
Ardabelle Capital Case Study
AI-Powered Deal Screening for Ardabelle
Ardabelle is a European private equity firm focused on value-chain resilience and industrial transition, and one of the first PE funds to build AI into its investment process from the ground up.
Challenge
Information on potential investments often arrives as a dense document from an investment bank, which can be 100 pages or more. Ardabelle takes a first pass through it to answer one vital question: does this asset fit the Fund's mandate, in particular Ardabelle's requirements around exclusion screens, sustainability practices, and European regulation? That read is slow, manual, and repetitive, and it delays the most valuable outcome in the process: an early no.
Development & Implementation
What We Built
A custom Claude plugin that reads a full deal document and screens it against Ardabelle's investment criteria before an analyst opens it. Each document comes back with one of four flags - Approved, Conditional pass, Needs review, or Fail - and the reasoning behind the call.
The same engine can run any codified screen, so the criteria set can extend as the fund's needs change.
How It Works
Runs in Ardabelle's own environment. Deal documents and their contents stay inside the firm's own account. Nothing sensitive leaves it.
A person always signs off. The tool never silently rejects a borderline deal. Anything ambiguous is flagged "needs review" and routed to an investment team member. It triages against fixed, published criteria; it does not make the judgment call, but rather augments the team's ability to make investment decisions.
Consistent and auditable. Every flag comes with its reasoning, so the basis for each call is visible, traceable and repeatable.
Results
First-pass screening, from roughly five hours to around twenty minutes per document.
Analysts spend their time on the deals that already clear the screen.



