L3 · Execution
Watch the work move and the outputs appear.
Pick a line and press play. Each step names the agent that holds it, its autonomy level, the gate that can stop it and the artifact it produces. Gates halt the run until a human releases them.
Steps simulated
98
Handoffs
91
Artifact classes
12
Gate stops
15
Choose the situation
The situation
A regulatory docket moves in one jurisdiction and immediately puts filed positions in two others at risk of inconsistency.
Trigger
Docket movement detected by the Policy Radar
Clock
Weeks — but the inconsistency window is immediate
Ceiling rationale
A3. Collection, normalization, analysis, narrative-share computation and draft production run unattended. What the board is told is decided by a person.
Run state
Steps done
0/18
Artifacts
0
Gates cleared
0
Human touches
0
Canon stores written
0
Overrides
0
Autonomous so far
—
Detect the development in its originating jurisdiction — docket, consultation, filing, enforcement action, legislative text
Map to affected business activity and quantify exposure — restaurants, franchise entities, cost impact, operational change required
Scan all other jurisdictions for parallel or contagion developments; forecast which markets follow and on what timeline
Retrieve the company's currently filed positions in every jurisdiction; flag inconsistencies between them
Map the stakeholder field — legislators, regulators, trade associations, opposing coalitions, allied industry, franchise association positions
Build scenario set with trigger conditions, trajectories and prepared positions for each branch
G1 POSITION — the global policy position, with explicit approval of where jurisdictional variation is permitted and where it is not
Draft the per-jurisdiction position papers, consultation responses and filings from the approved global position
Verify every claim per jurisdiction against Evidence Ledger scope; flag anything unsubstantiated in that jurisdiction
Check cross-jurisdiction consistency — the filing in Brussels against the filing in Sacramento, on the same facts
Simulate regulator, legislator, coalition and franchise-association reaction to each filing
G3 RISK — regulatory, litigation and franchise-relations exposure of the filed positions taken together
G5 RELEASE — filings authorized per jurisdiction against their deadlines
Produce the supporting communications set — franchise-operator briefing, employee explanation, press materials, allied-coalition materials
Track progression in every jurisdiction — hearings, amendments, votes, enforcement — and alert on deviation from the forecast
Monitor opposing coalition activity and narrative; flag when their framing is winning
Measure outcomes — legislative results, regulatory outcomes, narrative position, franchise-operator confidence
Write back — what arguments worked in which jurisdiction, with which audiences, so the next campaign starts from evidence
Output shelf
0Everything the run has actually produced. Click an item to read the output contract the producing agent is held to.
Nothing produced yet. Run the situation.
Canon touched
No write-backs yet.
Architectural rule R2: no output leaves the system without a Canon write-back. That is the difference between a tool that produces content and an organization that gets smarter.
Output taxonomy
What a step actually produces
Every specified step leaves an output. Grouping them into classes shows what was produced, by whom, under which level, and what it changed.
The work reached the audience it was built for.
Publishable language in the house voice.
The organization is permanently smarter for it.
A judgment with its drivers attached.
What the company says, and why it is defensible.
The same position, landed in-market.
A position of record with an external body.
A claim is substantiated and traceable.
What it changed, counted rather than claimed.
A named human accepted the consequence.
Something in the world was noticed and named.
A human is prepared to act or answer.
Why the bench behaves this way
Three rules govern every run
The simulation obeys the same three architectural constraints the platform is specified under.
R1
No agent-to-agent coordination. Every handoff goes through the Chief of Staff.
Swarms that self-coordinate cannot be audited, debugged or governed. The cost is a small latency penalty; the benefit is that every decision has a traceable path.
R2
No output leaves the system without a Canon write-back.
This is what compounds. Without it, the platform is a very expensive drafting tool.
R3
The autonomy ceiling is data, not code.
Legal must be able to ratchet down a ceiling at 23:00 on a Friday without a deployment.