- Original authors / creators
- Henry Maathuis; Marcel Stalenhoef; Sieuwert van Otterloo; Raymond Zwaal; Kees van Montfort; Danielle Sent
- Original publication
- Frontiers in Artificial Intelligence · Frontiers Media SA · 2026-03-05
- License & retained notices
- CC BY 4.0
Copyright © 2026 Maathuis, Stalenhoef, van Otterloo, Zwaal, van Montfort and Sent.
Article-specific license evidence ↗Exact license notice
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Article-specific license URI: https://creativecommons.org/licenses/by/4.0/ - Edition changes
- Narrative paragraphs and references retained with whitespace normalization; figures, tables and display equations remain in exact retained XML and original source, not the reader edition.
- Evidence limits
- Adjacent research, not a directly sampled family-office study. Applicability to a family office requires separate assessment. To address these limitations, our study conducts a human-centered evaluation of visual presentation formats for communicating explainable AI (XAI) outputs in the financial domain, framed as an exploratory analysis due to the small sample size. We adopt a model-agnostic approach, focusing not on generating explanations from specific algorithms but on how different types of pre-defined explanations—such as feature importance or counterfactuals—can be effectively presented and understood. Our investigation centers on two real-world use cases: (1) business credit provision and (2) automotive insurance claim fraud assessment. Since all participating end-users already work directly with these AI-supported decision processes, they possess an inherent understanding of the underlying use case, decision task, and role of the AI system. The visualizations and explanation designs were therefore aligned with their existing domain knowledge and everyday workflows. End-users evaluated explanations on explicitly given dimensions such as usefulness, clarity, and actionability, which reflect meaningful engagement with the context rather than mere visual appeal, and generally rated the designs positively, suggesting they supported understanding and decision-making within their specific workflows (Kim et al., 2024b). Despite these strengths, the study has limitations. The sample size was small and drawn from two financial organizations, which may limit generalizability; broader and more diverse samples across domains would strengthen future work. Additionally, the focus on financial decision-making may reduce applicability to other high-risk domains such as healthcare or emergency response. Finally, real-world deployment and longitudinal studies are needed to evaluate long-term effects on decision-making and trust in AI.
- Retraction / correction check
- No retracted-publication label in the retrieved Europe PMC record or Crossref update-to/updated-by notice. XML metadata screened for retraction. Publisher article page searched for explicit correction/retraction/concern notices; see verification record. These bounded checks do not cover unlinked or unregistered notices and are not exhaustive independent retraction clearance. Verified 2026-11-05. Automated/source checks are not human peer review.
- Preserved record
- Rights and full text checked 2026-11-05. Included in this volume on 2026-11-05. Original source dates remain unchanged.
Download the preserved source
SHA-256 (retrieved artifact): 1b3e5ca19b6d31a899540d68e47bb85de8e4a805f1cbe43caec5a49b4d74795c
SHA-256 (preserved text edition): 78b06f4edba2397cfc2f09e702bbca62f8f3c1c04888ad789730ed1928667468