New research from Hexagon reveals a growing disconnect in manufacturing: most companies believe their operations are connected, but far fewer have data flowing across the entire production process. For manufacturers looking to scale AI, that gap could become one of the industry’s biggest challenges.
Artificial intelligence is moving deeper into manufacturing.
But before manufacturers can expect AI to transform production, there is another problem to solve: getting the factory’s data to work together.
Hexagon’s 2026 America’s State of Manufacturing Report, based on a Dynata survey of 511 U.S. manufacturing professionals, found that 77% of manufacturers describe their operations as fully or mostly connected.
Only 40% say they are connected across design, production, quality and business systems.
That difference is important.
A factory can have connected machines, digital inspection systems and multiple software platforms without having a truly connected manufacturing workflow.
For AI, that distinction matters because an algorithm can only work with the information available to it.
And manufacturing has plenty of information.
The problem is making sure the right information reaches the right system at the right time.
AI Adoption Is Moving Beyond the Job Replacement Debate
The conversation around AI in manufacturing is changing.
Hexagon’s research found that only 7% of manufacturers expect AI to reduce headcount, down from 18% a year earlier.
At the same time, 79% expect their workforce either to grow through AI-created roles or remain stable while existing roles change.
Another 79% are actively training workers for AI-assisted or automated environments.
That suggests manufacturers are increasingly looking at AI as a tool for augmenting existing operations rather than simply replacing workers.
The question is becoming less about whether AI belongs in the factory and more about where it can create measurable value.
That could mean identifying quality problems earlier, supporting maintenance decisions, improving process control or helping workers make better decisions.
But each of those applications depends on one thing.
Data.
The Factory Is Full of Data. It Just Isn’t Always Connected
A modern manufacturing operation generates data at almost every stage.
CAD systems contain design information.
CAM systems contain programming and toolpath information.
CNC machines generate production data.
Inspection systems generate measurement results.
MES and ERP platforms contain production and business information.
Maintenance systems collect information about machine condition.
The problem is that these systems do not always communicate effectively.
Hexagon found that 39% of manufacturers report limited feedback between design, manufacturing and quality teams.
Another 38% say they discover quality issues too late, while 37% struggle to share expertise across teams.
This creates a familiar problem.
The information exists, but it is separated.
An inspection system may know that a measurement has moved outside an expected range.
The machine may have data showing a change in operating conditions.
The engineering team may know that a particular feature has caused problems before.
And an experienced operator may already recognise what is happening.
But if those pieces of information remain isolated, AI cannot easily connect them.
The factory can be data-rich without being data-connected.
77% Say Connected. Only 40% Say Fully Integrated.
This is the figure from Hexagon’s research that deserves particular attention.
77% of manufacturers describe their operations as connected.
But only 40% say they are connected across design, production, quality and business systems.
The distinction is more than a matter of terminology.
A machine being connected to a network does not mean its data is being used effectively across the manufacturing process.
A CMM can generate inspection results without automatically feeding those results back into production.
A CAM system can be connected to manufacturing without receiving meaningful feedback about what happened to the part after machining.
And a machine-monitoring system can collect huge amounts of information without that information being connected to quality or maintenance decisions.
For AI, connectivity alone is not enough.
The data has to be accessible, relevant and connected to context.
AI Needs the Digital Thread
Consider a typical machining workflow.
A part begins as a CAD model.
The design moves into CAM.
A toolpath is created.
The machine produces the component.
The finished part is inspected.
The inspection system identifies a variation.
That measurement could be useful to production.
It could also be useful to engineering.
It could provide another signal for maintenance.
And when combined with historical production data, it could help AI identify a pattern that would otherwise be difficult to see.
But that only works when the information can move through the process.
This is where the digital thread becomes increasingly important.
AI does not simply need more manufacturing data.
It needs connected data with context.
That could become one of the defining requirements for the next stage of industrial AI.
Quality Data Could Become an AI Input
Quality is one of the clearest examples of where connected data could change the role of AI.
Hexagon’s research found that 59% of manufacturers now carry out inspection on or near the shop floor, while 15% remain mostly or entirely laboratory based.
Moving inspection closer to production can reduce the distance between discovering a problem and acting on it.
But the data still has to move.
Hexagon found that 67% of manufacturers lose six or more hours every week to measurement and inspection bottlenecks, with 35% losing 11 hours or more.
That creates a significant opportunity for AI.
Today, inspection can tell a manufacturer whether a part meets its specification.
With the right data infrastructure, AI could potentially help identify whether the manufacturing process itself is beginning to drift.
That is a different question.
Instead of asking only:
Did we make the part correctly?
Manufacturers can start asking:
Is the process beginning to move in the wrong direction?
From Detecting Problems to Predicting Them
This is where connected manufacturing data becomes particularly interesting.
A small change in measurement results might not be significant on its own.
A change in measurement combined with machine conditions, tool life, historical production data and previous inspection results could tell a very different story.
AI can potentially identify those relationships across large datasets much faster than a person working with isolated reports.
That could support earlier intervention.
The result is a shift from AI detection toward AI-assisted prediction.
It does not mean every manufacturing problem can suddenly be predicted.
It means manufacturers have the possibility of moving from reacting to individual events toward identifying patterns across the production process.
That is a much more meaningful role for AI than simply adding a chatbot or another automated feature to an existing system.
The AI Skills Gap Is Part of the Same Problem
Technology is only one side of the equation.
Manufacturers also need people who understand how to use increasingly intelligent systems.
Hexagon found that 90% of manufacturers say their workforce needs are not fully met.
Seventy-nine percent identify a skills gap either alone or alongside a headcount shortage.
At the same time, 79% say they are actively training workers for AI-assisted or automated environments.
Manufacturers also identified upskilling existing workers as the area where they most believe they outperform the market, at 48%, followed by adopting automation at 41% and implementing AI at 38%.
That points to an important part of the AI transition.
The future factory does not simply need more technology.
It needs manufacturing workers who understand the processes behind the data and can work with the systems interpreting it.
Executives and Operators Do Not See the Same Factory
One of the more revealing findings in the report is the difference between management and shop-floor perspectives.
80% of executives say their data is integrated.
Only 33% of entry-level workers agree.
That is a substantial gap.
It matters because digital transformation is ultimately experienced on the shop floor.
Operators see the manual workarounds.
They know when information arrives too late.
They know which systems do not communicate.
They know which processes still depend on one experienced employee.
They also know when a digital workflow looks efficient on paper but creates additional work in practice.
For manufacturers developing AI strategies, those experiences are important sources of information.
An AI system can process data.
It cannot automatically know that a particular production report is being recreated manually every afternoon unless someone captures that reality in the digital workflow.
Automation Is Still Moving Forward
AI is arriving alongside a broader automation push.
Hexagon found that only 24% of manufacturers believe they are ahead on automation, while 46% say they are keeping pace and 30% say they are behind or have not started.
The barriers are also becoming more practical.
Cybersecurity is the most frequently cited barrier to accelerating automation, at 27%, followed by cost and ROI concerns at 22%.
These concerns are relevant to AI as well.
Manufacturers need to know whether new systems can work with existing infrastructure.
They need confidence in the data.
They need secure connections.
They need employees who can use the technology.
And they need a clear understanding of what the technology is actually improving.
That makes AI integration an operational issue, not simply a software decision.
The AI-Ready Factory May Not Need to Be Rebuilt
One of the most interesting implications of the research is that becoming more AI-ready does not necessarily mean replacing the entire manufacturing technology stack.
In many factories, the first step may simply be identifying where information stops moving.
It could be between design and production.
Production and quality.
Quality and engineering.
Machine monitoring and maintenance.
Or experienced employees and newer workers.
Fixing those connections can create a stronger foundation for AI without requiring manufacturers to start from scratch.
This is also where Hexagon is positioning its manufacturing portfolio, bringing together design, production, quality, automation and operational information.
The broader industry trend is clear: the value of AI increasingly depends on the infrastructure surrounding it.
AI’s Next Manufacturing Challenge Is Not the Algorithm
Manufacturing has moved beyond asking whether AI will enter the factory.
It already is.
The bigger question is whether manufacturers have connected the systems and information AI needs to understand what is actually happening.
Hexagon’s research highlights that challenge clearly.
77% say their operations are connected. Only 40% say they are connected across design, production, quality and business systems.
For manufacturers exploring AI for quality, predictive maintenance, process optimisation, automation and workforce support, that gap cannot be ignored.
AI cannot work with information it cannot access.
And a factory cannot become truly intelligent if its most important information remains trapped inside separate systems.
The next phase of manufacturing AI may therefore be less about adding another AI tool and more about connecting the factory that already exists.
The machines are generating the data.
The AI is becoming available.
The next challenge is connecting the two.
To see the complete findings, methodology and additional insights from Hexagon’s 2026 America’s State of Manufacturing Report, click here to read the full report.
MTN Analysis
Hexagon’s numbers point to a problem that is bigger than AI adoption itself.
Manufacturers can invest in AI without having the data foundation needed to make that investment useful.
The 77% versus 40% gap is therefore more interesting than the headline AI adoption figures. It shows that there is still a difference between having connected technology and having an integrated manufacturing information flow.
That distinction will become increasingly important.
AI can analyse a machine signal, but its value increases when that signal can be understood alongside tool life, production history, inspection results and the original design intent.
The same applies to quality.
A measurement result sitting inside a quality system is useful. A measurement result that can immediately provide context to production, engineering or maintenance is potentially much more valuable.
There is also a human side to this.
The gap between executive and entry-level perceptions of data integration suggests that manufacturers should not measure digital maturity only from the top of the organisation. The shop floor can reveal problems that are invisible in management dashboards.
For MTN, the bigger takeaway is that manufacturing AI is becoming an integration story.
The competitive question is increasingly not who has access to AI.
It is who can connect AI to the information, processes and people already operating the factory.
FAQ
What is the biggest AI challenge facing manufacturers?
One major challenge is connecting the data generated by different manufacturing systems. AI can be limited when machine, quality, production, engineering and business data remain isolated.
How many manufacturers say they are connected?
Hexagon’s 2026 research found that 77% of manufacturers describe their operations as fully or mostly connected.
How many manufacturers are connected across the full workflow?
Only 40% say they are connected across design, production, quality and business systems.
Why does data connectivity matter for AI in manufacturing?
AI needs relevant data and context to identify patterns and support decisions. Connecting data across the manufacturing workflow can give AI a more complete view of what is happening.
How can AI be used in manufacturing quality?
Connected AI systems can analyse inspection, machine and production data to identify patterns that may indicate process drift or emerging quality problems.
Does AI mean fewer manufacturing workers?
Hexagon found that 7% of manufacturers expect AI to reduce headcount, down from 18% in the previous year’s research. Meanwhile, 79% expect their workforce either to grow through AI-created roles or remain stable while existing roles change.
What is a digital thread in manufacturing?
A digital thread connects information across stages of the manufacturing lifecycle, including design, engineering, production, inspection, maintenance and business operations.
Are manufacturers training workers for AI?
Yes. 79% of manufacturers surveyed by Hexagon say they are actively training workers for AI-assisted or automated environments.
What is preventing manufacturers from accelerating automation?
Hexagon’s research identifies cybersecurity at 27% as the most frequently cited barrier, followed by cost and ROI concerns at 22%.
Further Reading
For more information about Hexagon’s manufacturing technologies, digital manufacturing and industrial solutions, visit the official Hexagon website.
- Hexagon Manufacturing Intelligence
- Hexagon: 77% of Manufacturers Call Their Operations Connected, Only 40% Really Are
- Hexagon’s 2026 America’s State of Manufacturing Report
- AI in Machining: The Real Value Starts Before the First Cut
- IMTS 2026 Post-Show Review: What the Biggest Manufacturing Show in the West Revealed About AI
