Mary Elizabeth's Hospital will treat some of the most critically ill children, young people and expecting families in Denmark. Børneriget Fonden is supporting it through private donations, with 96% of every krone going directly to the hospital. The foundation's team is lean, and what they do with their time has a direct line to what the hospital can do for its patients. That makes good prospecting more than an efficiency question.
Fundraising at scale depends less on effort than on sequence: knowing who to ask, and in what order. The CVR register is the obvious starting point, and a difficult one: the segment that matters most to a hospital foundation is almost entirely silent.
Of the 6,500+ companies registered as fonde og andre selvejende institutioner, the legal form covering most family foundations, almost none publish a purpose statement, board members or financial figures. The most relevant group to approach is also the hardest to assess. Dropping them was not an option, so the engine was built to work around the gap.
The scoring draws on two sources that are typically kept separate. For financial strength, the engine reads balance sheets, equity and asset figures from public filings, and fills in where records are absent using a chain of external data sources. For purpose fit, it uses GenAI to read the language each company uses to describe itself: its stated mission, its industry classification, the tone of how it talks about its work. A foundation whose language centres on children's health reads very differently from one that mentions health only in passing. Both signals combine into a single relevance score that determines the ranking.
The network piece works differently. Børneriget arrived with an existing set of contacts. The engine mapped those contacts against the boards, owners and founders of every candidate in the database, tracing connections up to four degrees of separation, and returned the shortest path to each prospect written in plain language. For every high-priority target, the fundraising team knows exactly who in their network can make the introduction.
The output is a Microsoft Excel workbook Børneriget controls directly. Scoring weights, financial thresholds and purpose criteria are all adjustable without touching code, so the team can redefine what counts as relevant as their priorities evolve. The shortlist does not go stale when the engagement ends.
Three weeks was enough because the question was kept to one. The engine ranks prospects; it does not try to be a CRM.