Skip to Content Skip to main content

AI: hype or edge? Numbers, not opinions.

Seven percent of Dutch SMEs have AI in a core process. Fifty percent haven't started. This page shows what makes the difference. Numbers with sources, four frictions you'll recognize, a foundation that works. Two minutes to read.

7% of Dutch SMEs have AI in a core process 43% experiments without anything in production 3% globally run agentic AI extensively

Three concrete next steps at the end. For founders and leadership of Dutch organizations between EUR 10M and EUR 300M revenue.

The market waits. The technology runs. That's where your edge sits.

You've seen enough presentations. Enough pilots that lead nowhere. Enough consultants with three-year roadmaps. This page is different: facts with sources, in two minutes.

AI is everywhere in the news. In practice, it's still rarely deployed. That's not a problem. That's your edge, if you move now.

7%

of Dutch SMEs have AI embedded in a core process

Source: CBS, AI-monitor 2025

43%

experimenting with AI. 50% haven't started.

Source: KVK Entrepreneur Panel AI, 2025

3%

globally run agentic AI extensively or fully in production

Source: Deloitte State of AI in the Enterprise, Jan 2026, n=3,199

The productivity gap

Indicator Growth per year Source
AI capability ~100% Various benchmarks 2024-2025
Dutch labor productivity 1.1% CBS, 2024

That difference compounds. Whoever invests in AI in a core process now builds a lead that can't be caught up later.

"Technology is sprinting. The market is still waiting. That gap is the opportunity."

Why AI projects stall before they start

Eighty percent of AI initiatives don't reach production. That's not a BrainStax statistic. That's the Dutch market. Ask five fellow founders what their AI pilot delivered last year. Count how often you hear "we learned from it." It doesn't stall on the technology. It stalls on what sits underneath: the three pillars of people, data, and technology. Working together, or past each other?

"AI projects don't stall because AI doesn't work. They stall because organizations start with the technology, not with the friction that costs the most. That's the difference."

The triangle underneath: People, Data, Technology

An organization only becomes effective when people, data, and technology reinforce each other. Not as three separate projects running in parallel.

People first, technology follows. The other way around doesn't work.

Triangle with three points: People, Data, and Technology. The three sides reinforce each other; if one is missing, the triangle collapses.

From insight to action. That's where the gap sits.

Every organization climbs the same data ladder. From Excel (people calculate manually), to BI dashboards (data becomes visible), to algorithms (data predicts), to AI assistants (data acts). Most organizations reach dashboards. The step to algorithms and assistants that do work independently? That's where most drop off. Not because it can't be done. Because the data isn't ready for it.

Four frictions you hear in every boardroom

DATA & IT

How do you know which data you can trust?

Employees call, email, and manually search for information that's already in your systems. Nobody feels responsible for data verification.

SUPPLY CHAIN

Why do shortages only become visible when the customer calls?

Inventory decisions on gut feeling, outdated reports, or manual Excel calculations.

PLANNING

Why does planning depend on one person?

Spreadsheets, disparate systems, or knowledge in one head. Changes are translated manually.

SALES

Quotes take too long and contain errors

Manually compiled from historical data, price lists, and product specs across different systems.

You hear this back in every organization above 100 FTE, regardless of industry, regardless of ERP system, regardless of how much you've already invested in IT. These are four of the seven frictions we eliminate.

And then there are the concerns that keep founders awake

73%

Data privacy and security

50%

Legal, IP, and compliance

46%

Governance and oversight

Source: Deloitte State of AI in the Enterprise, January 2026, n=3,199. The EU AI Act takes effect August 2026 for high-risk applications. Do you know which of your systems hit that threshold?

AI ACT READINESS

How prepared is your organization?

6 questions. 3 minutes. Personal compliance analysis based on People, Data, and Technology. For the August 2026 deadline.

One foundation. All data sources. No system replacement.

We call it The Three-Pillar Foundation. A data layer under your existing systems, built on the triangle of people, data, and technology. Three pillars, one layer, no replatform. This is what the seven percent run in production. The other ninety-three percent don't.

Technically, the foundation is called BrainGrounds. It's not a replacement for what you have. It's a layer underneath that brings all data sources together. Your ERP stays your ERP. Your WMS stays your WMS. Your CRM stays your CRM. BrainGrounds connects to what's there and turns it into one central, reliable source.

And it runs on a private AI language model: your company data isn't used to train public AI models, and stays in a closed environment that only your people see.

Enterprise tech. SME+ budget.

Databricks isn't a matter of taste. It's the standard Shell, ABN AMRO, and Heineken run their data on, because one platform handles every data format: tables, documents, sensor data, images. We make that layer accessible for organizations between EUR 10M and EUR 300M revenue, on SME+ budget, at fixed price. Data stays at the source. No copies, no shadow administration. Every query logged, every data source traceable. Ready for EU AI Act compliance.

Three myths that hold founders back

"This becomes a replatform project."

It doesn't. BrainGrounds is a layer underneath. Your ERP, WMS, and CRM stay what they are. Your IT department stays in charge of the existing systems.

"Our data isn't clean enough for AI."

Nobody starts with clean data. We build the foundation on the data as it stands now, and make visible which sources are ready for the next step.

"The model shares our data with the world."

It runs on a private AI language model. Your company data isn't used to train public AI models. Every query logged, every source traceable.

Three solutions that run on the foundation

BRAINCHAT

Ask your own systems in plain language

Answers from ERP, CRM, and documents in one place. No phone calls, no manual searching.

"What's the steel plate inventory in Eindhoven?"

BRAINDOC

Procedures findable in seconds

Work instructions, quality documents, and protocols on your own documentation. Directly available on the shop floor.

"What's the protocol for material deviations?"

BRAINPLAN

Planning no longer dependent on one person

Operational data translated to planning decisions in real time. Changes automatically propagated.

"Optimize tomorrow's loading."

Fixed price, live in production

1

Data assessment

Map data sources, identify opportunities, input for strategy.

2

First solution live

BrainGrounds is running, BrainChat answers operational questions.

3

Expand

Eliminate the next friction, more sources, more users.

4

Continuously grow

Scales with new systems and ERP replacement.

Fixed price per sprint. Live in production. No throwaway investment: the foundation adapts when data sources change. Every step funds the next.

Four facts, one choice.

  • AI capability doubles each year. Dutch labor productivity grows 1.1 percent. That gap won't close by itself.
  • Eighty percent stall on data, not on technology.
  • A foundation on your existing systems works. Replacement doesn't. No throwaway investment.
  • Fixed price. Live in production. Every step funds the next.

"Every month without a foundation is a month your competitor widens that gap."

Anton de Nijs, founder BrainStax

Anton de Nijs

Founder BrainStax · 10+ years data & AI · Databricks partner

Questions after reading? A short email works fastest. I reply personally within 24 hours.

anton@brainstax.com
Take the AI Readiness scan →

10 questions. 5 minutes. Get by email which of the seven frictions yields the most for your organization. No phone call, no sales pitch.

P.S. What you start today runs in production a few months from now. What you don't start, still won't be running a few months from now. That's the entire calculation.

Read more

The method

People, data, technology. How we go from friction to production.

Read about the M-D-T method →

BrainGrounds

The foundation explained technically. For IT leads and data architects.

Explore the BrainGrounds platform →

All solutions

BrainChat, BrainDoc, BrainPlan and what comes after.

See the overview →

Results

What it delivers in production. Cases and numbers.

View the cases →