Frequently Asked Questions

You upload a .docx or .pdf file of up to 10 MB — the same for every instrument. Pasting plain text is not possible. If your report is in another format, save it as .docx or .pdf first.

No. The analysis happens in real time and your document is never saved on our servers. Only your payment details are handled securely via Stripe.

Your document is used solely to produce the analysis and is not kept on our servers afterwards. Processing is done by Anthropic (Claude) as our processor, under a data processing agreement and EU standard contractual clauses — see the privacy policy. If your document falls under a confidentiality agreement that excludes third-party processing, weigh that up within your own organisation.

You pay once per analysis via Stripe (iDEAL or credit card): no subscription, no credit system. You first verify your email address with a code, then payment follows and the analysis starts.

All prices include 21% VAT (VAT number NL1798.00.103.B01). After payment you automatically receive a payment confirmation from Stripe. If you need an invoice made out to your organisation’s name for your records, email info@dutchmind.com — you’ll usually receive it within one business day.

Usually a few minutes. RisicoRadar, PlanScore and EvaluatieScore typically finish within two minutes; RapportageScore takes two to five minutes because of its more extensive assessment (21 criteria, multiple assessment rounds). You see the result on screen and also receive it by e-mail.

No. The analysis is an objective second opinion powered by Anthropic’s Claude AI. It sharpens your own professional judgement; it doesn’t replace it.

You can try, but you lose the fixed yardstick: fixed criteria with score anchors, two independent assessment rounds that dampen random outliers, and a rubric grounded in international standards — see the accountability page. A one-off chatbot prompt gives a different verdict every time, without traceable substantiation.

Each scale follows the method of the instrument: RapportageScore adds up 7 pillars with 3 criteria each to 0-100, PlanScore and EvaluatieScore add up seven building blocks or touchstones (each 0-3) to 0-21, and RisicoRadar uses risk language from 0 (critical) to 10 (low risk). The yardstick behind them is always the same: fixed criteria, score anchors and two independent assessment rounds.

Every analysis runs twice in full, as two separate assessments by the same AI model. If the two diverge too much, a third round runs automatically and the middle judgement counts. This dampens random outliers and safeguards the consistency of the score.

Yes — every analysis is a separate payment, and you can score a revised version as often as you like. Because the model reassesses from scratch each run, the score may differ slightly; the two assessment rounds per analysis keep that difference small.

First check your spam folder (sender: noreply@dutchmind.com). Your result also stays on screen as long as you keep the page open. Nothing received and the page closed? Email info@dutchmind.com — we will fix it or refund you.

On screen for as long as you keep the page open, and permanently in the results email with PDF attachment — that is your copy. We do not store your document or result; temporary processing data is deleted automatically within about 24 hours.

After a technical error you can retry immediately for free; your purchase remains valid. If it keeps failing, e-mail info@dutchmind.com: we will fix it or refund the full amount (see the terms, article 6.3).

RapportageScore evaluates your project report on 7 categories and 21 subcriteria, and gives you a score with concrete improvement points per category.

No, and that is deliberate. RapportageScore assesses your report, not your project. A low score does not prove your project is going badly, and a high score does not prove it is going well — even a perfectly sound project can report poorly, and a reassuring report can be missing key information. What the score does tell you: whether your steering committee can tell from this report how the project stands. Exactly what the score does and does not mean is described on the scientific foundation page.

Not exactly — the underlying language model reasons afresh each time. In an internal measurement (10 repeated analyses of the same test report), the overall score stayed within a range of about 3 points on the 100-point scale. Individual category scores can vary more than the overall score. An additional measurement with 6 test reports of pre-set, ascending quality placed those documents in the correct order every time, with no overlap between weak, medium and strong — for large, clear quality differences; subtler differences were not measured. Internal calibration measurement, August 2026; details available on request.

For project managers and leads who want to sharpen their report before it goes to the steering committee or client.

RisicoRadar analyses your project plan across 5 risk categories and 15 subcriteria, and delivers the top risks with concrete recommendations. The score runs from 0 (critical risk) to 10 (low risk).

No, and that is deliberate. RisicoRadar assesses the risk information in your document, not the project itself. A document with no risks listed does not prove the project carries little risk — and a long risk list does not prove the project is risky. What the score does tell you: whether your document contains the risk information your team and steering committee need to see and track the risks. Exactly what the score does and does not mean is described on the scientific foundation page.

Scope & Requirements, Planning & Resources, Budget & Finance, Quality & Compliance, and Stakeholders & Communication.

No. It’s a fast, independent second opinion, and it’s especially useful as preparation for a risk session with your team.

Whether your project plan is ready to launch: 7 categories, a score from 0 to 21, and a go/no-go recommendation.

Use PlanScore before the project starts, to check whether the plan itself is ready. Use RapportageScore during execution, to assess the progress report.

An explicit recommendation on whether the plan is ready to launch or not, with the points that need attention before you give the go-ahead.

No, and that is deliberate. PlanScore assesses the information in your plan, not the project’s future. A plan that is ready to launch is not a guarantee of success — it is a signal: research shows a link between the quality of the plan before the start and project performance, not a guarantee for the individual project. What the score does tell you: whether your plan contains the information needed for a well-founded start. Exactly what the score does and does not mean is described on the scientific foundation page.

EvaluatieScore checks the learning value of a project evaluation report: 7 categories, a score from 0 to 21, and a verdict.

EvaluatieScore is the youngest instrument in the Dutchmind suite. The methodological basis is the same as for the other instruments; the calibration set and the scientific foundation are still being built — the Body of Knowledge documents precisely what is established and what is not yet.

Before publishing your final evaluation: are you drawing out the lessons that actually help the organisation move forward?