What the data actually says.
In the 2024 ICT Usage in Enterprises survey, companies that had considered AI but decided not to adopt it were asked why. Lack of experience was by far the most common reason, named by 74.6 percent of these companies, ahead of privacy concerns at 44 percent and worries about legal consequences at 37.4 percent, according to the AI Monitor from Statistics Netherlands (CBS). CBS notes the finding held for companies of every size class. For small and medium businesses, where one blocked team can stall the whole company, that number describes the single biggest brake on adoption.
Why lack of experience is named as a barrier.
When Dutch SMBs report lack of experience as the top adoption barrier, they are describing a real phenomenon. Employees do not know what to do with the tools, middle managers cannot model new behavior, and senior leaders cannot distinguish between vendor promises and operational reality. The instinct is to treat this as a training deficit. The pattern we see is different. The experience gap is not about missing technical knowledge. It is about the absence of a safe, structured way to build judgment.
Most AI training in SMBs is delivered as a workshop or a vendor demo. Someone shows the tool, walks through features, and expects employees to translate that into daily work. What actually happens is paralysis. Employees do not know which tasks are safe to hand over, which outputs need checking, or what to do when the tool produces something that looks right but feels wrong. They have been given a capability without a framework for using it. Lack of experience becomes a polite way to describe organizational abandonment.
The hidden cost of treating experience as a credential.
When organizations treat AI experience as a credential, they create two problems. First, they assume that hiring someone with AI on their resume solves the adoption problem. It does not. One person with tool experience cannot transfer judgment across a workforce that does not yet have permission to experiment. Second, they defer adoption until enough employees have been trained. This guarantees that the organization waits while competitors who designed around inexperience move forward.
The real cost is cultural. Framing experience as a prerequisite signals that only certain people are allowed to engage with AI. Everyone else is expected to wait, comply, or risk getting it wrong in public. In Dutch SMBs where hierarchy is flatter and collaboration is assumed, this creates quiet resistance. Employees do not push back loudly. They simply do not adopt. The tools sit unused, and leaders interpret the stall as lack of interest rather than lack of design.
What people need instead of experience.
Employees do not need years of AI experience to adopt effectively. They need four things: clear boundaries, visible judgment criteria, safe space to experiment, and a feedback loop that treats mistakes as data. Most SMBs provide none of these. Instead, they provide access, a short training session, and an expectation that people will figure it out.
Clear boundaries mean defining which tasks are in scope for AI assistance and which are not. This is not a technical decision. It is a business decision about risk, quality standards, and customer expectations. Visible judgment criteria mean showing employees how to evaluate AI output. Not whether it sounds good, but whether it meets the standard the business actually holds. Safe space to experiment means creating low-stakes environments where employees can test AI on real work without fear of public failure or wasted time. A feedback loop means regular, structured reflection on what worked, what did not, and why. These four elements turn inexperience into capability faster than any training program.
The Millennial and Gen Z adoption dynamic.
In Dutch SMBs, a significant portion of the workforce is Millennial or Gen Z. These employees do not interpret lack of experience the same way older cohorts do. They assume tools should be intuitive, that learning should happen through use, and that organizations should design for fast iteration. When an SMB treats AI adoption as a formal launch with prerequisite training, it signals distrust. Younger employees hear: you are not ready, you might break something, wait for permission.
This creates a secondary barrier. Employees who would adopt quickly if given room to experiment instead disengage. They see the distance between the organization's stated commitment to AI and its actual tolerance for experimentation. The result is not loud resistance. It is quiet non-adoption. Leaders interpret this as lack of interest or capability, when it is actually a response to risk-averse design. The fix is not more training. It is creating space for judgment to develop through supervised use, fast feedback, and visible modeling by managers.
How to design around inexperience.
Start by naming inexperience as a design constraint, not a skills deficit. The organization has licensed tools that most employees have never used in a work context. That is a fact, not a failure. The question is how to design adoption so that judgment develops faster than risk.
First, identify three to five high-frequency, low-risk tasks where AI can assist without catastrophic failure. These are the tasks employees repeat daily, where mistakes are visible and fixable, and where output quality is easy to evaluate. Examples in Dutch SMBs: drafting customer emails, summarizing meeting notes, generating first-pass reports, translating internal documents, formatting data for analysis. Choose tasks where employees already have domain expertise. AI becomes the assistant, not the decider.
Second, create task-specific guidance that shows employees what good output looks like and what to check before using it. Not a feature manual. A decision guide. If you use AI to draft a customer email, here is what you check: tone, accuracy of product details, compliance with company policy, clarity for a non-expert reader. If you use AI to summarize a meeting, here is what you verify: no missing action items, no misattributed decisions, no invented next steps. This is not training. It is scaffolding that lets employees build judgment through repetition.
Third, create a weekly feedback ritual where employees share what they tried, what worked, and what they learned. Not a formal review. A fifteen-minute huddle where three people talk about one thing they did with AI that week. This makes experimentation visible, normalizes mistakes, and transfers judgment faster than any training session. It also signals that the organization values learning over perfection.
Fourth, give middle managers a clear role. They model experimentation, share their own mistakes, and coach employees through uncertain decisions. If managers are not using AI themselves, employees will not adopt. If managers treat every AI misstep as a quality failure, employees will stop trying. The manager's job is to create the conditions where inexperience becomes experience without penalty.
When to bring in external support.
Most Dutch SMBs try to solve the experience barrier internally. They assign ownership to HR, IT, or a single enthusiastic manager. This works if the organization already has a culture of fast feedback and low-penalty experimentation. Most do not. The result is that adoption stalls, the internal owner becomes the scapegoat, and the tools remain underused.
External support makes sense when the organization cannot name the real barriers, when internal attempts have failed without clear learning, or when leadership needs a credible outside read to justify a different approach. The value is not in bringing AI expertise. It is in designing the scaffolding that turns inexperience into capability. That means ground truth first: an honest read on why adoption has stalled, who is blocked and why, and what the real risk tolerance is. Then a plan built around the workforce's actual starting point, not the one leadership wishes they had.
A tool like the AI Profit Readiness Assessment gives a clear read in about two minutes, showing where adoption is blocked and what is driving resistance. For SMBs that need a full plan, the AI Profit Sprint delivers a structured approach to designing change around real capability, not assumed experience.
Questions people ask.
Is lack of AI experience really the main barrier in Dutch SMBs?
It is the most commonly named barrier, but it is usually a symptom. The real barrier is that organizations provide access without boundaries, judgment criteria, or safe space to experiment. Employees do not lack capability. They lack design.
Should we hire someone with AI experience to solve this?
Hiring one person with AI experience does not solve an organizational adoption problem. That person cannot transfer judgment to a workforce that does not yet have permission to experiment. The fix is designing scaffolding that lets inexperience become capability through use.
What is the fastest way to build AI experience in an SMB?
Identify three to five high-frequency, low-risk tasks, create task-specific guidance on what to check, and build a weekly feedback ritual where employees share what they tried. Judgment develops through repetition with scaffolding, not through training.
Why do younger employees resist AI adoption in some Dutch SMBs?
They do not resist the technology. They resist risk-averse launches that require formal training and permission before experimentation. When organizations design for control instead of iteration, younger employees disengage quietly.
When should a Dutch SMB bring in external help for AI adoption?
When internal attempts have stalled without clear learning, when leadership cannot name the real barriers, or when the organization needs an outside read to justify a different approach. External support works when it starts with ground truth, not prescription.