AI-driven micro-donation integration: Enhancing social impact through seamless financial workflow
Loading...
Supplementary material
Other Title
Authors
Ikeda, A.
Pan, J.
Sharma, S.
Song, Lei
Shakiba, Masoud
Pan, J.
Sharma, S.
Song, Lei
Shakiba, Masoud
Author ORCID Profiles (clickable)
Degree
Grantor
Date
2025
Supervisors
Type
Conference Contribution - Oral Presentation
Conference Contribution - Poster Presentation
Conference Contribution - Poster Presentation
Ngā Upoko Tukutuku (Māori subject headings)
Keyword
charities
fund raising
donors and donations
microfinance
invoices
recommender systems (information filtering)
dashboards (management information systems)
AI in accounting
fund raising
donors and donations
microfinance
invoices
recommender systems (information filtering)
dashboards (management information systems)
AI in accounting
ANZSRC Field of Research Code (2020)
Citation
Ikeda, A., Pan, J., Sharma, S., Song, L., & Shakiba, M. (2025, December, 1-5). AI-driven micro-donation integration: Enhancing social impact through seamless financial workflow [Paper presentation] [Poster presentation] ITP Rangahau & Research Symposium 2025 + OPSITARA 2025, New Zealand
https://hdl.handle.net/10652/7143
Abstract
This project integrates an AI-driven micro-donation system into an invoicing process, allowing business users to contribute to local charities by donating a portion of their revenue. Specifically, this research investigates the AI-driven recommendations that match donors with causes that align with their values and past behaviour, which enhances transparency and user engagement. The traditional donation system on financial platforms often experiences a lack of personalisation, transparency, and integration with user workflows, such as invoicing. This project was managed by agile practices, involving iterative sprints to refine features based on stakeholder input and technical constraints. We developed a real-time donation impact dashboard and an agent for user support. The agent is conducted on multimodal GPT4.1 with a supportive Retrieval-Augmented Generation (RAG) and collaborative prompt engineering tuning. A responsive web application has been built to test the integration of donation features with invoicing workflows, ensuring usability and transparency. Due to the project being confidential, we set up an internal evaluation survey community with five experts. The agent received exceptional feedback for its responsiveness, clarity, and integration into the donation process, with an average score of 4.92 out of 5. The overall user acceptance testing (UAT) achieved an average score of 4.66 out of 5, reflecting intense user satisfaction across donation interactions, dashboard metrics, and chatbot support. Most user stories were successfully fulfilled, confirming that the system met its core functional and usability goals. In addition, the agent doesn’t collect any personally identifiable information (PII), only donation and invoice amount; no ethical concerns are raised in this research. The research discovered the potential of AI in donation systems to optimise user behaviour and enhance donor trust, offering a model for similar applications in other scenarios, such as education, healthcare, etc.
Publisher
Permanent link
Link to ePress publication
DOI
Copyright holder
Authors
Copyright notice
All rights reserved