Vague goals
No clear hypothesis or success criteria to genuinely test something. So any outcome can be justified.
AI & Governance
Proof of concept for AI. From pilot to proven step.
Most AI pilots don’t end with a decision, but with silence. No goal, no measurement, no point when someone says: we will or won’t do this. We set up pilot projects with a clear hypothesis, a test design and criteria agreed in advance—so there’s a decision at the end.
Clear goalsFrom idea to a testable hypothesis.
Fairly assessedData, impact and risks assessed objectively.
Decide with confidenceStop, adjust or scale up.
Pilot scorecard
exampleOverall score72out of 100 · ready for a decision
Time savings30%target was 20%
Adoption58%threshold was 60%
| Component | Outcome | Score |
|---|---|---|
| Goal and hypothesis | clear and verifiable | 9/10 |
| Data readiness | complete and representative | 8/10 |
| Feasibility | stable and scalable | 7/10 |
| Risk and compliance | manageable | 7/10 |
| Impact and value | 30% time savings achieved | 9/10 |
| Adoption | 58%, just below the threshold | 6/10 |
No pilot without
goal, measurement
and decision.
goal, measurement
and decision.
Where pilots go wrong
Scope too broad
The pilot tries to do everything at once, making it impossible to compare or steer.
Unclear data
Data is incomplete, unrepresentative or of insufficient quality.
No measurement
Nothing is measured, or only output rather than impact.
No decision
No criteria agreed in advance, so experiments drag on forever.
What a good pilot setup includes
Eight questions. If you can answer them beforehand, the pilot will lead to a decision.
Goal
What do we want to learn or validate?
Hypothesis
Which assumption are we testing, and how will we know if it’s correct?
Scope
What’s included, and what’s explicitly out of scope?
Data
What data is needed, and is it available and representative?
Risk and ethics
What risks do we see, and how do we address them?
Success criteria
When is the pilot successful? Agreed in advance, not afterward.
Owners
Who owns it, and which team is involved?
Final decision
What do we do for each outcome? We agree on that in advance, too.
Our approach
1Kickoff and intakeWe understand the challenge and identify opportunities and constraints.
2Hypothesis and designWe define the hypothesis, success criteria and test setup with a minimal scope.
3Data and setupWe validate the data and set up the environment.
4Execute and measureWe run the pilot, measure impact and gather feedback.
5EvaluateWe score the results against the criteria agreed in advance.
6Decision and next stepsStop, adjust or scale up—with lessons learned documented.
Deciding after the pilot
Three outcomes, all three good. Only "we'll take another look at it" is not an outcome.
Adjust
Changes needed to the scope, data or approach.
- Adjust data scope
- Improve the design
- New test round
Scale
Goals achieved and the business case is positive.
- Plan scaling
- Put governance in place
- Measure and adjust
Stopping is the cheapest outcome. A pilot that is honestly called off after three months costs less than one that drags on half-heartedly for two years because nobody dares to make the decision. That’s why we agree on the criteria up front, before anyone is emotionally invested in the outcome.
From isolated experiments to a portfolio
One pilot is an experiment. Ten pilots without overview is a problem. This is how you maintain control once there are more.
Portfolio
One overview of all ideas, experiments and initiatives, with their phase and owner.
Prioritization
Weigh options transparently based on value versus effort, not on who shouts the loudest.
Ownership
Clear who decides and who is responsible—for each idea, not just in general.
Learn and repeat
A structure for learning, sharing and doing better next time.
What you sort by
- Quick wins — high value, easy to achieve. Start here.
- Big leap — high value, difficult. Plan carefully; don’t do everything at once.
- In the queue — low value, easy. Only if there’s room left.
- Don't start — low value, difficult. Say it out loud so it doesn’t get stuck in your head.
What you document for each idea
- Who contributed it, and who owns it
- What value it should deliver, in hours or euros
- What stage it’s at: idea, validation, experiment, or scaling
- When it will be reviewed again if nothing happens
Who this works for
Innovation leads
Validate ideas faster and build a strong pipeline of promising applications.
IT and data
Working within clear guidelines, data requirements and technical feasibility checks.
Process owners
Assess impact, adoption and actual value before scaling up.
Executive team and management team
Make decisions based on facts, risks and actual value rather than enthusiasm.
Take the first step toward a pilot that delivers results.
Even if the outcome is "don't do it". Start with the free online Fresh perspective to see where the opportunities lie.
Free online scanCriteria agreed in advanceA decision at the end
Ask this site
Ask your question about AI, the AI Act, integrations or digitalization. The answer comes from the pages of this site, with references to where it can be found.
The way it works now
- The appointment is in one colleague's mailbox
- Data is manually retyped from one system to another
- The person who knows is on vacation
- No one can see afterwards why this decision was made
As it is
- The agreement is stored with the client, findable for everyone
- Once entered, information flows through to the rest
- Work continues, even if someone is absent
- Every decision has a source, an owner, and a date
drag with your finger
What if Peter leaves?
Two out of five rules on this page disappear. That's how it feels when knowledge isn't part of the company.
What is your next step?
Free self-assessmentTwelve click questions, three minutes, immediate score and your biggest risk.
Fresh LookBusiness case, benchmark and priority. Settled once we build after this.
What it costsFixed price per package, no price per user, no hourly billing.
From the fieldWhat changed at companies that started with this.
Also good to know: frequently asked questions · security and data · terms and conditions · partners · research and figures · templates · what's changing · where you start
Where does the process break down in practice?
An AI pilot only adds value if the source data are reliable, someone is accountable and the output is used in the actual workflow.
Direct answer
Start with a concrete business problem and a measurable outcome, not with an AI tool. Arrange data, ownership, risk and human control before you scale up.
What we do
We connect analysis, business intelligence, data and execution so you don't get stuck in advice or loose tooling.
How it works
- We clarify the current situation, data and the real bottleneck.
- We determine the smallest working solution, owner and measurable outcome.
- We implement or prioritize the next step and verify the result.
- We secure what works and stop or fix what doesn't work.
What you get
A concrete analysis, priorities, decision criteria, an executable next step and visibility into what needs to be built, integrated or secured next.
For whom and whom not
Right for Dutch SMB organizations that want to solve a concrete process, knowledge, data or AI problem. Not for organizations looking only for a generic tool demo or unproven AI promise.
Costs and choice
Price and scope depend on the chosen next step. View the current fixed packages and pricing logic at pricing for digitization.
Risks and concerns
We work step by step, make assumptions and ownership explicit, use only the minimum necessary data, and maintain human control over decisions and execution.
Frequently asked questions before you decide
Does everything have to be done at once?
No. The preference is one clearly defined problem with a measurable outcome, then scale up.
What does it cost?
The current pricing and package logic is on the pricing page; custom work is only chosen if the scope really requires it.
How do you prevent a trajectory without results?
By defining the problem, owner, success metric and stop criteria in advance, and doing readback and verification after each step.
How we assess your situation
We map your current process, systems, exceptions and costs. Only then do we advise whether an off-the-shelf solution, a custom integration or another approach makes sense.
From comparison to business value
Continue with the implementation, governance or commercial step that fits this topic.