MELvan: An AI-Assisted Approach to Capturing Program Impact, Reducing Cognitive Load for Leaders Documenting Impact on the Go

Challenge: Unintended and emergent program outcomes are crucial to capture for accountability and learning, yet they are easily lost without the right tools and processes. Program leaders often notice evaluation-relevant moments as they unfold — an unintended result mentioned in a meeting, an informal conversation — but lack the time to reflect or document it on the spot. One program leader (co-author) described this challenge: "The moment passes — and so does the chance to document consequences of our program." She wished she had an assistant — something simple on her phone she could take everywhere to help capture these emerging outcomes on the go.

Solution: MELvan — a custom AI assistant prototype built on ChatGPT — operationalizes a division of cognitive labor across time and expertise. Leaders capture short notes, photos, or slides during daily work. MELvan then provisionally classifies their input into outputs, outcomes, or indicators, flags gaps, and proposes new ones. Entries are then reviewed during dedicated sessions in which program leadership and the evaluation specialist engage fully with the content, bringing program and evaluation expertise. (*MELvan is a wordplay on Monitoring, Evaluation, & Learning)

Methodology, Rubric, & Lessons: Forthcoming