Builder's Fund - AI ImpactQF - ReFi DAO & Regen Coordination

[paul]

Status: Executed
Funded Amount: 2500 USDGLO
Authors: @montymerlin & Luiz
Created: 30 Apr 2025
Proposal Link in Gardens: https://app.gardens.fund/gardens/10/0x1eba7a6a72c894026cd654ac5cdcf83a46445b08/0xd3345828914b740fddd1b8ae4f4d2ce03d1e0960/123/0x4ceda4f34d3512900cc03c813e7eff4619ce5cfa-12

Could Common Approach + AI + ImpactQF (along with the supporting tools, methodologies and communities to enable it) be the holy grail for ecosystem grant funding?

1. TLDR

ReFi DAO and Regen Coordination are pioneering an integrated methodology that combines:

• Common Approach to impact measurement
• AI-augmented evaluation processes
• ImpactQF (Impact-weighted Quadratic Funding)

Together, these components enable scalable, transparent, and equitable funding decisions that prioritize real outcomes over marketing hype. After successfully piloting this system in the Regen Coordination Global GG23 round ($100K distributed), we now seek support to expand, document, and refine this methodology for broader adoption in the Allo and Gitcoin ecosystems.

2. Project Summary

We believe that funding decisions should be driven by clear, evidence-driven evaluations that reward real impact, not just good intentions.

As such, ReFi DAO has been leading the implementation and further development of a methodology that combines the Common Approach + AI + ImpactQF. The first deployment of this has been with the Regen Coordination GG23, distributing over $100k in matching and donations to regenerative projects across the globe.

Through continued development and evolution of this methodology, we aim to:

• Support projects that demonstrate tangible contributions to regenerative finance (ReFi) and Web3 ecosystems.
• Strengthen trust among funders, communities, and participants.
• Align blockchain ecosystem growth with measurable ecological, social, and economic impact.

Methodologies and approach will continue to be developed to reflect evolving best practices, ecosystem needs, and feedback from practitioners and evaluators.

3. Contextualization

Common Approach

Common Approach to Impact Measurement is an initiative and set of frameworks designed to support projects in measuring, managing, and communicating their impact in a way that is both rigorous and flexible. It addresses the tension between standardized metrics and context-specific approaches by creating interoperable, adaptable tools and standards for impact measurement.

CIDS (Common Impact Data Standard) is a data ontology that allows organizations to represent and share their impact data in a machine-readable, structured way. It enables:

• Flexible representation of any impact model (logic model, impact thesis, etc.)
• Interoperability between systems and standards (e.g., SDGs, IRIS+)
• Integration of the five Impact Management Project (IMP) dimensions: What, Who, How much, Contribution, and Risk
• Use by multiple software platforms for consistent and shareable data.

For those interested in exploring further, the Common Approach Impact Ontology resource offers detailed information.

Through comprehensive research and experimentation with Common Approach frameworks as well as working actively with partners, ReFi DAO & Regen Coordination aims to support adoption across the ReFi and wider Web3 grants ecosystem.

KarmaGAP Activities Reporting

Through close collaboration with KarmaGAP, ReFi DAO is directly supporting the integration of Common Approach into the platform, enabling more effective tracking and evaluation of project activities and outputs (Common Approach classes).

So far this integration has already enabled:

• Flexible activity and output reporting across diverse projects
• Structured data collection based on Common Approach classes
• User-friendly interface for projects to submit impact data

Projects can efficiently report activities, track progress, and demonstrate impact, creating a standardized dataset that can be evaluated consistently. ReFi DAO has also developed templates and guides to support projects in effectively using the platform.

Human + AI Evaluation

For the Regen Coordination Global GG23 round, we used a hybrid Human + AI evaluation process.

Through careful prompt engineering and system design (see Regen Coordination GG23 AI Evaluation Methodology), multiple AI models were tasked with running a standardized prompt sequence to deeply analyze:

• Application data submitted through the Gitcoin portal
• Project data on Karma GAP, with a focus on Activity & Output reports

Each model generated independent and detailed project reports, providing a consistent, fast, and reproducible baseline across the entire pool. You can browse the full reports here:

• Evaluation Reports (GPT-4o)
• Evaluation Reports (Claude 3.7)

Running multiple models allowed us to compare outputs, check for consistency across different AI providers, and mitigate potential model-specific biases. For example, see Evaluation Comparison: Claude 3.7 vs GPT-4o

Once the AI-generated Baseline Reports and Scoring are produced:

• Coordination Council Members (experts from CeloPG, ReFi DAO, and GreenPill) individually review each project’s evaluation.
• Humans make final scoring adjustments based on nuanced judgment, context, and Regen Coordination’s North Star Outcomes.
• Human scores are averaged to determine the final impact rating.

This approach is inspired by the emerging paradigm described by Vitalik Buterin as “AI as the engine, humans as the steering wheel”:

“AI can tirelessly and scalably perform large amounts of structured evaluation work, but humans provide the critical strategic steering — ensuring judgment, alignment, and integrity in final decisions.”

Rather than enshrining a single AI system, we use AI as a high-efficiency analysis engine, while maintaining human stewardship to uphold credible neutrality, discernment, and ecosystem integrity.

This ensures that:

• High-quality, verifiable evidence is consistently assessed.
• Biases or errors from any single model or evaluator are mitigated.
• Final decisions reflect both systematic evaluation and human-centered strategic judgment.

See more details here The AI workspace that works for you. | Notion

ImpactQF Calculation

The Impact Score generated through the evaluation process is integrated into the funding mechanism using Tunable Quadratic Funding (TQF). This system:

• Modifies matching fund calculations based on verified impact data
• Creates more equitable funding distribution prioritizing high-impact projects
• Addresses inefficiencies in traditional quadratic funding mechanisms
• Provides a data-driven approach to capital allocation

ImpactQF extends the standard quadratic funding model by incorporating impact metrics directly into the matching calculation. By adjusting the matching power based on verified impact scores, the system rewards projects that demonstrate measurable impact while maintaining the crowdfunding and democratized nature of community funding.

Allo Distribution

The final step connects our methodology with Allo-based distribution. Regen Coordination has successfully implemented this approach in the Regen Coordination Global GG23 round (with a $96k matching pool) and localized rounds such as ReFi Mediterranean ($10k matching pool). These rounds serve as real-world demonstrations of how quadratic funding—when paired with verified impact evaluations—can support mission-aligned initiatives while significantly improving the transparency and effectiveness of outcomes.

Looking ahead, we plan to scale this methodology into a global network of bioregional and local community funding rounds, each using Allo infrastructure, Karma GAP data integration, and AI-enhanced evaluation to coordinate capital toward regenerative work. This expansion not only extends Allo’s footprint as a decentralized coordination layer for impact funding, but also provides strong, composable primitives for open-source, impact-driven capital allocation across Web3 ecosystems.

Through this approach, Allo becomes the distribution backbone for cosmo-local regenerative economies—where communities globally can access interoperable tools, run their own rounds, and direct resources to local needs with clear and trusted reporting.

4. Project Deliverables

1. Common Approach Research & Documentation (Nov 1, 2024 - Mar 13, 2025)

• Comprehensive exploration…

2. Karma GAP Development (Nov 1, 2024 - Mar 13, 2025 and ongoing)

• Activities & Output reporting…
• Further integrating Common Approach framework into Karma GAP platform
• Use and explore KarmaGAP’s AI agent, to enhance evaluation capabilities for the Coordination Council
• Providing training and preparing materials for project reporting

3. GG23 Round Execution (Feb 1, 2025 - Apr 16, 2025)

• Successful first deployment of Common Approach + AI + ImpactQF in practice
• Workshops, Presentations, and Webinars: Engaging the ReFi community to explain impact measurement techniques, the Karma GAP platform, and the Common Approach framework, ensuring effective use during the GG23 round and beyond. Training 50+ projects in impact reporting.
• ImpactQF Calculation & Matching Funds Distribution (Apr 17, 2025 - May 1, 2025). Evaluating (Human + AI) project impact reports. Calculate & execute ImpactQF funding distribution.

4. Methodology Playbook Development (May 2025 - June 2025)

• Publishing a comprehensive GG23 round retrospective and analysis
• Developing a detailed playbook/toolkit for local communities and grant rounds
• Documenting implementation steps for ImpactQF + Karma GAP methodology to expand out to many other rounds.

5. Community Support & Adoption (July 2025 - ongoing)

• Providing direct support to communities adopting the methodology
• Facilitating knowledge transfer to new funding rounds
• Building a network of implementers across the ecosystem

5. Team Bios

• Monty Merlin Bryant – Founder at ReFi DAO, board member at Trusted Seed, steward at Celo Public Goods.
• Luiz Fernando – Operations Lead at ReFi DAO, founder of ReFi Barcelona, steward (gardener) at Trusted Seed.

6. Funding Request

$2,500. This micro-grant will allow us to refine, further develop, and disseminate our model across the Allo ecosystem and beyond.

7. Conclusion

ImpactQF, powered by Common Approach and AI-augmented evaluation, represents a leap forward in regenerative funding. By integrating standardized reporting, scalable evaluation, and transparent methodology, we help unlock capital for the projects and communities delivering the most meaningful impact.

Let’s fund what matters — together.