Skip to main content

Actions And Consequences

What action should this artifact cause, and what consequence should that action create?

Actions And Consequences is the agent-facing discipline for artifact design. It says every page, agent, journey, and venture surface must name the reader, the tension, the desired state, the first action, and the consequence chain before polish begins.

Operating Rule​

Use this when an artifact can look complete while still failing to move anyone.

Answer five lines before writing:

  • Profile — who arrives, and at what moment?
  • Tension — what friction came with them?
  • Dream — what state would feel solved?
  • First action — what small verb-led move should happen next?
  • Consequence — what direct, downstream, system, cultural, and structural effects follow?

Agent Profiles​

Owner / Principal:

  • tension: wants agent leverage without losing line of sight.
  • dream: a business that runs with receipts worth trusting.
  • first action: authorize one bounded delegation.
  • proof: a receipt arrives inside scope, spend, and time limits.
  • failure mode: scope is skipped, surprise cost appears, and trust collapses.

Developer / Integrator:

  • tension: wants fast integration without inheriting false capability claims.
  • dream: integrations ship against live contracts.
  • first action: read the machine manifest before coding.
  • proof: the integration calls only live capabilities.
  • failure mode: code ships against planned capability and breaks trust.

Operations Director:

  • tension: needs FTE relief without hidden failure modes.
  • dream: workflows move from manual execution to review and improvement.
  • first action: map one workflow to one declared agent capability.
  • proof: every step emits a receipt.
  • failure mode: automation hides the work instead of making it auditable.

Checks​

  • The artifact names one primary reader.
  • The artifact names the arrival moment.
  • The first action starts with a verb.
  • The first action is smaller than the desired outcome.
  • The consequence chain names proof, not mood.
  • A wrong action is named so the surface can steer away from it.

Failure Modes​

  • Reader blur — the artifact says "users" and never names the person.
  • Dream blur — the desired state is a slogan, not a state change.
  • Action drift — the next step asks for too much.
  • Proof gap — the artifact claims impact without a receipt, metric, or signal.
  • Loop break — the action cannot feed learning back into the next artifact.

Context​

  • applies-to Agent Knowledge Contract — this discipline decides whether the source contract causes the right action.
  • depends-on Purpose — a surface must know why it exists before it can ask for action.
  • pairs-with Principles — principles decide which action is worth causing.
  • proved-by Performance — outcomes need a visible proof surface.

Questions​

What action should the next reader take?

  • Who is the reader?
  • What tension did they bring?
  • What state do they want?
  • What is the smallest useful action?
  • What proof shows the action created value?