AI for SMEs · for owner-directors
AI for SMEs: start with the work, not the hype
Your business is growing, your people are at full capacity and everyone has something to say about AI. Here you'll read what your team can do itself tomorrow, when it's better to have something built, what that costs and how quickly it pays for itself. Honest and with numbers, including about when it doesn't pay off yet.
- 160+ organizations helped
- At least 2x back in year one
- All prices on the website
12 months with an AI team
More work, same team
+40%
Worked with us
01 / The short answer
Where should an SME start with AI?
Don't start with a tool. Start with the work that comes back every week. Look for the tasks where people retype, search or write the same email again and again. Teach your team to use AI for those first. Only have something built when a process comes back often and runs through several systems. And always do the maths first.
That sounds simpler than it feels. As an owner-director you're offered a new AI tool every week, you read everywhere that you're falling behind and you watch your best people work late every day. The temptation to buy something is strong. Don't. First spend a week looking at what your people actually do all day.
The good news: in a growing business the gains are often close at hand. Lines are short, you make the decisions yourself, and a process that costs a colleague half a day every week can be set up differently in no time. You don't need an IT department for it. You do need someone to drive it. The figures back this up: among companies with 50 to 250 persons employed, AI use rose from 20 percent in 2023 to 45 percent in 2025, the fastest growth of any size class (Statistics Netherlands, CBS). More figures in AI in the Netherlands in numbers.
“Don't let it drive you crazy. Start small.”
02 / Sound familiar?
What we hear from owner-directors.
They rarely have questions about AI itself. They do have questions about time, people and money. We hear these six again and again, in intro calls and after a talk.
Too much choice
“Every week someone calls with an AI tool. I no longer know what to believe.”
Growing without people
“There's plenty of work, but I can't find the people. And the people I have are already at full capacity.”
Everything depends on one person
“When our planner wants to go on holiday, half the business grinds to a halt.”
Everyone does their own thing
“Half the team uses ChatGPT. Which accounts, and with which customer data? No idea.”
No time to get started
“I'd like to get going, but I'm in the day-to-day operation all day myself.”
Afraid of a costly mistake
“I once had a software project that ended up costing three times as much. I don't want that again.”
03 / Tomorrow, on your own
What your team can already do with AI tomorrow.
Your team handles build layers 1 and 2 itself. You don't need a developer for that, just a few hours of practice and clear agreements about what is and isn't allowed. A workshop at your premises brings the whole team to the same level in one half-day session. You set out the agreements in a short AI policy.
Build layer 1
Prompting well
A customer email in your tone, a summary of a long contract, a job advert. With a good prompt the first draft is there in a minute. You finish it off.
Build layer 1
Providing knowledge
Give AI your price list, your house style or your terms and conditions, and suddenly the answers are usable. The difference between 'a text' and 'our text'.
Build layer 2
Your own assistants
An assistant that knows your products and writes quotes in your style. Set it up properly once, and the whole team benefits every day.
Build layer 2
Skills for the team
Capture your standard ways of working, so a new colleague gets the same good answer as someone who has been there ten years. No more 'ask Pete', but 'ask the assistant Pete filled'.
04 / Where do you stand?
Take the free self-assessment: where does your business stand with AI?
Seven questions, one per dimension, in two minutes. You immediately see where your lowest score is: that's usually where the biggest gain lies. We don't store your answers.
Free AI self-scan
1 = nothing in place · 5 = fully sorted
Strategy and Vision
Is there a shared AI vision and a mandate at board level?
Data readiness
Is data central, structured and accessible?
Technology and Infrastructure
How easily do you connect new systems to existing ones?
People and Skills
Does the team know AI tools, and is there room to learn?
Processes and Adoption
Are processes documented and actively optimised?
Governance and Risk
Is there policy for AI use, privacy and security?
Culture and Change
Is there room to experiment, to fail and to learn?
3.0
your score
4.0
target · after 12 months
Start here
Strategy and Vision
your lowest score, so the biggest gain
You have loose pieces in place
Enough to start with, too little to build on. The gain is in tying together what is already there.
Free, no strings
Send us your profile and get an analysis back
- ✓Which of the seven dimensions we would tackle first, and why that one
- ✓Two examples of what that delivered at a comparable organisation
- ✓Whether a single workshop makes more sense here than a partnership, or the other way round
Your mail app opens with your seven scores in it, addressed to noud@nina-ai.nl. You send it yourself, so we store nothing until you hit send. Reply within 24 hours, from a human.
Rather look around first?
Indicative and not stored: we do not keep your answers. The real baseline we do on site in the AI Quick Scan, together with your team.
05 / Do it yourselves or have it built?
When should you have AI built? The four build layers side by side.
Not every opportunity is a build project. We place every opportunity in one of four build layers. Your team handles layers 1 and 2 itself. We build layers 3 and 4, within the AI Partnership. Rule of thumb: only have something built when a task comes back often, runs through several systems and you can express it in hours.
| 1. Prompts and knowledge | 2. Assistants and skills | 3. Workflows | 4. Custom work | |
|---|---|---|---|---|
| Who does it | Your team | Your team | NinA builds | NinA builds |
| Example | Answering a customer email in your tone | An assistant that writes quotes with your prices | Orders from email go into the ERP automatically, a person approves | An agent that reads requests, assesses them and puts them into your systems |
| When to choose it | One-off tasks, per person | Recurring work where knowledge and tone matter | High volume, fixed steps, several systems | Lots of variation, your own logic or a platform of your own |
| With what | ChatGPT, Claude or Copilot | Custom GPTs, Claude Projects or Copilot agents | n8n, connected to Exact, AFAS or HubSpot, for example | Custom work, connected to your own systems |
| How to start | Workshop or practise on your own | Workshop, then build it yourselves | AI Quick Scan | AI Quick Scan |
Swipe sideways to see all columns →
More on workflows in n8n and AI automation, more on agents in AI Agents.
06 / From idea to result
From idea to result in four steps.
This is how growing businesses tackle it with us. You can stop after any step, and every step delivers something on its own.
- 01
Teach your team to work with AI
Week 1
A workshop or talk at your premises, or practise on your own. Everyone learns to prompt and to create their own assistants. You capture the first gains right here, and your team will naturally see where work gets stuck.
- 02
Find the work that keeps coming back
Weeks 2 to 4
Ask everyone to note down for a week what they retype, look up or write again. Give each task three numbers: how long, with how many people, how often. Then you see where the hours are.
- 03
Do the maths
Within a week
A business case in euros per opportunity, and an order. You can do that yourself, or with the AI Quick Scan: one day on site, and within a week a roadmap for twelve months.
- 04
Build one process, and measure
Then, month by month
Start with one process, not five. A person approves until the numbers add up, and every month you see what it delivers. Only when the first process is running do you take on the next.
07 / Do the maths yourself
What would AI deliver for your business?
Enter how many people do office work, how many hours a week they spend on repetitive work and what an hour costs. The calculator shows what that's worth and whether a partnership pays off. If it comes in below the guarantee threshold, it tells you so, and recommends a workshop.
Calculator
Retyping, searching, answering emails, reporting
Share AI takes over
We always advise on the expected scenario, never on the best case.
Value per year
€ 264,960
Hours back per year
5,888
3.5FTE of extra capacity, without hiring anyone.
This fits package M.
The savings clear the 2x threshold of € 192,000. In the Quick Scan we check it against your own processes.
An estimate based on your input and the assumptions above. Not a quote and not a promise: the real numbers come out of the Quick Scan.
08 / Payback
How quickly does AI pay for itself?
That depends on one thing: how many hours currently go into the work. That's why we always do the maths first. A worked example with one task, preparing a quote:
- It takes 40 minutes each time
- 2 people do it
- 80 times a month
That's 1,280 hours a year. At € 45 per hour, that one task costs € 57,600 a year. If AI takes over 40 percent of it, the expected scenario in the calculator above, that's € 23,040 a year. For one task, with a person who keeps checking.
Set the investment against that. An AI Partnership S costs € 60,000 in the first year, including the Quick Scan. Our guarantee: the partnership as a whole delivers at least twice the investment in the first year, or we don't start. One task won't get you there. Five or six tasks like this will. That's why in the Quick Scan we look at a whole department, not at one isolated job.
And in practice? For one of our clients the build cost € 28,000 and delivered € 77,000: paid back in about 4.5 months. Want to work it out for your own business? Read making an AI business case.
01€ 57,600 a year
What one task of 40 minutes, 2 people and 80 times a month costs at € 45 per hour.
02About 4.5 months
Payback time at one client: € 77,000 saved against € 28,000 in build costs.
03From 4 hours to 1
For a sales proposal, at another client.
0412 hours a week
Saved on invoice processing: a day and a half of work, every week.
05At least 2x
Our minimum for the partnership as a whole in the first year.
09 / In practice
What AI delivers at growing businesses.
Three businesses, three processes that came back every day: product copy, requests and administration.
10 / What it costs
What does AI cost for an SME?
All our prices are simply on the website, excluding VAT. One rate for advice and building: € 150 per hour. To be fair: the partnership is a serious investment. If that's too big a step right now, start with a workshop or a talk. Your team captures the first gains itself, and afterwards you know better what you want. What we build, we keep running with a maintenance contract from € 245 per month. Everything side by side: what does AI cost?. And whether there's a subsidy for it, we answer honestly in AI subsidies in the Netherlands.
Inspire
AI talk: € 2,450
A talk about what AI can do today, with examples from practice. A good way to bring the whole team, or your customers, along in one go.
Do it yourselves
Workshop: € 4,500 per half day
At your premises. Your team gets to work on its own tasks: build layers 1 and 2. Invoiced in advance.
Do the maths
AI Quick Scan: € 6,000
One-off. One day on site, and within a week a business case per opportunity and a roadmap. Also available on its own.
Have it built
AI Partnership: from € 4,500 per month
S € 4,500, M € 7,500 or L € 12,000 per month, minimum 12 months. Year 1 including Quick Scan: € 60,000, € 96,000 or € 150,000.
The guarantee
Better an honest no than a project that delivers nothing.
The partnership as a whole delivers at least twice the investment in the first year. If the business case doesn't show that, we advise a smaller package, a workshop or doing nothing. And we say so plainly.
11 / Security and GDPR
Is AI safe for your business? The GDPR in outline.
Yes, if you arrange a few things. The biggest risk rarely lies in the technology. It lies with the colleague who quickly pastes a customer list into a free AI tool, because there's no business alternative.
For your team: work with business licences, don't put personal data into tools where that hasn't been agreed, and have a person check whatever goes out of the door. Put that in a short AI policy. What can and can't go into a prompt is covered in using ChatGPT safely in your business.
For what we build: the data stays in the EU, on your own infrastructure or on servers with a European provider. We work in line with ISO 27001 and NEN 7510; certification is in preparation. There's always a human in the loop, and code, data and documentation are yours.
And the AI Act? Among other things, it asks you to work on the AI literacy of your people. Read the EU AI Act for SMEs and AI literacy. On the legal details, a lawyer has the final say.
01Business licences
No personal accounts for work with customer or company data.
02Agreements on paper
A short AI policy: which tools, which data, who checks.
03A person checks
Anything that goes to a customer has been seen by a person.
04Data in the EU
What NinA builds runs on your own infrastructure or with a European provider.
12 / Questions
Frequently asked questions about AI for SMEs.
Is AI worthwhile for an SME?
Yes, especially for an SME. In a growing SME the lines are short and the director makes the decisions, so an improvement is in place quickly. The gains are in work that comes back every week: quotes, orders, customer emails, planning and administration. How much it delivers depends on how many hours currently go into that work. Do the maths before you invest.
Where do you start with AI in your business?
With the work, not with a tool. First teach your team to work with AI themselves, with good prompts and their own assistants. Then find the tasks that come back often and give them three numbers: how long, with how many people, how often. Work out per opportunity what it delivers. Only have something built when a process comes back often and runs through several systems.
What does AI cost for an SME?
Working with AI yourselves costs licences and time. At NinA a talk costs € 2,450, a workshop € 4,500 per half day and the AI Quick Scan € 6,000 one-off. The AI Partnership, in which we build, costs € 4,500, € 7,500 or € 12,000 per month, for a minimum of twelve months. Maintenance costs € 245, € 445 or € 895 per month. All amounts exclude VAT. Everything side by side is on pricing.
How quickly does AI pay for itself?
That depends on the hours currently going into the work. One task of 40 minutes, with 2 people, 80 times a month, costs € 57,600 a year at € 45 per hour. For one of our clients a solution paid for itself in about 4.5 months: € 77,000 saved against € 28,000 in build costs. Our guarantee: the partnership as a whole delivers at least twice the investment in the first year, or we don't start.
What can my team do with AI itself, without a developer?
A lot. Prompting well, providing knowledge such as your price list or house style, and creating your own assistants that write quotes, customer emails or summaries in your style. Those are build layers 1 and 2. All you need is ChatGPT, Claude or Copilot and a few hours of practice. A workshop at your premises brings the whole team to the same level in one half-day session.
When should you have AI built?
When a task comes back often, runs through several systems and you can express it in hours. Think of orders that need to go from email into the ERP, requests that are assessed and filed, or product copy in several languages. Those are build layer 3 (workflows, for example in n8n) and build layer 4 (custom work and agents). We build those within the AI Partnership, after the AI Quick Scan.
Can I use ChatGPT with customer data?
Not just like that. Not with a free version or a personal account. With a business subscription, a data processing agreement and clear agreements it can be done, and even then with as little personal data as possible. Set out in an AI policy what is and isn't allowed. More in using ChatGPT safely in your business.
What if a partnership is too big for us?
Then we'll tell you, and you start smaller. The partnership is a serious investment: at least € 60,000 in the first year, including the Quick Scan. If that's too much right now, start with a workshop at € 4,500 per half day or a talk at € 2,450. Your team captures the first gains itself with prompts and assistants, and afterwards you know much better where building pays off. The AI Quick Scan is also available on its own.
Further reading
- What does AI cost?All prices side by side, and how to work out the payback time.
- Creating an AI strategyNot an 80-page report, but a roadmap for twelve months.
- Talks and workshopsYour team at the same level in one half-day session, at your premises.
- What is AI?Artificial intelligence in plain English, with figures on Dutch companies.
Start small, with an intro call
Fifteen minutes, no strings attached. Where does the work get stuck, and what's a logical first step: a workshop, the Quick Scan or straight into the partnership?
01
Intro call
Fifteen minutes, no strings. Where does work get stuck and what do you want to achieve?
02
AI Quick Scan
One day on site. Within a week, a business case and a roadmap.
03
AI Partnership
A dedicated team that builds, trains and shows what it delivers every month.
160+ organizations · Sessions rated 9.3 · Based in Amsterdam