Engineering for Change
Turning 15 Years of Sustainable Development Knowledge into Decision-Ready AI
E4C challenged builders to unlock their knowledge base with AI. Origin built E4CInsights and won Best Overall out of all competing teams.
Result
1st Place
Track
Track 1: Knowledge Access
Year
2026
01 / The Challenge
A decade and a half of knowledge, largely untapped
Engineering for Change (E4C) has spent over 15 years building one of the most comprehensive vetted libraries of sustainable development resources on the web. The problem: that knowledge lived in a static form. Practitioners, students, and community members could search it but could not get it synthesised into something immediately useful at the point of a decision.
E4C put up a $7,500 prize and issued a challenge across three tracks asking builders to answer three core questions:
- How can AI improve access to the insights E4C has built over the years?
- How might AI increase the relevance of their data for practitioners and students?
- How can AI improve human interactions within E4C's global community?
Track 1, which Origin entered, focused specifically on the first two questions: making knowledge accessible and turning it into something decision-ready.
02 / Why Origin Was Selected
Trust, transparency, and a human in the loop
Most AI approaches to knowledge retrieval summarise and present. Origin took a different position: every claim in any output had to be traceable to a verified source. We called it trust-but-verify.
The judges at E4C selected E4CInsights as the Best Overall winner because it directly served E4C's mission of making engineering knowledge actionable for people making real decisions, while staying rigorous and transparent about where every piece of information came from.
The human-in-the-loop review step was also a key differentiator. Before any policy brief reaches the end user, a human reviewer signs off on it. This was recognised as a responsible approach to deploying AI in high-stakes knowledge domains.
03 / What We Built
E4CInsights: an AI pipeline for decision-grade policy briefs
E4CInsights is an AI pipeline with four core steps:
01
Knowledge Query
The pipeline queries E4C's vetted sustainable development knowledge base to surface relevant, peer-reviewed material for any given topic.
02
Live Data Enrichment
Real-time statistics are pulled from World Bank and WHO APIs, grounding every brief in current, authoritative figures rather than static snapshots.
03
Synthesis
An AI layer synthesises the retrieved knowledge and live data into a structured, fully cited policy brief. No unsourced claims.
04
Human Review
Before delivery, a human reviewer checks the output for accuracy and relevance. The system is designed so the AI assists the human, not the other way around.
04 / The Results
Best Overall, out of all teams across all tracks
E4CInsights won Best Overall at the 2026 E4C AI Pilot Competition, the top award across all competing teams regardless of track.
The judges highlighted the combination of citation rigour, live data integration, and the human-in-the-loop design as the standout qualities. The pipeline demonstrated that AI does not have to choose between being useful and being trustworthy.
Best Overall
Top award across all tracks and teams
Track 1
Knowledge access and relevance
References
From E4C directly
We believe great technology and responsible innovation are the same thing. If you are working on something that matters in sustainability, development, or beyond, we want to hear from you.