Date: 2026-10-02 3:01 PM EDT Source: https://notes.granola.ai/d/c7c5a22e-a9cc-44db-87b6-015703a5d406
Attendees
- Michael Houck (Rye Valley)
- Emil Barr (Candid Network / Flashpass)
- Hunter (micro1, technical lead for government team)
Summary
Exploratory call between Emil/Michael and Hunter from micro1 about potential data brokering: Emil’s company has relationships with state governments and can broker non-PII state agency data; micro1 does AI work for government clients and may have appetite for structured operational data. Ended with a soft pilot interest and a request to follow up Monday with volume/pricing details.
Emil’s Company (Candid Network / state gov SaaS)
- SaaS tools live across 25 states; lobbyists in nearly every state
- ~$2M/month top-line revenue
- Core niche: filling the gap left by poor Accenture/Deloitte deployments at state level
- Competition is essentially just those two incumbents
- Have previously licensed data from Ohio, Montana, and Louisiana to train a waste/fraud/abuse AI agent
- Data types available: training manuals, email records, weekly reports, operational docs
- Explicitly non-PII: anything that doesn’t directly identify individuals
- Structure: raw files (Google Drive links, zip files); engineering team can manipulate if needed
- Have pulled 100TB in the past; theoretically can go up to that across all 50 states
- Connected via mutual contact (Ali, friend of the founder)
What micro1 Wants
- Operational, connectable data reflecting how a state agency actually works day-to-day
- Outlook, Slack, Microsoft Teams, Google Drive, Figma
- Structured to be representative of a real operating environment
- NOT interested in: static PDFs, old presentations, low-signal documents
- Primary focus is federal (bigger budgets and GPUs), but state-level datasets have a clear upstream use case
micro1 Context
- Hunter leads technical direction for the government team at micro1
- Work includes PII anonymization, dataset curation, hill climbing for government
- Most work at federal level given massive budgets and GPU resources
- State-level as a target is not immediate but there is need for datasets that flow up to federal
Pilot Structure Discussed
- Start: one state, one agency, small volume
- Pricing TBD: Emil to follow up after call with volume estimate and pricing
- micro1 has a systematic, volume-based pricing approach for large-scale corpus data
- No payment until data validated as useful
- Potential buyers: micro1’s internal government team, and possibly the labs
Decisions
No formal decisions. Both sides expressed soft interest in a small pilot.
Action Items
- Emil: Send pilot scope and pricing estimate (one state, one agency to start) with volume and pricing structure
- Hunter: Send over micro1’s standardized corpus data purchasing approach so Emil’s team can back into pricing
- Emil/Michael: Ping Hunter on Monday to align on pilot details and confirm whether internal government team or labs have appetite
Notable Transcript Excerpts
“State governments are kind of like the ugly duckling where it’s not quite as sexy as federal, so they don’t get the same amount of dev talent and attention. And basically Accenture and Deloitte do like these really crappy deployments across state. That’s like the entire competition.”
“[What we’d want is] Outlook, Slack connectors, Google Drive, Figma, anything that makes a company a company in terms of how it operationally works…to be representative of a state agency, that becomes very valuable to us. A PDF of a presentation that they gave 6 years ago is probably not very valuable to us.”
“Most of our work is at the federal level. Just given that they have massive amounts of money and GPUs and stuff for us. But yeah, definitely state as a target for us at some point in time. Definitely not immediate, but there’s definitely need for datasets on the state level to kind of flow up to the federal level.”