The Boring Parts Problem: Why Manufacturing Leaders Can't Get Excited About What Actually Works

Here’s a behavioral puzzle hiding in plain sight: manufacturing COOs are systematically investing in AI tools that look impressive while under-investing in the capabilities that determine whether those tools actually work. They know this. The data is clear. And yet the pattern persists.

This isn’t about ignorance or incompetence. It’s about psychology—the cognitive and organizational forces that make certain investments irresistible and others invisible, regardless of what the evidence says about effectiveness.

Understanding this dynamic matters because it explains why 66% of manufacturers remain stuck in pilot purgatory despite massive AI spending commitments. It also suggests what would need to change for the industry to break through.

The Attention Economy of Executive Decision-Making

McKinsey’s recent COO100 Survey asked manufacturing executives to allocate 100 points across different AI investment categories. The results reveal a consistent pattern: applications beat enablers, every time.

Shop floor automation and robotics: 15 points (highest). AI-driven process optimization: 14 points. Factory operations and control: 13 points.

Workforce enablement: 9 points. IT/OT infrastructure: 9 points. Cybersecurity: 7 points.

The disparity isn’t random. It reflects how organizational attention actually works.

Robots are tangible. You can photograph them for the annual report, tour them with the board, feature them in recruitment videos. When a CEO asks “what are we doing with AI?” the answer “we’ve deployed collaborative robots on three production lines” lands very differently than “we’ve been upgrading our data infrastructure and training frontline workers.”

The visibility bias compounds at every level of decision-making. Investment proposals for automation include dramatic before-and-after comparisons, impressive ROI projections, and clear narratives about competitive advantage. Proposals for workforce training describe activities that are harder to visualize and outcomes that are harder to measure.

Same budget competition. Different psychological weight.

The Conference Keynote Test

There’s an informal heuristic I’ve noticed in technology investment: if you can’t imagine describing the investment in a conference keynote, it’s harder to get funded.

Imagine a manufacturing executive at an industry conference. Which of these generates more audience interest and personal satisfaction?

“We’ve deployed digital twins across our production network, enabling real-time simulation of entire production scenarios…”

Or:

“We’ve spent two years cleaning up our data architecture and training our workforce to actually use AI recommendations…”

The first sounds like leadership. The second sounds like table stakes—necessary but unremarkable, the kind of work that happens invisibly if it happens at all.

This isn’t vanity. It’s how executive attention is structured in large organizations. Leaders need narratives that communicate strategic direction internally and externally. They need investments that signal vision and competence. Unglamorous enablers struggle to serve these purposes even when they’re essential to outcomes.

The Pilot as Organizational Comfort Zone

The survey reveals that 24% of manufacturers are still in “exploration and testing” mode, while 42% have achieved only “targeted implementation”—pilots in specific areas that haven’t scaled. Combined, that’s two-thirds of the industry stuck short of meaningful integration.

Many have been stuck for years. The pilot has become something more than a testing phase—it’s become an organizational comfort zone.

A pilot demonstrates innovation. It satisfies boards and shareholders who expect AI investment. It generates learning, produces case studies, wins internal recognition. And crucially, it avoids the hard work of organizational transformation.

Scaling AI isn’t a technical challenge. It’s an organizational one. It requires changing how people work, what they’re measured on, how decisions get made, what skills matter. This is uncomfortable, politically risky, and slow. It creates winners and losers within the organization.

A pilot creates none of these problems. It sits safely in a designated innovation space, staffed by willing volunteers, insulated from the organizational antibodies that attack real change. Success can be celebrated. Failure can be quietly absorbed as “learning.”

The psychological comfort of perpetual piloting explains why companies with substantial AI budgets remain stuck in exploration mode. They’ve found a sustainable steady state—enough investment to claim innovation credentials, not enough transformation to face real organizational disruption.

Why We Ignore What the Data Tells Us

Half of surveyed COOs cite cultural shift as their biggest implementation challenge. Nearly as many point to reskilling needs. These are the problems. This is where transformation stalls.

Yet workforce enablement receives bottom-tier prioritization.

The disconnect illuminates something deeper about organizational decision-making. Executives can acknowledge problems intellectually while systematically under-investing in solutions behaviorally. The phenomenon has several reinforcing causes.

First, abstraction distance: “cultural shift” is abstract and systemic, while “robot deployment” is concrete and specific. Abstract problems are harder to act on than concrete solutions, even when the abstract problem determines whether concrete solutions work.

Second, time horizon mismatch: training investments pay off over years, while automation investments can show short-term productivity wins. Quarterly pressure favors fast returns even when slow returns are more valuable.

Third, accountability diffusion: if a robot deployment fails, there’s clear ownership and visible evidence. If culture change fails, responsibility is diffuse and failure is gradual. Organizations systematically underweight problems with diffuse accountability.

Fourth, expertise gaps: COOs understand manufacturing operations deeply. They may understand people and culture less deeply—or at least feel less confident evaluating workforce transformation proposals than technology proposals.

The Companies That Escaped

The exceptions prove instructive. Chilean lithium producer SQM has successfully deployed AI that enables frontline employees to make real-time production decisions optimizing output while minimizing water and energy use. Their leaders credit training investments as critical to success.

A global pharmaceutical company recognized its legacy IT/OT systems were too siloed for AI to scale. Rather than layering more applications on broken infrastructure, they invested in three integrated data platforms connecting multiple systems before deploying advanced applications.

What these companies share isn’t superior technology. It’s superior organizational self-awareness. They recognized the enabler gap before their investments became stranded, and they prioritized boring fundamentals even when flashier alternatives beckoned.

What Would Change the Pattern

For individual companies, the path forward involves structural changes that override natural biases:

Making enablers visible by tracking and reporting workforce AI capabilities, data quality metrics, and integration progress alongside application deployments. What gets measured gets attention.

Bundling enablers with applications so that investment proposals must include workforce development and infrastructure requirements as part of the business case, not optional add-ons.

Creating accountability for scale by evaluating AI leaders on integration achieved, not pilots launched. This shifts incentives toward the hard work of organizational transformation.

For the industry broadly, change requires new narratives. When training investments and data infrastructure become as prestigious as robotics deployments—when conference keynotes celebrate the boring parts—organizational attention will follow.

The Deeper Question

Beneath the organizational dynamics lies a more fundamental question about how we evaluate progress.

We’re drawn to visible transformation—the factory floor that looks different, the process that runs faster, the technology that impresses visitors. These changes feel like progress because we can perceive them directly.

The enablers that make visible transformation sustainable are largely invisible. Better-trained workers making better decisions. Cleaner data flowing through integrated systems. Governance structures that catch problems before they compound. These capabilities don’t photograph well.

The manufacturing AI experience offers a lens on this broader pattern. Billions of dollars flowing toward impressive technology, stuck short of impact because the boring parts weren’t exciting enough to fund adequately.

The companies that recognize this pattern—and structure their organizations to invest against it—will separate from the pilot-purgatory majority. Not because they bought better technology, but because they understood what actually makes technology work.

Based on findings from McKinsey’s COO100 Survey conducted June-July 2025 among 101 COOs at manufacturers with $1B+ revenues.

Source: McKinsey COO100 Survey, June-July 2025