AI strategy · for leadership teams
Creating an AI strategy: short, concrete and with a number
A good AI strategy isn't an 80-page report. It's a list of processes, a business case in euros for each one, and a roadmap for the next twelve months. Here you'll read how to put one together yourself, which mistakes you can skip and how the AI Quick Scan puts it on the table for you within a week.
- 12-month roadmap
- AI Quick Scan: € 6,000 one-off
- 160+ organizations helped
7
dimensions in the baseline, from data to culture
3
numbers per opportunity: duration, people and frequency
12
month roadmap, worked out per quarter
2x
back in the first year, or we don't start
01 / The short answer
What is an AI strategy?
An AI strategy is a decision about where AI saves your business time and money, in which order you tackle it and what you need for it. That's all. It fits on a few pages and has five parts: where you stand now, a list of opportunities, a business case per opportunity, an order and a twelve-month roadmap.
What it isn't: a vision paper on 'the impact of AI on our sector', a list of tools you still want to try, or a report that disappears into a drawer after the presentation. Those documents aren't wrong. They just don't help you choose.
The test is simple. After reading it, can you say which process you'll tackle first on Monday, what that delivers and who's driving it? Then you have an AI strategy. If you can't, you have a wish list.
01Where you stand
A baseline on seven dimensions: strategy, data, technology, people, processes, governance and culture.
02What's possible
A list of opportunities, gathered from the people who do the work. Not from the vendor.
03What it delivers
A business case per opportunity in hours and euros, calculated on the expected scenario.
04Where you start
All opportunities on an impact-effort matrix. Quick wins first.
05What you do
A twelve-month roadmap, per quarter, with an owner and a number.
02 / Where it goes wrong
Six mistakes we see again and again.
In large and small organizations, in every sector. The good news: all six can be avoided. Recognise two or more? Then you know where your strategy needs attention first.
Mistake 1
Starting with the tool
'We need to do something with Copilot.' Then you're looking for a problem to fit a solution. Start with the work: where does someone retype things, where does a customer wait, where is the planner still at work at ten in the evening? Only then choose the tool. Often it turns out you already have it. More in stop buying AI.
Mistake 2
Pilots that never land
Five enthusiastic colleagues, a nice demo, and then nothing. Because nobody said up front when the pilot counts as a success, who manages it afterwards and which system it needs to connect to. Agree on that before you start, not after.
Mistake 3
No number
'That saves a lot of time' isn't a business case. 'Forty minutes, two people, eighty times a month' is. Without a number you can't compare opportunities, you won't get budget released and afterwards nobody knows whether it worked.
Mistake 4
Forgetting the data
AI is only as good as the information it can find. If your prices are in three Excel files and only Pete knows which one is right, AI will doubt just as much as a new colleague. Also read the first step in AI is data.
Mistake 5
Forgetting the people
A workflow nobody trusts gets worked around. Involve the people who do the work from day one: they know where it hurts, and they'll be the ones working with it. An afternoon of training often does more for adoption than a month of building.
Mistake 6
No owner
If AI belongs to everyone, it belongs to no one. Appoint one owner per opportunity in the business, not just in IT. That person guards the number, says when it's good enough and keeps the team on board.
03 / Do it yourself
How to put together an AI strategy yourself, in six steps.
You can use this plan without us. It's the same order we follow in the AI Quick Scan, except we do it in one day on site and do the maths in the days after. On your own it'll take a few weeks, alongside your normal work.
- 01
Ambition and frameworks
An afternoon with the leadership team
What do you want to have achieved a year from now? Growing without extra people, delivering faster, fewer mistakes? And what are the limits: which data may go where, what's the budget, who decides? Write it down on half a page.
- 02
Measure where you stand
An hour
Score your organization on seven dimensions, from 1 to 5. Be honest. The lowest score is usually where you start: without your data in order, every AI project gets stuck, however good the opportunity.
- 03
Gather opportunities
A week
Don't ask 'where can AI help?', because then you'll get a list of tools. Ask: 'what do you do every week that you find boring, what do you retype, what are you waiting for?' Spend a morning alongside someone at work. The best opportunities are in work so ordinary that nobody sees it any more.
- 04
Three numbers per opportunity
A few days
How long does the step take, with how many people, how often? Forty minutes, two people, eighty times a month is 1,280 hours a year: € 57,600 at € 45 per hour. Calculate with the expected scenario and at most 90 percent saving per step.
- 05
Build layer and matrix
An afternoon
Choose the build layer for each opportunity: a prompt, an assistant, a workflow or custom work. Then put the opportunities on a matrix of impact against effort. Top right is where you start. Anything that delivers little and costs a lot, you drop.
- 06
12-month roadmap
An afternoon
Per quarter: which processes, which owner, which budget and which number you want to see. Don't plan more than your team can handle. One process that's live is worth more than four half-finished pilots.
04 / Step 2 in pictures
This is what a baseline on seven dimensions looks like.
An example, not client data. The radar chart shows where you stand now and where you want to be in twelve months. If one dimension lags far behind, usually data or governance, that's where your roadmap starts. That way you don't build on a foundation that isn't there yet.
Profile · now vs. target
2.7
now
4.1
target after 12 months
Biggest jumpData readiness (2 to 4)
What we measure, on a scale of 1 to 5
Strategy and Vision+1 jump
How clear and shared is the AI vision at board level? Is the strategy written down and who holds the mandate?
now 3, target 4.Data readiness+2 jump
Is data centrally available, structured and accessible for automation and AI?
now 2, target 4.Technology and Infrastructure+1 jump
Which tools, cloud platforms and integrations are in place? How easily do you connect new systems?
now 3, target 4.People and Skills+2 jump
Does the team know AI tools, prompting and automation? Is there capacity and willingness to learn?
now 2, target 4.Processes and Adoption+1 jump
Are processes clearly described and actively optimised? How did earlier change programmes go?
now 3, target 4.Governance and Risk+2 jump
Is there policy for AI use, data ethics, privacy and security? How is risk weighed against speed?
now 2, target 4.Culture and Change+1 jump
How does the organisation deal with change? Is there room to experiment, fail and learn?
now 4, target 5.
05 / Not everything is building
Which opportunities can your team handle, and what do you have built?
A good AI strategy says not only what you tackle, but also how. For that we place every opportunity in one of four build layers. That saves money: an opportunity a good prompt can handle doesn't need a custom build. And an opportunity that genuinely needs custom work won't be solved with a prompt.
Build layers 1 and 2: your team does it
Build layers 3 and 4: having it built
Rule of thumb: only have something built when a task comes back often, runs through several systems and you can express it in hours.
01Layer 1: prompts and knowledge
Your team, themselves. One-off tasks, per person.
02Layer 2: assistants and skills
Your team, themselves. Recurring work where knowledge and tone matter.
03Layer 3: workflows
NinA builds. High volume, fixed steps, several systems.
04Layer 4: custom work
NinA builds. Lots of variation, your own logic or a platform of your own.
06 / The four phases
The four phases of AI adoption: where are you?
Organizations go through four phases: prompting, assistants, workflows and agents. Most are in phase 1: people use ChatGPT or Copilot, each on their own. The value is in phases 3 and 4, because that's where the work itself changes. A good AI strategy says which phase you're in now and how to reach the next. Click a phase to see what goes with it. More in the 4 phases of AI adoption.
Where are you? Click a phase.
- What it looks like
- People use ChatGPT or Copilot for an email or a summary. The handy ones save time, the rest watch.
- What you need
- Licences, a few ground rules and basic prompting skills.
- What it delivers
- Personal and hard to measure. The organization barely notices.
Our advice if you're here
Get your team to the same level in one afternoon. Then you'll see where work piles up.
07 / Example
What does a twelve-month AI roadmap look like?
An AI roadmap is a plan per quarter: which processes you tackle, in which order, with which owner and which result. Below is an example for a wholesaler with an inside sales team of twelve people. It's an example to show the idea, not a client.
Quarter 1: foundation and quick wins
Quarter 2: the first process is running
Quarter 3: the second process
Quarter 4: broaden and recalibrate
See the pattern? Every quarter one thing goes live, with a number attached. Not ten pilots at once.
01One process at a time
The next one only starts when the first is running. That way things get finished.
02A number every quarter
Hours back, fewer errors, faster delivery. Measured, not estimated.
03People and technology together
Training and building run side by side, so the team grows with it.
04Recalibrate every quarter
If the numbers don't add up, you adjust. A roadmap isn't a contract with yourself.
Our rule
An AI strategy without a number is a wish list.
That's why every opportunity gets three numbers and a business case from us before a single line of code is written. And we only start a partnership when the business case shows it pays back at least twice the investment in the first year.
08 / Three routes
Strategy project, AI Quick Scan or do it yourself: what suits you?
Three ways to arrive at an AI strategy. All three work, in different situations. We don't quote other agencies' prices: they vary too much per agency and per assignment.
| Strategy project with a large agency | AI Quick Scan from NinA | Do it yourself with the step-by-step plan | |
|---|---|---|---|
| Lead time | Varies per assignment, often several sessions | 1 day on site, results within a week | A few weeks alongside your work |
| What you get | Analysis and recommendations, often as an extensive report | Baseline, business case per opportunity, matrix, roadmap and package advice | Your own list of opportunities and a first roadmap |
| Business case in euros per opportunity | Varies per agency | ✓ | If you do the maths yourself |
| Conversations on the work floor | Varies per agency | ✓ | If you make time for it |
| Build layer per opportunity | Varies per agency | ✓ | With the explanation on this page |
| Who carries it out afterwards | Often another party | The same team, within the AI Partnership | You, or a party you choose |
| Price | On request from the agency | € 6,000 one-off | Your own time |
| Suits | A broad transformation question across the whole organization | Anyone who wants to know within a week where AI pays off and what it costs | A team with time and someone to drive it |
Swipe sideways to see all columns →
Amounts excluding VAT. The outcome of the Quick Scan is yours, even if you don't continue with us afterwards.
09 / What NinA does
The AI Quick Scan: your AI strategy on the table within a week.
The Quick Scan covers the six steps above in one day at your site: conversations with leadership, interviews on the work floor, processes viewed live and a brainstorm with the team. Within a week we present the outcome. Cost: € 6,000 one-off, excluding VAT, travel within the Netherlands included. More on the AI Quick Scan.
Where you stand
Baseline on seven dimensions
Strategy, data, technology, people, processes, governance and culture. Scored and substantiated together, in one radar chart.
What it delivers
Business case per opportunity
In hours and euros, with three scenarios. We advise on the expected scenario, never on the most favourable one.
Where you start
Impact-effort matrix
All opportunities on one page. Chosen on numbers, not on who shouts loudest in the leadership team.
What you do
Twelve-month roadmap
Which processes first, in which order, with the return and the build layer alongside.
What next
Package advice
Which AI Partnership package fits, or whether you're better off continuing on your own. 'Do nothing' is also possible advice.
Honest maths
A brake on nice stories
If data is missing or there are no examples of good work, we set the confidence to 40 percent at most.
10 / And then?
From AI strategy to execution.
A strategy is only worth something when it's carried out. That's often where it goes wrong: the advice comes from one party, the build from another, and half of it gets lost in the handover.
If you choose the AI Partnership after the Quick Scan, the same team carries out the roadmap. We train your people, build the workflows and agents in your own systems and show every month what it delivers. There are three packages, all for a minimum of twelve months:
- S: € 4,500 per month, 30 hours, one process at a time
- M: € 7,500 per month, 50 hours, two processes at a time (our advice)
- L: € 12,000 per month, 80 hours, four processes at a time
The investment in the first year, including the Quick Scan, is € 60,000, € 96,000 or € 150,000. That's serious money. Hence the guarantee: the partnership as a whole delivers at least twice the investment in the first year, or we don't start.
Would you rather continue with the roadmap yourselves? Fine. The outcome is yours. All prices side by side are on pricing.
01One team
The same people from strategy to build. No handover where half of it disappears.
02One rate
€ 150 per hour for advice and building. Invoiced quarterly in advance.
03Data in the EU
On your own infrastructure or with a European provider, with a human in the loop.
04No lock-in
Code, data and documentation are yours. You take them with you within 30 days if you stop.
Worked with us
11 / Questions
Frequently asked questions about an AI strategy.
What is an AI strategy?
An AI strategy is a decision about where AI saves your business time and money, in which order you tackle it and what you need for it. It has five parts: a baseline of where you stand, a list of opportunities, a business case per opportunity, an order based on impact and effort, and a twelve-month roadmap. A good AI strategy fits on a few pages.
How do you create an AI strategy?
In six steps. Set out the ambition and the frameworks. Measure where you stand on seven dimensions. Gather opportunities from the people who do the work. Give each opportunity three numbers: duration, headcount and frequency. Choose the build layer for each opportunity and put the opportunities on an impact-effort matrix. Finally, make a twelve-month roadmap. On your own that takes a few weeks alongside your work. With the AI Quick Scan the outcome is ready within a week.
What's the difference between an AI strategy and an AI roadmap?
The AI strategy is the why and the what: where AI creates value and which opportunities you choose. The AI roadmap is the when and the who: per quarter, which processes you tackle, with which owner, which budget and which result. A strategy without a roadmap stays an idea. A roadmap without a strategy is a list of loose projects.
How long does it take to create an AI strategy?
On your own, with the step-by-step plan on this page, it takes a few weeks alongside your normal work. With NinA's AI Quick Scan it costs you one day on site and the outcome is ready within a week: baseline, business case per opportunity, matrix, roadmap and package advice. From agreement to presentation, that's usually about three weeks.
What does an AI strategy cost?
Doing it yourself mostly costs time, yours and your team's. NinA's AI Quick Scan costs € 6,000 one-off, excluding VAT. That covers the preparation, the day on site, the calculations and the presentation. Travel within the Netherlands is included. The outcome is yours, even if you don't continue with us afterwards.
Does a small business need an AI strategy?
Yes, but not a thick one. Especially when budget and time are limited, you want to know which process delivers the most before you invest. One page with your five best opportunities, the numbers alongside and an order is often enough. For a step-by-step plan to get started, read implementing AI in 7 steps.
Who should be involved in an AI strategy?
Leadership, for the ambition, the frameworks and the budget. The people who do the work, because they know where the time goes. Someone from IT or your system administrator, for data, integrations and security. And one owner per opportunity in the business, who guards the number. An AI strategy written only by IT or only by leadership rarely lands.
What are the four phases of AI adoption?
Prompting: people use ChatGPT or Copilot, each on their own. Assistants: your own assistants that know the work and the house style. Workflows: automations that run by themselves, with a person who approves. Agents: digital colleagues that take steps in your systems by themselves. Most organizations are in phase 1. The value is in phases 3 and 4, because that's where the work itself changes.
Do AI policy and the AI Act belong in an AI strategy?
Yes. Your strategy says what you want to achieve with AI; your AI policy says what is and isn't allowed: which tools, which data and who checks. The AI Act also asks for AI literacy among your people: our AI training page explains what that means. On the legal details, a lawyer has the final say.
What if the strategy shows that AI delivers little?
Then that's an outcome too, and a valuable one. We only start an AI Partnership when the business case shows that the partnership as a whole pays back at least twice the investment in the first year. If it doesn't, we advise a smaller package, a different process, a workshop so your team captures the first gains themselves, or doing nothing.
Your AI strategy on the table within a week
Start with a fifteen-minute intro call. Then we schedule the Quick Scan, usually within two weeks.
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