By Marc Broers, Co-founder and CTO FindNFund

In this article, I outlined why AI has evolved in just a few years from a demo to infrastructure. The logical follow-up question I hear everywhere is: what should we do with it?

My answer is almost always the same. The question is not whether AI will impact the work of funders and initiative makers, but how we choose to use it. And that “how” is not a technological question — it's a substantive choice that should align with your organization's mission, values, and processes.

First, the framework: three principles

Before we look at where AI can help, I want to share the three principles we apply at FindNFund. Not because they are mandatory, but because they make the difference between AI that adds value and AI that simply creates noise.

1. Human in the loop

AI does not need to replace human judgment — it can strengthen it. Think of how the medical field operates: an algorithm can pre-screen hundreds of scans, but the doctor remains responsible for the diagnosis. That's exactly how we approach application assessment: AI helps structure, compare, and summarize, but the evaluator retains judgment.

2. Serving a meaningful purpose

Using AI to make administration more efficient is valuable. But efficiency is not the end goal. The real goal is to create space. Less time spent on forms, more time for meaningful conversations. Less searching, more substantive work. Less administrative burden, higher quality in assessment. If an AI application does not serve that purpose, it does not belong in our sector.

3. Safe and responsible

This means complying with regulations such as the EU AI Act, storing data securely, and being transparent. Explainability, traceability, and data ownership are essential conditions for trust.

The process: five steps, five types of opportunities

The funding journey between funders and initiative makers can roughly be divided into five steps. Each step presents a different type of opportunity for AI. Let's walk through them.

Step 1 — Finding each other

It all starts here: the right initiative finds the right fund — and vice versa. This is often where the barrier is highest, especially for initiative makers who are not used to thinking in policy language and keywords.

From searching to finding: more accurate matching through smarter algorithms that understand context, not just keywords.
Lowering the threshold: chatbots that help sharpen a project plan and suggest suitable funding options.
Discovering initiatives: proactive suggestions for funds based on their policy, history, and current portfolio.

With Fundy, our AI assistant, we demonstrate what's possible here. It's not generic AI — it understands the philanthropic sector and can directly assess whether an initiative aligns with a fund's current criteria.

Step 2 — The funding application

For initiative makers, this is often the most demanding step. Forms, attachments, reporting requirements — and each fund does things slightly differently.

Draft applications: automatically pre-filling proposals based on a project plan. What used to take days can now take hours.
Reducing human error: scanning applications before submission to flag incomplete or unclear information.
Increasing accessibility: speech-to-text and video-to-text, so language or administrative skills are no longer barriers.

The beauty of this step: the time savings benefit the applicant. That's exactly where the burden is often greatest.

Step 3 — Assessing applications

This is perhaps the most sensitive domain — and therefore the most interesting, when applied carefully.

Identifying improvement insights in project plans based on a fund's historical experience.
Pre-advice: assessing whether an application aligns with vision, mission, and strategy, and comparing it with past decisions.
Better steering: recommendations based on portfolio analysis, enabling more intentional distribution of grants.

Again: assessment remains human work. AI provides the groundwork, context, and comparisons. Decisions are made by people with mandate.

Step 4 — Implementing the initiative

There are many opportunities here, depending on the type of project. I'd like to explore these in more detail in a future article.

Step 5 — Reporting

Saving time: AI supports the creation of reports for both initiative makers and funders.
Advanced impact analysis: based on sector-wide, opt-in data that is currently fragmented.
Pre-advice on outcomes: supporting the assessment of reports, again strengthening human judgment.

How should funds approach this?

What I see working in practice: start small, start together, and start with the steps where risks are lowest and gains are clearest. Search and administrative support are good entry points—you learn quickly without major consequences.

This is exactly why we created the AI Lab. Together with funds already working with FindNFund, we translate the possibilities of AI into concrete applications within their assessment and decision-making processes, with attention to policy, governance, and ethics. In a small, trusted group. Learning, testing, and reflecting together.