80% Report AI Productivity Gains. Only 37% Report Positive EBIT Impact.
AI is making people more productive. So why isn’t that productivity showing up more clearly in business performance?
McKinsey’s latest research puts unusually sharp numbers around the problem:
- 80% of respondents say AI has improved their individual productivity.
- Only 37% report positive EBIT impact.
- Just 6% of organizations qualify as AI high performers. (McKinsey & Company)
That gap tells us something critical about where the operational bottleneck is moving. The technology is increasingly capable, and employees are increasingly adept at using it. But the organization around them often remains largely unchanged.
The missing link is the organization itself.
Productivity Is Not Enterprise Value
It is easy to understand why companies focused first on individual productivity. AI can draft faster, analyze faster, research faster, code faster, and summarize faster. Those tactical gains are visible almost immediately.
But saving an employee an hour does not tell leadership what happens to that hour next:
- Does the company serve more customers?
- Improve quality?
- Reduce operational cost?
- Take on higher-margin work?
- Increase revenue?
- Remove an operational bottleneck?
- Or simply produce more of the same work?
McKinsey makes the distinction explicit: individual productivity gains do not automatically translate into enterprise value. Capturing that value requires leaders to decide where the released capacity should go. (McKinsey & Company)
Capturing enterprise value is not primarily an AI question. It is a management decision.
The High Performers Are Redesigning Work
The operational divide between AI high performers and everyone else is revealing. Nearly three-quarters of high performers report substantially redesigning workflows, compared with only about one-quarter of other organizations.
Meanwhile, only 11% of organizations have reached what McKinsey calls the “reinvention” stage—where companies rethink how work gets done and how value gets created rather than simply bolting AI onto legacy processes. (McKinsey & Company)
AI tool → Productivity gain → Workflow redesign → Organizational change → Enterprise value
Many enterprises are getting stuck near the beginning of this progression. They have changed the tool, but they have not yet changed the system.
From the Org Chart to the Orchestration Chart
One of McKinsey’s most actionable concepts is the shift from the traditional org chart toward what it terms an orchestration chart. Instead of beginning with existing departments and static roles, leadership starts with the target business outcome and asks:
Which humans and agents should come together to produce this outcome?
Once the outcome becomes the starting point, the entire operating system opens for evaluation:
- Which steps are still necessary?
- Which handoffs should disappear?
- Which execution steps can move to autonomous agents?
- Where does human judgment create the most value?
- Who owns the final decision?
- Where should decision authority sit?
- Which legacy approvals are purely friction?
- Which management layers exist primarily to route information that AI can coordinate directly?
Agent autonomy, system access, oversight, escalation protocols, human judgment, and accountability increasingly need to be designed directly into operational workflows. (McKinsey & Company) That goes far beyond task automation; it is deliberate organizational redesign.
AI May Remove Coordination—but It Cannot Remove Decision Architecture
Internal coordinator roles—positions focused heavily on shepherding decisions through organizations—have grown roughly 1.5 to 2 times as fast as line roles over the past five years. AI creates an opportunity to strip out that coordination overhead. (McKinsey & Company)
However, if companies flatten coordination layers, something robust must replace the alignment those layers provided. McKinsey identifies three fundamental requirements:
- Better intelligence
- Clearer decision rights
- Stronger human judgment
A flatter organization without explicit decision architecture is not more efficient—it is simply more ambiguous.
People May Be More Ready Than Their Organizations
The disconnect between individual adoption and enterprise readiness is striking:
- 70% of respondents say they are personally ready for AI.
- Only 27% of leaders believe their organizations are ready for the changes required by an agentic operating model. (McKinsey & Company)
For years, executive discussions focused on employee adoption: Will people use it? Will they resist? Do they have the skills? While those factors still matter, the critical challenge has inverted: Is the organization ready to restructure around the workforce already using AI?
Widespread AI usage routinely coexists with near-zero organizational transformation. Employees leverage AI daily while workflows, governance, incentives, and operational frameworks remain frozen in legacy formats. This structural stagnation directly explains the productivity-to-EBIT gap.
Start With the Mission, Not the Tool
McKinsey’s proposed “agentic mission factory” provides a clear design principle: do not start by asking which agents to deploy. Start with a measurable business outcome:
- Approve a qualified loan in minutes rather than days.
- Resolve an enterprise support ticket in a single interaction.
- Eliminate claims leakage.
- Accelerate technical product development cycles.
- Strengthen supply-chain resilience against supply shocks.
From that outcome, engineer the combined human-agent workflow, financial model, controls, access boundaries, accountability structures, and enabling technology. (McKinsey & Company)
Start with the mission, not the tools.
The Bottleneck Is Moving
Raw model capability and individual efficiency will continue to compound, but neither guarantees bottom-line impact. The operational leverage now sits in the execution chain between capability and outcome:
Outcome → Workflow → Decision → Authority → Human/AI Allocation → Controls → Action → Measurement → Learning
This is organizational decision design. The next competitive phase will not be won by organizations deploying the highest volume of agents, but by those that rethink value creation and make the hard structural choices required to deliver it.
The leadership question has shifted:
Not: How much more productive is AI making our people?
Rather: What have we redesigned in the organization so that productivity converts directly into enterprise value?