Cross-Function Demand Coordination
Planning, procurement, and logistics agents now negotiate the next day's production schedule every night through an agent-to-agent protocol, retiring the weekly Monday planning meeting.
The challenge
Across 180 active SKUs, the plant's production schedule was fixed once a week in a Monday planning meeting between planning, procurement, and logistics leads. In between meetings:
- A schedule locked on Monday had to absorb five days of overnight demand and supply shifts with no formal adjustment mechanism
- Planning, procurement, and logistics each worked from a separate system with no shared negotiation process
- Late-breaking supplier delays were discovered only when the next shipment failed to arrive on schedule
- Expedited freight was routinely used to cover gaps the weekly plan hadn't anticipated
- Schedule adherence sat at 68%, with most misses traceable to information that was already known but not yet incorporated
- Exception handling consumed a growing share of the Monday meeting itself, crowding out actual planning
How it works
Replacing a weekly meeting with a nightly negotiation
Instead of adding more meetings, the three functions were modeled as agents that negotiate directly:
- 01
Modeled planning, procurement, and logistics as autonomous agents with explicit objectives and constraints
- 02
Built an agent-to-agent negotiation protocol that exchanges proposals and constraints every night
- 03
Fed each agent the latest overnight demand signals, supplier ETAs, and transport capacity as they update
- 04
Set the agents to converge on a committed next-day production schedule by a fixed midnight cutoff
- 05
Routed only cases where agents can't converge to a human planner, with the disagreement summarized
- 06
Automated schedule publication to shop floor systems immediately after the midnight commit
- 07
Kept a lightweight weekly meeting focused solely on reviewing unresolved exceptions
What we built
Key capabilities
Nightly re-planning
The production schedule is renegotiated every night against current demand and supply data, not held fixed for five days.
Autonomous cross-function negotiation
Planning, procurement, and logistics agents resolve tradeoffs directly instead of waiting for a weekly meeting to surface them.
Exception-only human involvement
Planners now engage only on the cases agents genuinely can't resolve, with the disagreement already summarized.
Same-day incorporation of new information
A supplier delay discovered at 9pm is reflected in the following morning's schedule instead of waiting for next Monday.
Before vs after
What changed in production planning
- Schedule commitment cadence
- Weekly Monday meeting → nightly midnight commit
- Schedule adherence
- 68% → 91%
- Expedited freight spend
- Baseline → –24%
- Planning data freshness
- Up to 5 days stale → under 24 hours
Business impact
What it changed
Schedule adherence up 23 points
Adherence rose from 68% to 91% once the plan incorporated overnight demand and supply signals nightly instead of weekly.
24% less expedited freight
Fewer late-breaking gaps meant fewer schedule fixes had to be covered with expedited shipping.
Planning meetings shrank to exceptions only
The weekly meeting now reviews the handful of cases agents couldn't resolve, instead of rebuilding the whole week's plan from scratch.
Technology stack
“The plant didn't need a better meeting — it needed the schedule to keep learning between meetings. Agents negotiating overnight did exactly that.”
Keep reading