01
A buyer has an agent concept but no system boundary
The opportunity is described in features rather than decisions, authority, data, and operating constraints.
Discovery service
Create enough shared technical context to decide whether a machine intelligence or agentic swarm workstream is viable—without pretending a workshop is implementation.
Triggers
The engagement is appropriate when one or more of these conditions is blocking a decision.
01
The opportunity is described in features rather than decisions, authority, data, and operating constraints.
02
Workshare, interfaces, evidence, and client access need to be tested quickly.
03
A small paid step is needed before architecture or implementation commitment.
What the engagement does
The scope is narrowed to one decision boundary and a representative system slice.
How the work proceeds
Each step can narrow the next one as evidence changes the system understanding.
Collect the problem statement, workflow, constraints, and decision date.
Identify actors, agents, data, tools, approvals, outcomes, and failure concerns.
Explore whether multiple agents are justified and where control belongs.
Name responsibilities, access, artifacts, acceptance, and exclusions.
Proceed to a bounded service, perform prerequisite work, or stop.
Boundaries
Explicit inclusions and exclusions keep the engagement decision-focused.
Included
Work that is expected within the agreed system and decision scope.
Excluded
Work that requires a separate decision, scope, authority, or engagement.
Start with a bounded decision
Bring public-safe context about the workflow, agents, data, tools, authority, current evidence, and decision date.