When Remix cohort 2 began, we posed a question: What happens when partners take patterns they have seen across supply chain transformations and turn them into agents?
Fourteen partner teams took on the challenge. After two weeks of building, the winning submissions gave us a few answers.
What stood out was not that teams had uncovered entirely new planning problems. Instead, the winners focused on familiar moments where decisions become difficult: when assumptions change faster than the planning cycle, when a commercial opportunity creates operational uncertainty, or when identifying risk is only the beginning of resolving it.
That distinction matters.
Familiar problems, reframed
EY, the overall winner of this cohort, explored what happens when events invalidate the assumptions behind an S&OP plan between formal planning cycles.
Accenture, winner of the Best Business Impact award, focused on the tension between capturing new demand and understanding whether the supply chain can realistically support it.
Capgemini, winner of the Most Innovative Agent award, looked at revenue risk and how planners can move from identifying exposure to evaluating possible responses and taking action.
Different use cases, but with one common thread: each addresses a planning pattern that appears repeatedly across organizations.
The opportunity for agentic AI is not simply to automate an individual task within those processes. It is to help connect the reasoning around them: understanding what changed, determining what matters, evaluating alternatives and helping a planner decide what to do next.
Experience becomes a design input
That is where the partner perspective becomes particularly interesting.
Working across multiple supply chains creates a form of pattern recognition. A capacity issue in one organization may look different from a service -risk problem in another, but the underlying decision journey can be remarkably similar.
Where is the exception? What is driving it? What options are available? What are the consequences of each response?
Those recurring questions can become useful starting points for agent design.
Rather than asking, “Where can we use an agent?”, the stronger question may be, “Which decisions do planners repeatedly struggle to move through?”
The cohort 2 winners approached that question from different directions, but each translated experience into a more focused role for the agent.
From novelty to usefulness
That may be one of the more important lessons from this cohort.
The value of agentic AI does not depend on finding problems nobody has seen before. In many cases, the opportunity is the opposite: taking a problem people know extremely well and reconsidering how the work around it could be orchestrated.
Partners are well positioned to do that because they have seen what repeats.
Remix gives them a place to turn those observations into something tangible: to test how an agent might support a decision, where human judgment still matters and how different parts of a planning process could come together differently.
Cohort 2 brought the perspective of partners into that experiment. Next, Remix shifts back to customers, bringing another group of supply chain teams into Agent Studio to build from the realities they encounter every day.
Different perspectives. Different patterns. And more to build from. Explore more highlights from the second cohort here.
