Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: Why Disconnected P&Ls Remain Biggest Barrier to Supply Chain AI in Simple Termsand what it means for users..
Supply chain leaders have spent the past few years talking about agility, responsiveness, and the transformative potential of artificial intelligence. Walk into any industry conference and you’ll hear executives describe their vision for AI-optimized inventory positioning, real-time network adjustments, and overcoming disruptions before they snowball.
Then those same executives return to organizations where demand planning reports to sales in one region and supply chain in another. Where three different business units maintain three different P&Ls, each incentivized to hoard inventory or cut corners in ways that make sense locally but hurt the network as a whole; where nobody below the VP of supply chain has authority over more than one node.
This disconnect has been around for decades. But 2026 may be the year it becomes impossible to ignore.
Fragmentation by design
A survey of 54 companies across various industries found that in almost one-third, demand planning is not systematically organized. It reports to supply chain leaders in some geographies and to sales or business leads in others. In about one in ten companies, order management and logistics have the same problem. One-fifth of organizations reported acute struggles with silos and difficulty in cross-business execution.
Even companies where “supply chain” nominally covers the entire plan-source-make-deliver system don’t solve the coordination problem through reporting structure alone. Competing incentives, capability gaps, and the sheer effort of pulling data together from different systems continue to get in the way.
A separate McKinsey case study shows how far this can go. A global industrial company had 60 businesses, each with its own P&L, often devising their own versions of similar processes but with different language and terminology for nearly every process. In one division, half of the job titles in a commercial function were unique to a single person. Sharing information or transferring skills across business units was nearly impossible.
Scorecards that work against each other
McKinsey also documented a consumer goods company where the sales force was measured on top-line revenue, which encouraged inflated sales projections. Supply chain employees were measured on inventory and write-offs, which led them to keep stock levels as low as possible. The result was frequent stock-outs and lost sales. Each function hit its numbers while the company missed its targets.
Procurement teams get rewarded for unit cost reductions that fragment the supply base. Manufacturing gets measured on utilization rates that encourage building ahead of demand. Distribution optimizes for warehouse efficiency while transportation optimizes for load consolidation. The goals conflict more often than they align, and nobody has both the visibility and the authority to make trade-offs across them.
Supply chain leaders know these dynamics exist. Most have been working around them for years. The question is what might finally force a reckoning.
AI exposes the fault lines
The push to implement AI across supply chain operations is accelerating, and these projects have a way of surfacing organizational problems that were previously easy to ignore.
AI systems need integrated data to function. They need aligned incentives to deliver business value. They need someone with cross-functional authority to act on their recommendations. When those conditions don’t exist, pilots stall, implementations fail to scale, and organizations end up with expensive technology that produces impressive demos but limited results.
A 2025 survey from Writer, cited by Slalom, found that 71% of executives report AI applications being created in silos; 68% note tension between IT teams and other business units over AI initiatives; and 42% of C-suite executives believe AI adoption is actively creating organizational rifts.
Getting AI to work in a controlled environment turns out to be the easy part. Getting it to work across an organization with fragmented authority and competing incentives is harder. As a result, too many large-scale change programs don’t reach their stated goals due to poor cross-functional collaboration and a lack of accountability. AI implementations are transformation programs, whether organizations recognize them as such or not.
The conversation that keeps getting postponed
Most supply chain organizations are structured in ways that prevent the outcomes their leaders say they want. The agility and responsiveness that executives describe at conferences would require someone to have authority across functions that currently operate as separate territories. The AI-optimized network would require incentive structures that reward enterprise outcomes rather than local metrics.
Addressing this means difficult conversations with business unit leaders who have built careers around their current authority. It means revisiting compensation systems that have been in place for years. It means acknowledging that organizational design choices made in a previous era may not serve the current strategy.
For years, organizations have been able to avoid these conversations by keeping transformation efforts modest enough to fit within existing structures. The AI push of 2026 may not allow that approach to continue. As organizations move from isolated pilots to enterprise-wide implementation, they will run into the structural contradictions they have been working around.
The technology will work. The data integration challenges are solvable. The harder question is whether anyone has both the authority and the incentive to act on what the systems recommend.
Somebody is going to have to ask whether the organization is actually structured to execute the strategy it claims to be pursuing. That conversation has been postponed long enough.