no human has ever met, and reframes data quality as a question of timeliness, location, and fitness for purpose rather than flawless numbers. It is a clear-eyed look at why so many AI pilots get written off as failures when they are actually doing their job, and what it takes to move from hand-holding to trust to genuine autonomy.
🎧 Episode Highlights
Why AI's inaccuracies get all the scrutiny while human inaccuracy goes unquestioned.
Why a "failed" AI pilot that exposes bad data was never really a failure.
The "hands off the wheel" case study: 30 days of tracking that convinced a customer to trust the system over their own team.
Who should own the data: resolving the IT vs. business tension and putting AI on the RACI chart.
Teaching AI to "trust and verify": what it takes to move from human-in-the-loop to real autonomy.
Brian's closing advice: experiment small, because "the quarters and the dimes" add up to millions.
🔑 Key Takeaways
-
Measure AI against humans, not perfection. The bar for a system is not flawlessness, it is whether it beats the fallible person it replaces on speed and accuracy. Brian's autonomous planning project proved the point: over 30 days of tracking, the humans consistently broke the system's answer. Perfection can stay the goal without being the benchmark you measure against to move forward.
-
A failed pilot usually isn't one. A program that reveals your data is awful just did something valuable, it found the problem you didn't know you had. The work isn't perfecting every field before you start; it's prioritizing which data actually needs bit-for-bit accuracy and which is good enough for the decision at hand. Eat the elephant one bite at a time instead of waiting for a fully stocked data lake.
-
Data ownership is joint, and the machine is now on the RACI. The business owns data creation and the value drawn from it; IT is the enabler in the middle keeping it centralized, clean, and usable. Those two groups belong in nearly every conversation together, and AI is now a third party on the chart, responsible for storage and processing, informed by inputs across the business.
💬 Notable Quotes
"We talk all about the data and the inaccuracies of things about AI. We don't talk about the inaccuracy of humans." Brian Reed
"You didn't know your data was awful before you ran the program. Now you know your data's awful, so it has some value." Brian Reed
"Perfection can always be the goal, but it doesn't have to be the bar at which you measure to go forward." Brian Reed
"Small value adds up to big value. We call it the quarters and the dimes. We added up enough quarters and dimes, and they turned into millions of dollars." Brian Reed
👤 About The Guest
Brian Reed is a tenured supply chain expert with 20+ years of global experience spanning every aspect of supply chain, from customer service through final-mile fulfillment, across all industries and on both the business and IT sides of digital and physical supply chain solutions. Beyond technology and consulting organizations, he has led a high-growth supply chain organization focused on execution at a large U.S. CPG manufacturer. Brian has helped some of the world's largest technology and manufacturing companies transform their supply chains with technology as a tool for automation, insight, and resilience, including broad transformations delivering cost savings and revenue enablement measured in the hundreds of millions to billions of dollars.
He was most recently the AWS Supply Chain and Pan-Amazon Supply Chain GTM Leader, working with Amazon's largest and most strategic customers and partners to develop and deliver new supply chain capabilities. He now advises the C-suite of Oracle's largest and most strategic customers, helping them deploy Oracle Applications and AI at scale.
Outside of work, Brian and his family enjoy traveling internationally, food and wine, and hitting the mountain bike trails as often as possible, both at home in Southern California and while traveling.
🎙️ About The Host
Derek Aranda spent over two decades as a global executive operating across commercial, supply chain, and digital transformation roles at scale. That span across the full value chain shapes his lens: the decisions, incentives, trust, economics, and governance that determine whether technology actually works inside real supply chains. On Supply ChAInge, he pressure-tests the autonomous future and helps leaders shape the framework to build it on their own terms.
Stay Connected
Produced by Speakerbox Media.