Open Research Institute Pitch Jam

The Pitch Jam

I entered the Open Research Institute's Pitch Jam in March 2026 with a project called the Governance Memory System. GMS is a system that builds queryable institutional memory for organizations. It captures why decisions were made, what happened afterward, and which patterns keep recurring across leadership cycles. Essentially, version control for governance!

GMS tied for first out of 15 entries. That part is nice, but it is not why I am writing this.

From DAOs to Parliaments

A participant in the competition works in government digitalization in a developing democracy. He saw my pitch and reached out because his country's parliament has the exact problem I built GMS to solve. It's a multi-party democracy, and every election has resulted in a change of the ruling party or coalition. Each incoming administration takes the previous government's legislative work, rebrands it, and reintroduces it as its own — a cycle that has produced over a decade of stagnation, with the same initiatives being recycled across parliaments.

Most young people are politically disengaged, not out of apathy but because the system has given them no reason to believe participation changes anything. The monitoring infrastructure already exists. What does not exist is an analytical layer that can prove these patterns are happening and make them visible to citizens.

The Prototype

I built a working prototype only using publicly available data from their parliament. Session transcripts I downloaded manually, published budget tables, and a government monitoring website. No insider access, no API keys, no government cooperation. The system:

• Scored the parliament's governance health across five dimensions
• Mapped informal power dynamics inside the ruling supermajority
• Detected intra-party factions through co-voting network analysis
• Flagged budget areas where new laws are creating obligations that the budget is not keeping up with
• Classified every member of parliament into behavioral archetypes based on what they actually do: sponsorship, attendance, committee work, voting patterns

The government contact independently validated the output. He confirmed the system picked up on politically sensitive dynamics without any prior knowledge of his country's politics. He is now sharing GMS with sitting members of parliament and technical colleagues across government agencies. The goal is to pilot at the provincial level first, then scale to the national parliament.

Three Connections

The Pitch Jam produced two other connections on top of that. Bonfires.ai, a community archival platform whose data infrastructure aligns with one of GMS's analytical layers, is a potential technical collaborator and is now reviewing the NEAR architecture spec. Three connections from one event: a deployment partner, an integration partner, and a build partner.

The Proof

This parliamentary work is one of two active deployment contexts. The other is blockchain governance, where GMS started, and where I have five years of experience embedded across fifteen different ecosystems. Running the same framework in both contexts proved something I had been hypothesizing on paper: institutional memory loss is a structural problem, rather than a domain-specific one. The five-layer framework was transferred to a sovereign parliament without any architectural changes.

GMS is entirely self-funded. No external backing, no institutional affiliation, no revenue. The prototype, the research paper, and the working dashboard, along with all three partnerships, came out of personal savings and credit. This isn't a sustainable model.

The full write-up with technical details, deployment plans, and funding breakdown is available, along with funding support options through OpenCollective and Artizen Fund.

Prepared by: Othman Gbadamassi