Most supply chain teams operate against metrics tied to their specific role. Procurement focuses on unit cost reductions. Manufacturing prioritizes utilization and output. Distribution emphasizes throughput and efficiency. Transportation teams optimize for consolidation and on-time delivery.
What’s Related
Each of these goals makes sense in isolation. Together, they often pull the organization in different directions.
When demand shifts or disruption hits, teams respond according to what they’re held accountable for, not what the operation needs most in that moment. Inventory is reduced to protect working capital even as service levels suffer. Production runs ahead of demand to preserve utilization. Transportation plans prioritize efficiency over flexibility.
These outcomes are rarely the result of poor judgment. They are the predictable consequence of scorecards that reward local success without accounting for downstream impact, even as organizations gain access to more data, better visibility, and increasingly sophisticated decision support.
That tension is playing out at scale: in McKinsey’s global supply chain risk report, 82% of leaders said new tariffs are already affecting their supply chains, with up to 40% of supply chain activity disrupted in some organizations. These disruptions create the kind of cross-cutting trade-offs most companies struggle to resolve quickly.
Understanding where execution slows, and why authority, incentives, and decision design matter more than insight alone, is key to closing the gap.
Why this structure held for so long
For many organizations, this operating model was a rational response to growth. As supply chains expanded across regions and product lines, local ownership made complexity manageable. Teams knew what they owned, how success was measured, and where responsibility began and ended.
Decision cycles also moved more slowly. Trade-offs developed over time, not in real time. That gave organizations room to escalate issues, negotiate priorities, and absorb inefficiencies without immediate consequences. Informal coordination filled the gaps.
Over time, those workarounds became normalized. Structural friction was treated as part of operating at scale, not as a problem to solve.
When insight starts moving faster than decisions
That balance changes once insight accelerates. Real-time data and advanced analytics surface conflicts earlier, often before teams feel their downstream impact. Instead of reacting after the fact, organizations can see pressure building as conditions shift.
What’s different now is not just the volume of information, but how directly it points to action. AI-driven systems increasingly surface specific recommendations rather than broad signals. Some go a step further, using agents to continuously evaluate conditions and propose adjustments as situations evolve. Decisions that once waited for weekly reviews or post-mortems are now presented in near-real time, with clear implications for inventory, transportation, customer commitments, and costs.
Yet that speed gap is widening. McKinsey found that while 75% of companies are planning or piloting AI use cases, only 19% are deploying AI at scale, suggesting that insights are advancing faster than the decision structures needed to act on them. When that acceleration collides with organizations that lack clear authority across ownership boundaries, execution friction becomes unavoidable.
A practical way out of the stalemate
For organizations stuck in this pattern, the way forward isn’t a sweeping reorganization. Most don’t have the appetite, time, or political capital for that.
It’s also not a multi-year digitization project or a data normalization effort that must be completed before AI can deliver value. Those approaches feel like progress, but mostly delay it.
What does work is narrowing the problem. Instead of trying to align the entire organization, teams can focus on the workflows that break down when conditions change. Appointment scheduling during volume spikes. Carrier coordination when delays cascade. Document collection that holds up invoicing. These are usually well-known internally.
Once those moments are identified, the question shifts from “who decides” to “does a human need to decide at all?” Many cross-functional breakdowns aren’t judgment calls. They’re coordination tasks that fall between teams and wait for someone to pick them up.
AI agents are useful here because they execute. They reschedule appointments, follow up with carriers, and pull documents together. The work gets done without requiring org charts to change.
Stephen Dyke is Principal Solutions Consultant Manager at FourKites.
