From Task Automation to AI Workforce Operations
We’re grateful to AppWorks, OpenAI, and Anthropic for supporting our work with ChatGPT API and Claude credits.
This support enables us to move beyond task-level automation and operate AI systems under real organizational conditions.
From Tasks to Departments
Our current focus is not on building isolated AI tools, but on running department-level AI Workforce operations.
This includes:
- an agentic R&D function, operating with its own execution rhythm and review loop
- an AI-native back office, handling operational and administrative workflows
- authorization flows anchored in EUDI Wallet and TWDI Wallet, treating identity and delegation as first-class infrastructure
These systems are not demos. They are deployed, exercised, and stress-tested as part of daily operations.
The Question We’re Testing
The core question is not which model performs better.
The real question is:
Can AI agents operate as auditable, governable organizational units?
This shifts the evaluation criteria away from benchmark scores and toward operational properties:
- traceable decision paths
- clear authorization boundaries
- reproducible behavior under defined constraints
- failure modes that can be inspected, reported, and improved
AI Workforce Incident Reports
To support this, we are developing and using a structured AI Workforce Incident Report framework.
Its role is not post-mortem storytelling, but enterprise risk control and deployment readiness:
- capturing abnormal agent behavior in operational terms
- linking incidents to identity, delegation, and policy context
- enabling compliance review, insurance alignment, and deployment gating
In practice, this functions as the missing layer between AI capability and organizational trust.
Why This Matters
As AI systems move into core business functions, the limiting factor is no longer model intelligence.
It is organizational compatibility.
We believe the next phase of AI adoption will be decided by whether AI agents can be integrated into existing governance, audit, and responsibility structures—without exception handling becoming the norm.
That is the problem we are working on.
More to come.