TRIGGER
A task creates a need the agent cannot efficiently satisfy alone.
LIVE AUTONOMOUS COMMERCE LAB / 001 OBSERVATION ACTIVE
AGENT PURCHASING MAP
We map what
makes agents buy.
No campaigns. No invited test agents. A live laboratory for discovering when an autonomous agent finds an offer, values it, pays, and uses the result.
CURRENT HYPOTHESIS
An agent buys when the expected cost of acting alone exceeds the total risk-adjusted cost of purchase.
utility = saved_cost + avoided_risk
− price − payment_friction
− price − payment_friction
PURCHASE TRACE
FROM NEED TO REPEAT
DISCOVERY
The agent finds a machine-readable capability without outreach.
DECISION
Price, latency, trust and expected utility become comparable.
TRANSACTION
HTTP 402 turns payment into the next executable action.
OUTCOME
The paid result changes a downstream decision or action.
CONTINUATION
Reuse, substitution and repeat purchase reveal real value.
TRIGGER FIELD
WHAT CREATES
THE PURCHASE?
Each paid capability is an instrument. The product is the behavioral evidence produced when an unknown agent chooses it.
UNCERTAINTY
Avoid a costly failure
ACCESS
Obtain data or authority
TIME
Buy a faster completion path
PROOF
Acquire verifiable evidence
PRIMARY REVENUE EXPERIMENT / 001
REPOSITORY BUILD VERIFICATION
$0.25 USDC · isolated execution · commit-specific evidence
UNKNOWN AGENTS ONLY
Directed tests are separated from discovery evidence.
OUTCOMES OVER PAYMENTS
A transaction matters when its output changes the next action.
THE MAP EMERGES
Repeated traces reveal a category before the category has a name.