Chris Bradley said Veritiv manually keys 1.7 million orders a year through its order entry process.
“order entry within our company, where we manually key 1.7 million orders a year”
Person
Chris Bradley
1 episodes · 18 quoted claims
Chris Bradley said Veritiv ran its first test of agentic order entry in roughly late November or early December 2025.
“we did kind of our first test of that back in probably late November, early December of 2025”
Veritiv records a replayable screenshot at every step of each agentic mission it deploys.
“every single agentic mission that we deploy, there is a screenshot at every step. And you can play it back to be able to see exactly what it it did.”
Chris Bradley said large language models are non-deterministic, so an agent can sometimes do something unexpected.
“Because, you know, L L Ms are non-deterministic, it can do something unexpected sometimes.”
Chris Bradley said Veritiv has no screen recordings of human work on computers, but it does have them for its AI agents.
“we don't have screen recordings of every time a human does anything on a computer, but we do with agents”
Chris Bradley said each call an agent makes to a large language model is a fresh model turn that must re-read the context, skills and guardrails.
“If it takes twenty different agent call or like large language calls per session, you're actually getting a fresh model every time.”
Chris Bradley said Veritiv's order agent worked well at low volume but showed anomalies once it handled thousands of orders.
“because it was operating really well at low volume. But once we got to, like, over thousands of orders, then all of a sudden you start to see these anomalies.”
Veritiv built a second agent that inspects the end of each agent run to confirm the first agent did what was expected.
“So now we have to actually build another agent that would inspect at the very end of an agent run to make sure that the agent did what we expected it to do.”
Veritiv customer service staff assign cases to the AI agent in their CRM and exchange case comments with it.
“they will sign a case and then they will have case comments back and forth with the agent who's actually doing the work on their behalf.”
Chris Bradley described decision leverage as one of two primary types of AI leverage, where AI helps with better analysis.
“there's two primary types of leverage that you can get with A I. So you can get decision leverage, where it's really helping you do better analysis.”
Chris Bradley said Veritiv deliberately chose labor leverage, automating repetitive manual work, over decision leverage.
“We deliberately picked the latter, and it's it's tempting to pick the the other one first.”
Chris Bradley obtained from Veritiv's chief HR officer a list of jobs across the company's roughly 8,000 employees to find automation targets.
“I started with our chief H R officer and I got him to give me a list of all of the jobs for like our 8,000 employees.”
Chris Bradley said Veritiv found job types such as customer service, financial operations and sales where five or six hundred people hold the same role.
“we found, like customer service, financial operations, sales, where you literally could have five, 600 people that all have the same job type.”
Chris Bradley said change management has been the biggest obstacle in his past transformations, including before AI.
“in all of the big transformations I've done, even prior to A I, the biggest thing that you is always gonna hold you back is gonna be change management.”
Chris Bradley said Veritiv used agentic tools and computer use models to automate manual processes, which he said was not possible five or six months earlier.
“how do we use agentic tools to take some of those manual processes and really figure out how to use computer use models”
Chris Bradley said large B2B companies run a patchwork of legacy ERP and homegrown systems, with some Veritiv staff using up to four monitors.
“most of us live in this world with a patchwork of legacy systems, E R P systems, homegrown systems. And you have a large number of users who have two or three or even I see in our company sometimes four monitors.”
Chris Bradley defined true agentic AI as AI that is given a goal and performs all the steps that only a human could previously do.
“True agentic A I has to be, here's what needs to get done, and the A I goes and do does all the steps that the human was the only one that could do that previously.”
Chris Bradley said AI work and AI leadership at most companies usually come from AI, machine learning, analytics and business intelligence functions.
“a lot of A I, and even the leadership in A I, usually comes from the A I, M L, analytics, B I functions at most companies.”