Relational AI: From Principles to Practice in Empowering College Advisors
Tools and Resources|AIData and EvidenceEquity
At Project Evident, we believe AI can help college advisors reach more students, but only if it strengthens the human relationships at the heart of advising. Multi-year intensive advising programs have the strongest evidence of impact on college access outcomes, and they work because advisors offer both practical and emotional support. Our Hybrid Advising work asked one question: can AI chatbots safely provide some of that emotional and motivational support, so advisors can focus on what they do best?
The resources below capture that work and help college access organizations, school districts, and funders move from curiosity to responsible action as they expand college access resources through relational AI.
The white paper documents the full project: why relational AI matters for college access, how we developed our relational AI in college advising design principles with practitioners, and what five pilot organizations learned when they put the Technical Quickstart Guide into practice. It also introduces a measurement framework that goes beyond engagement metrics, tracking whether students leave an AI interaction more confident in their ability to persist in their college journey.
The Technical Quickstart Guide helps college advising organizations ensure their chatbots best support their students motivationally and emotionally while keeping each organization’s mission and values at the center. Upload your chat transcripts and get a readable report that flags conversations that need a human eye, based on the guardrails your organization sets.
Project Evident hosted a virtual convening on October 28, 2026, bringing together college access practitioners, funders, and technical experts. Sessions covered the foundations of responsible relational AI design, how to measure success, two tracks outlining technical approaches for implementing relational AI, and demos in which each pilot organization shared its work.