CMOs Are Auditing Legacy Workflows To Unlock Agentic AI's Full Value

Enterprise marketing teams have started deploying agentic AI on top of the workflows they already have, and the results are exposing a problem the tools cannot fix on their own. Automating a workflow that was inefficient before AI arrived doesn't remove the inefficiency, it accelerates it. The friction that used to live inside slow human handoffs now lands on top of executives in the form of AI-generated output they have to review at volume. A growing set of enterprise CMOs are arriving at the same conclusion: the productive move is to redesign the process first and then let the agents run on the cleaner version.

Chris Bradley is the Chief Marketing Officer and Head of AI Transformation at Veritiv, a multi-billion-dollar specialty packaging distributor. He holds more than 40 patents in product and packaging technologies, and describes his current role as that of a "forward-deployed engineer for marketing," a CMO who prototypes agentic tooling personally and lives inside the systems before recommending them to his team or his board. This combination gives him a working view of where enterprise marketing teams are absorbing AI's productivity gains, and where they are still generating operational problems in the process.

"Get your hands dirty and make sure you really understand the art of what's possible. You can't be an effective leader driving an AI strategy if you don't know what the technology can actually do," Bradley says. The reason he pushes CMOs into direct contact with the tools is that the volume and quality of AI-generated output are changing what marketing leadership does day-to-day. As agents produce more of the initial draft work, marketing executives are becoming the review layer that decides which drafts meet the brand's standards and which need to be rebuilt. Bradley's team audits the underlying legacy processes before adding agents to them so the executive review load lands only on decisions that are worth the review time.

14 days to 15 minutes

Bradley's approach to the audit borrows directly from lean manufacturing: before agents get deployed, his team runs a formal Kaizen review of the existing workflow to identify the steps that exist only because human handoffs used to require them. "As we looked at the sheer number of steps and how brittle the system started to become, we realized we needed to do a full Kaizen," Bradley explains. "We've now reduced a 14-day process down to 15 minutes by taking 34 of the 36 steps out. It goes directly from an intake form to the end result, whereas before it would bounce between four different departments doing one thing at a time."

Bradley's marketing organization is currently running at roughly 12 times its 2025 rate, completing more projects in the last month than the team completed in all of 2025 combined. Using AI as a strategic sparring partner has let the team clear a backlog of foundational assets that never had headroom before, including battle books and ideal customer profiles. "We used to have clear, logical breaks between completing one project and starting the next," he notes. "Now there are so many things constantly in flight. Even when I go to bed, I have Claude and other agents continuing to run overnight and in the morning, I pick it back up."

The productivity gains marketing organizations are reporting come with an equally large increase in the amount of work those organizations are now able to take on, and the two forces cancel each other out at the human level. Bradley's framing is that AI adoption in enterprise marketing is separating into two kinds of projects: efficiency work aimed at doing existing tasks faster, and opportunity work aimed at completing tasks that used to be impossible. "There is efficiency AI, where you can simply be faster at what you already do, but the opportunity side of AI means you can accomplish so much more than you ever could before," he explains. "Because of that, the human workload will either stay the same or get bigger."

The hands-on CMO requirement

The same approach Bradley applies to marketing workflow redesign is now moving outward into Veritiv's commercial operations. His team is deploying computer-use agents against the order-entry backlog, where AI's operational impact is starting to compound at scale. "Now we use AI to quickly see and view the entire worldview of a given segment, identify every possible target, generate the total addressable market for each one, and prioritize them," Bradley says. "If a human is going through a user interface, a task might take 50 steps. We have the agent log in through that same interface, but once they're in, they use an API through the browser to complete all the steps in half a second."

The specific project the team is running is a computer-use application targeting one million manually keyed orders that Veritiv processes each year, with the goal of redirecting the associated hours toward customer-facing work. Bradley's position is that most marketing leaders will only recognize this level of opportunity if they have direct working knowledge of what the tools can do, which means personally experimenting with agents, understanding the technical ceiling, and functioning as hands-on architects of AI-enabled operations. The scale of the outcome is what he uses to make the case concrete. "Instead of doing manual data entry, we're putting humans onto higher-level work and serving our customers better," Bradley concludes.