98% of Manufacturers Are Exploring AI. Only 20% Are Ready. Here Is Why That Gap Exists.
Ninety-eight percent of manufacturers are exploring AI and only 20 percent say they are ready. The gap follows the same pattern as ERP and CRM rollouts, where the tool arrived before the foundation did.
A new report from Redwood Software surveyed 300 manufacturing professionals and found that 98% are investing in or exploring AI. Only 20% consider themselves fully prepared to use it at scale. That is a 78-point gap between intent and readiness. If you have been around manufacturing long enough, that number is not surprising. It is familiar.
This is not an AI problem. It is the same problem that has followed every major technology implementation for the past 30 years. The tool arrives before the foundation is ready. And the result is always the same: money spent, expectations unmet, and a team left wondering what went wrong.
This Has Happened Before
Think about ERP implementations. CRM rollouts. Lean programs that never stuck. In each case, the pattern was the same. Leadership saw the potential. They bought the platform. They brought in the trainers. Then they ran into the real problem: the company's systems, structures, and workflows were built for how things already worked. Not for what the new tool required.
AI is different in what it can do. It is not different in what it requires before it can do it.
Your company's workflows were designed for people. People can handle ambiguity. People can fill in gaps. People can figure out what someone meant even when the data is incomplete or the process is inconsistent. AI cannot. It needs clean inputs, documented steps, and consistent execution to function the way the demo promised.
When those things are not in place, the AI tool does not fail because the technology is bad. It fails because it inherited a foundation it cannot work with.
The Real Reason for the Gap
The Redwood report points to fragmented workflows, unreliable data flows, and manual exception handling as the culprits. That is accurate. But those are symptoms. The root cause is simpler: most manufacturers sorted out the technology decision before they sorted out their own house.
In Lean terms, Sort is the first step for a reason. Before you can set anything in order, before you can standardize, before you can sustain, you have to look at what you actually have. What is working. What is not. What is documented and what only exists in someone's head. Most organizations skip that step because it is unglamorous and it does not show up in a vendor demo.
I have watched this play out across different industries and different technologies. A program looks promising on paper. Leadership is committed. The budget is approved. Then implementation begins and the team discovers that the process they planned to automate has three variations depending on who is running the shift, the data lives in two systems that do not talk to each other, and nobody has written down the actual steps in years. The tool cannot run on that. No tool can.
What to Do Before You Go Any Further
You do not need to solve everything before you start with AI. You need to solve one thing. Here is how to find it.
Pick the area where you most want to use AI. Not the flashiest use case. The one with the most pain right now. The workflow that costs you the most time, causes the most errors, or creates the most friction for your team.
Then ask one honest question: would you trust AI to run that process today?
Not eventually. Not after some adjustments. Right now, with the process as it exists, the data as it sits, and the steps as they are currently documented.
If the answer is yes, you have found your starting point. If the answer is no, that hesitation is telling you something. The work you need to do is not selecting an AI tool. It is getting that process to the point where you would trust it enough to hand it off.
That is the real readiness question. Not which AI platform. Not what the budget looks like. Whether the process is clean enough, documented enough, and consistent enough that a system without human judgment could run it.
The Thing Most Assessments Miss
The Redwood report frames readiness as a technical problem: data integration, workflow orchestration, automation maturity. Those things matter. But there is a layer underneath them that is harder to see in a survey.
Company culture. The people who run the process every day have to believe the change is worth making. If they do not, the AI tool becomes one more system they work around. The technology sits on top of the same broken foundation, and the 80% who are not ready stay right where they are.
Getting your process ready for AI is not just a documentation exercise. It is a conversation with your team about what is actually happening versus what is supposed to happen. That conversation takes longer than a tool evaluation. It also determines whether any of this works.
The 80% who are not ready are not behind because they chose the wrong AI vendor. They are behind because they started with the tool instead of the foundation. One process, cleaned up and trusted enough to hand off, is worth more than a roadmap full of AI use cases built on shaky ground.