Why slow AI adoption is more expensive in manufacturing

In some industries, you can afford to test at the edges, and it can be best practice. In manufacturing, you can’t, and it isn’t.

As part of NTT DATA’s 2026 Global AI Report research, which spans 35 countries, five regions and 15 industries, we’ve published the 2026 Global AI Report — Manufacturing and Automotive, revealing how AI leaders in these industries translate strategy into measurable impact for software-defined vehicles, engineering and intelligent manufacturing.

Our data shows that AI leaders, so classified based on their reported AI strategy, level of AI adoption and financial outcomes achieved from AI,are far more likely than other organizations to apply AI directly to core operations. Over 90% are already doing so, compared with less than 70% of laggards, the opposite cohort. Because most enterprise value in these industries is found in production, supply chain and operations, you simply won’t move the needle significantly if you don’t apply AI to those core workflows.

In other words, effective organizations focus less on scaling AI and more on finding the most productive places to apply it.

• ALSO READ → From software-defined to AI-native: Automotive’s next operating model

Focus, more than scale, is where AI starts to deliver

Manufacturing AI leaders start with a small number of high-impact domains such as production planning, quality, maintenance and supply chain execution. In these areas, even incremental improvements translate into meaningful gains in throughput, cost and reliability.

More importantly, they don’t treat AI as a layer on top of existing processes. They redesign workflows end to end, a shift that’s subtle but important.

Our data shows that 38.6% of leaders, versus only 12% of laggards, are rebuilding core systems with embedded AI. While you might create local efficiencies by automating individual steps, improving system-level performance means reworking the entire flow.

When you do that, you see the payoff in areas such as more consistency in how your organization makes decisions about quality, scheduling and maintenance. You’re also better positioned to respond in real time to issues like line disruptions, demand shifts or equipment failures.

AI from pilot projects to production: How leaders differ

Leaders are moving decisively from pilot projects into production environments — by which I mean they’re embedding AI directly into operational systems. They’re also more likely to invest based on early success, creating a flywheel effect where initial wins justify further investment and accelerate progress.

Laggards, on the other hand, tend to remain in an unproductive loop of pilot projects and fragmented initiatives. They wait for better data, more certainty or clearer use cases, which often delays the precise learning required to make AI effective at scale.

Rethinking risk: Speed with discipline

Another difference is how leaders think about risk.

There’s a perception that moving faster with AI introduces more risk. Quite often, it can. But in manufacturing, the opposite can be true as well, especially for manufacturing AI leaders who establish governance, guardrails and ownership early.

In safety-critical environments, centralized governance, clear accountability and a stronger connection between AI initiatives and plant-level outcomes enable organizations to scale AI without compromising control. In fact, the data shows that nearly two-thirds of leaders have centralized governance models in place.

Adoption happens when the frontline sees value

Effective manufacturers use AI to augment experienced engineers, operators and planners. That approach improves decision quality while keeping humans firmly in the loop — which matters more than it might seem.

In my experience, adoption comes from trust, and trust is built when frontline teams see AI as helping them perform better, not as threatening their jobs. That is reflected in adoption. Over 80% of leaders report positive sentiment toward AI across their workforce, nearly double that of laggards.

Once that confidence takes hold, progress accelerates. You start to see a shift from push to pull. Teams begin asking for more use cases, and investments become easier to justify as the AI adoption momentum builds. AI moves from being an initiative to becoming part of how the business runs.

The broader implication is clear: In manufacturing, AI leadership is quickly becoming indistinguishable from operational leadership.

• ALSO READ → The new automotive mandate: Embed AI throughout the value chain