Research
Papers and briefs, free to read.
Methods, maps and design principles for agents that act. No forms, no gates, and every number sourced or labelled as a design target.
Latest · Research
Agents at work: where enterprise automation needs a person
A practical map of which agent actions can run alone, and which need a named person.
elan1 research team · · 4 min read
Read the paper : Agents at work: where enterprise automation needs a personMore research
Showing 1–3 of 3 papers and briefs.
- Research
The confidence line: where AI hands a claim to a person
Calibrated confidence and a signed cost matrix decide which claims a handler sees.
· 4 min read
- Research
Fail toward a human, never toward an action
What should happen when an AI system is unsure, or goes down.
· 3 min read
- Research
A record a stranger can check: designing an audit trail for agents
Why an agent's record should be verifiable offline, without trusting the vendor.
· 3 min read
Readiness check
Is your first process ready for an agent?
12 questions about one process you have in mind: what it changes, who approves it today and what your auditor will ask.
Use it to choose where to start, then bring the answers to a working session.
Subscribe
One email a month. No hype.
New perspectives, research and regulation watch, gathered in one plain email. Unsubscribe in one click.
Intelligence you can answer for.
Start with one process. Add the next when it holds.