Project: Reasoning Agents - SESAP (Self-Executing Social Agreements Platform)

[princepspolycap]

Track

Reasoning Agents (Azure AI Foundry)

Project Name

SESAP - Self-Executing Social Agreements Platform

GitHub Username

@princepspolycap

Repository URL

https://github.com/Poly186-AI-DAO/sesap

Project Description

SESAP is a platform that transforms unstructured meeting transcripts into structured, executable Smart Social Contracts using the Accord Project technology stack.

Most AI tools merely summarize transcripts into plain text. SESAP bridges the gap between natural language and executable code. It places a 3-step reasoning agent pipeline (using Azure OpenAI's GPT-5 tier models) between raw transcripts and strictly-typed Accord Project schemas.

The core breakthrough is that these contracts act as the foundation for an economy of autonomous agents. Instead of humans doing the signing, we provide a "DocuSign for Agents" execution environment, allowing AI workforces to evaluate, execute, and be held accountable to smart contracts. This forms the base layer for Smart Social Contracts as a Service, using economics and provable execution as a foundation.

Demo Video or Screenshots

Repo & Screenshots: https://github.com/Poly186-AI-DAO/sesap

Primary Programming Language

TypeScript/JavaScript

Key Technologies Used

• Azure OpenAI Service (GPT-5.1, GPT-5-mini, GPT-5-nano)
• React / Vite / Ant Design
• Accord Project (Concerto, TemplateMark)
• Model Context Protocol (MCP)
• Zod

Submission Type

Individual

Submission Requirements

• My project meets the track-specific challenge requirements
• My repository includes a comprehensive README.md with setup instructions
• My code does not contain hardcoded API keys or secrets
• I have included demo materials (video or screenshots)
• My project is my own work with proper attribution for any third-party code
• I agree to the Code of Conduct
• I have read and agree to the Disclaimer
• My submission does NOT contain any confidential, proprietary, or sensitive information
• I confirm I have the rights to submit this content and grant the necessary licenses

Quick Setup Summary

1. Clone the repo: git clone https://github.com/Poly186-AI-DAO/sesap.git
2. npm install
3. Configure AZURE_AI_FOUNDRY keys via cp .env.example .env
4. Run frontend: npm run dev
5. Run backend: npm run server
6. Access Playground at http://localhost:5173

Technical Highlights

• Multi-Agent Reasoning: Uses a 3-tier LLM approach (Extractor -> Generator -> Validator) to coerce messy transcripts into rigorous Accord Project schemas.
• MCP Integration: Built an MCP server so GitHub Copilot Chat can generate, render, and validate smart contracts directly inside VS Code environments natively.
• Execution Environment: A DocuSign-inspired UI designed for future autonomous agents to evaluate, govern, and sign contracts, grounded in economic incentive structures.

Challenges & Learnings

Bridging the gap between LLM hallucination and the strict schema requirements of the Accord Project Concerto models was incredibly difficult. We learned that a single zero-shot prompt fails consistently in this domain, but a multi-tiered reasoning approach with a dedicated validation loop ensures valid artifacts capable of rendering verifiable legal/social agreements.

Contact Information

https://github.com/princepspolycap

Country/Region

United States