Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{199497,
        author = {Aryav Jain and Akshat Kumar and Pratibha Gautam and Sweta Yadav},
        title = {Sarthi – The Helping Hand: An AI-Assisted Framework for Optimizing NGO–Volunteer Coordination},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {14235-14245},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=199497},
        abstract = {Efficient mobilization of volunteer labor within India's fragmented civil society sector remains a consequential socio-technical challenge, compounded by persistent informational asymmetries and behavioral attrition that undermine even algorithmically sound matching systems. Prior work on the Sarthi AI-assisted coordination framework established content-based filtering as a viable baseline, achieving a matching precision of 0.71; however, a 22% volunteer no-show rate exposed a critical reliability gap between digital assignment and physical task execution. This paper presents an architectural evolution of the Sarthi framework that transitions the system from passive recommendation toward high-fidelity, behaviorally grounded coordination. The proposed methodology integrates four interconnected mechanisms: a 24-hour pre-commitment protocol functioning as a Human-in-the-Loop behavioral filter; Structured Task Specification templates employing Jaccard similarity for skill alignment and Haversine distance for geospatial proximity; a Bayesian Reputation Score modeled as the posterior mean of a Beta distribution over historical task completions and failures; and an epsilon-greedy Multi-Armed Bandit allocation reserving 15% of opportunities for a Shadowing Tier to ethically onboard unverified volunteers and mitigate cold-start inequity. The modified compatibility scoring function incorporates reputation as a multiplicative reliability weight alongside skill, availability, location, and cause-alignment features. Experimental projections derived from a 12-week pilot dataset indicate that these interventions are expected to reduce volunteer attrition from 22% to 12–15% and elevate matching precision from 0.71 to 0.85. The framework offers a replicable, equity-conscious methodology for optimizing human capital allocation in resource-constrained civic environments.},
        keywords = {Artificial Intelligence, Bayesian Inference, Behavioral Verification, Civil Society Technology, Content-Based Filtering, Digital Volunteerism, Exploration-Exploitation Tradeoff, Geospatial Matching, Human-in-the-Loop Systems, Multi-Armed Bandit, NGO Coordination, Recommender Systems, Reputation Scoring, Resource Optimization, Socio-Technical Systems, Volunteer Management},
        month = {April},
        }