Manufacturers have invested in digital systems, but information still fails to travel reliably between design, production and quality. New research into US manufacturing reveals why these gaps are distorting assessments of readiness and limiting the returns from automation and AI.
A manufacturer can invest heavily in digital technology and still depend on employees manually transferring information between systems. These workarounds can keep production moving, but they also conceal weaknesses in how departments communicate and share information. The result is a factory that appears connected at management level while the people working on the production floor experience something quite different.
This contradiction runs through the 2026 America’s State of Manufacturing Report: Manufacturing’s Reality Check, published by Hexagon Manufacturing Intelligence. Based on a survey of 511 US manufacturing professionals conducted by Dynata in July 2026, 77 percent regard their operations as fully or mostly connected, but only 40 percent describe a complete connection between design, production, quality and business systems. The survey records respondents’ assessments rather than independently verifying their technology, but the difference highlights how elastic the description ‘connected’ has become.
The practical difficulty becomes apparent when a design change reaches production without the corresponding information being available to the inspection team. Although the change may be recorded in several digital systems, employees can still be left to establish manually which requirements apply and whether the production process has been adjusted accordingly. These workarounds can keep a factory operating, allowing weaknesses in communication to persist long after new technology has been introduced.
Factory connectivity looks different across the workforce
One of the report’s most revealing findings concerns who believes the systems are working. Eighty per cent of executives consider their manufacturing data integrated, compared with 33 percent of entry-level employees. The groups may be judging different things: senior leaders may be thinking about platforms and management reports, while operators judge whether information is available quickly enough to complete a task. Yet that difference is important because operational readiness depends on the information available when a decision must be made, not simply on what a management dashboard can display.
The report also identifies differences between seniority groups in their confidence about responding to disruption and their assessments of lights-out production. These responses cannot establish which group has a more accurate picture, particularly given the relatively small numbers of executive and entry-level respondents. The research also does not demonstrate that employees at different levels were describing the same factories. Nevertheless, the findings reveal considerable differences in how manufacturing professionals assess connectivity and operational readiness, raising questions about whether management perceptions adequately reflect the practical experience of production employees.
Part of the difficulty lies in the informal work that keeps production moving. Employees learn which colleague has the latest specification, which spreadsheet reconciles two incompatible systems and how to interpret an inspection result that arrives without sufficient context. Such knowledge is valuable, but depending on it leaves the process exposed when experienced workers change roles or production requirements become more complex. The report identifies the transfer of expertise between teams as another persistent challenge, suggesting that the loss of context can be as damaging as the loss of data.
For connectivity to deliver practical benefits, manufacturers need to preserve the relationship between inspection results, design requirements and production conditions. An inspection engineer should be able to trace a component back to the relevant design revision, while the production team needs evidence showing whether a corrective adjustment has reduced variation. Making that information accessible without manually reconstructing records would give both teams a more reliable basis for improving processes and assessing operational readiness.
Quality reveals the cost of missing connections
The consequences become tangible at the point where production and inspection meet. Hexagon reports that manufacturers are increasingly bringing measurement closer to the production line: 59 percent of respondents say inspection now happens on or near the shop floor. That provides an opportunity to discover problems before more material, machine time and labour have been committed. However, proximity alone does not mean that a result reaches the people or systems capable of preventing the next defect.
Consider a measurement that shows a gradual movement towards the edge of an acceptable tolerance. If the result is available to the operator and linked to the machine and production conditions, the team may be able to investigate before parts begin to fail inspection. If it is held in a separate report, the same information may be reviewed only after a batch is complete. The measurement equipment has done its job in both cases, but the wider manufacturing process has benefited from it only in the first.
The accumulated cost appears in the responses about inspection delays. Two-thirds of those surveyed report losing at least six hours a week to measurement and inspection bottlenecks, while rework and late discovery of quality problems also interrupt production schedules. These are self-reported experiences rather than a quantified industry-wide productivity loss. Even so, they show why treating quality as a separate department creates consequences for planning, capacity and delivery, especially when teams cannot distinguish an isolated defect from a persistent process problem.
Closing this gap requires more than transferring inspection files into a central repository. Measurements need to retain their relationship to the component, design revision, production operation and conditions under which they were taken. Engineers must then be able to use that evidence to decide whether to adjust a process, review a tolerance or change a future design. A connected digital thread is useful precisely to the extent that it makes these relationships accessible and keeps feedback moving in both directions, rather than simply preserving a longer electronic record.
Automation depends on what happens between systems
Manufacturers attempting to extend automation beyond individual machines face a more complicated challenge than improving the efficiency of a single operation. Automated equipment depends on accurate information from other parts of the production process, and the data it generates must be available to the teams responsible for managing subsequent activities. Where systems remain disconnected, automating one operation may simply transfer delays and manual intervention to another part of the factory. As manufacturers give software and equipment greater responsibility for operational decisions, the reliability of the information passing between systems becomes increasingly important.
Cybersecurity has emerged as the most cited obstacle to faster automation in Hexagon’s survey, ahead of cost and return-on-investment concerns. That finding sits alongside the connectivity problem. Exchanges between operational and business systems can widen the consequences of failure or unauthorised changes. Manufacturers need to know not only that information can be shared, but who is allowed to alter it, how changes can be traced and what happens when one part of the process becomes unavailable. Reliable connections must be governed as well as established.
AI increases the pressure to solve these familiar problems. The report finds that around a third of manufacturers do not know where AI could deliver value in their operations. It is difficult to identify a worthwhile application when evidence about recurring delays, process variation or inspection outcomes remains scattered across different tools. A model may recognise a pattern in the information it receives, but it cannot provide a complete account of the production process if important context is missing or if its findings never reach someone able to respond.
The workforce challenge belongs in this discussion because people still provide much of that missing context. Manufacturers are training employees for AI-assisted and automated environments, yet 90 percent of respondents say their workforce needs are not fully met. The report draws particular attention to metrology, where employees need both practical measurement knowledge and the ability to work with increasingly digital equipment. It describes a partnership between Hexagon and Fullerton College in California in which students first learn to work with manual measuring instruments before moving to software, an approach intended to develop judgement alongside technical competence.
Bringing together data, expertise and accountability requires more than a technology rollout. It asks manufacturers to document how decisions are made, make information usable beyond the department that produced it and involve production employees in identifying where those links fail. Rather than treating every system as a candidate for replacement, Hexagon recommends beginning with the specific handover that obstructs useful information and addressing subsequent gaps in turn.
For all the investment in manufacturing technology, the research exposes a persistent weakness in how factories use the information they collect. Too much still depends on employees recognising that information is missing and finding a way to retrieve it. Addressing those gaps would give manufacturers a more credible measure of digital progress than the number of connected systems they have installed.
