This site publishes independent analysis of Intersect MBO committee governance — examining whether official records accurately reflect what was discussed, how elected members participated, whether commitments were followed through, and whether non-elected staff stayed within their defined roles.
Every finding is grounded in the official meeting transcripts and summaries. The analysis pipeline processes raw committee records through five structured prompts, producing evidence-based reports on each committee's performance. Reports cover: truth audits of official summaries, member participation analysis, leadership assessment, promise tracking, and critical observations.
The Five Reports
01 — Summaries vs. Transcripts
- Forensic comparison of official summaries against raw transcripts
- Identifies fabricated decisions, invented action items, tone laundering
- Flags omitted conflicts and sanitised accountability failures
- Requires transcripts — skipped for summaries-only committees
02 — Member Participation
- Attendance grid for every elected member across all meetings
- Speaking frequency, substantive contributions, between-meeting work
- Power dynamics — who dominates, who is sidelined
- Ranked from most to least effective with transcript evidence
03 — Leadership Assessment
- Chair: facilitation, decision-making, preparation, conflict management
- Vice Chair: support role, independent contributions, accountability
- Secretary: minutes accuracy, boundary respect, tool management
- Overall verdict on whether the committee is well-led
04 — Promises vs. Reality
- Every commitment and action item tracked from source to outcome
- Delivery rate calculated — delivered, partial, failed, recurring
- Most and least reliable promise-keepers identified
- Assessment of whether the committee delivers value to Cardano
Committees
How It Works
Analysis Pipeline
Committee meeting transcripts and summaries are processed through a structured Python pipeline using the Anthropic API. Each committee's source material is fed through five analysis prompts, producing independent reports with full audit trails.
For committees with both transcripts and summaries, a forensic truth audit (Prompt 1) compares every summary against its corresponding transcript, identifying fabrications, omissions, and systematic biases. Prompts 2-5 analyse participation, leadership, promise delivery, and governance dynamics.
Every API call is logged with its exact input, output, and token counts. Any finding can be verified by checking what the model actually received. The pipeline supports automatic chunking for large datasets and synthesis passes for multi-part analyses.
Methodology & Source Notes
All reports are based on official meeting transcripts and summaries sourced from Intersect's public documentation channels. Every factual claim is traceable to a specific transcript passage. Participation data is derived from transcript evidence — attendance lists in summaries are cross-referenced against actual speaking records.
These reports represent independent community analysis and are not affiliated with Intersect or any official body.