Toolkit — Facilitation Script¶

This is a field-ready script for running a 60–90 minute session to diagnose a failure and either adapt an existing system or design a Minimal Viable System (MVS). It is designed to prevent the most common failure mode: producing conversation and artifacts without decisions.

You can run this with a whiteboard and a shared markdown doc.

Preconditions¶

Do not run this session unless:

You have at least one real, recurring failure people can point to.
Someone present can enforce at least one constraint within the unit of analysis.

Roles (lightweight):

Facilitator: runs prompts and enforces timeboxes
Scribe: writes artifacts live (shared doc)
Decider (optional but ideal): person/team with authority for the target decision
Participants: people closest to the failure and the object of control

Hard rule: if you can’t name a decision output by minute 25, stop and re-scope.

Session Outputs (Artifacts)¶

By the end, you must produce:

Observable Failure Statement
Dominant problem frame + decision type
Object of control selection
One of:
Decomposition table (if adopting/adapting), or
System Contract + MVS builder (if inventing)
Constraint + default (non-negotiable)
Misuse warning + mitigation
Next-run plan (who/when/where)

If any output is missing, the session did not complete.

60–90 Minute Agenda¶

0) Setup (0–5 minutes)¶

Facilitator says:

“We are here to improve a specific decision by designing or adjusting a system. We will not leave without a decision output.”
“If we can’t name an observable failure, we stop.”
“If we can’t name a decision output, we stop.”
“Artifacts are mandatory. Constraints are mandatory.”

Scribe opens a doc with headings:

Failure statement
Frame + decision type
Object of control
Candidate systems (optional)
MVS (artifact + constraint + default)
Misuse + mitigation
Next run

1) Failure anchoring (5–15 minutes)¶

Prompt sequence (write answers verbatim):

“What is repeatedly going wrong?”
“Give two recent examples with dates or incidents.”
“What does it cost?” (time, risk, money, morale, customer impact)
“Who is impacted?”
“How often does it happen?”

Scribe writes the Observable Failure Statement (3–5 sentences).

Stop rule: if participants can’t cite concrete examples, stop and assign observation tasks.

2) Frame selection (15–20 minutes)¶

Facilitator:

“Which frame is dominant: strategy, discovery, delivery, cooperation, or evolution?”

Ask:

“If we solved only one thing, which frame would matter most?”

Record:

dominant frame
secondary frame (optional)
one supporting piece of evidence

3) Decision type declaration (20–30 minutes)¶

Facilitator:

“Which decision are we failing to make repeatedly?”

Force one primary decision type:

priority / scope / ownership / sequencing / investment / diagnosis / repair

Prompts:

“If this session succeeds, what decision becomes easier next week?”
“What decision keeps getting avoided or re-litigated?”
“What decision gets forced only by urgency?”

Write:

decision sentence: “We are optimizing the ______ decision.”

Stop rule: if the group insists the goal is “alignment,” translate it into a decision type or stop.

4) Object of control selection (30–40 minutes)¶

Facilitator:

“What can we actually control that influences this failure?”

Choose 1–2 objects max:

goals / work items / interfaces / domains / constraints / incentives / information flow

Prompts:

“What would we edit on Monday to influence this?”
“Who has authority to change it?”
“How often does it change?”

Write:

chosen object(s)
authority holder(s)

Stop rule: if the object requires changing incentives/org structure and nobody present can do it, choose a different object or redesign the unit of analysis.

5) Choose path: adopt/adapt vs invent (40–45 minutes)¶

Facilitator asks two questions:

“Is there an existing system already in play that claims to solve this decision?”
“Is the current landscape colliding (two systems owning the same decision) or is this a gap (no owner)?”

Decision:

Path A: Adopt/Adapt (decompose a candidate system)
Path B: Invent (design an MVS using the System Contract)

If uncertain, default to decomposition first.

Path A — Adopt or Adapt (45–70 minutes)¶

A1) One-sentence system spec (45–50)¶

Prompt:

“In one sentence: what does this system do?”

Use:

“System X reduces __ failure by optimizing _ decisions through control of , producing artifacts, enforced by constraints, at the ___ unit of analysis.”

If you can’t write this, you don’t understand it yet.

A2) Decompose quickly (50–65)¶

Fill the 10 dimensions at high speed:

problem frame
object of control
unit of analysis
causality model
decision type
artifacts
boundary rules
operating mode
misuse model
adoption path

Focus prompts:

“What artifact does it actually produce?”
“What happens when people avoid it?”
“How does it get gamed?”
“Does its causality model match our failure?”

A3) Fit decision (65–70)¶

Choose one:

adopt / adapt / subordinate / reject

If adapt, write the changes as a delta:

artifact change
constraint/default change
authority boundary change
cadence/trigger change

Then jump to sections: Constraint + Misuse + Next run.

Path B — Invent an MVS (45–80 minutes)¶

B1) System Contract (45–60)¶

Fill in:

target situation
observable failure (already)
root-cause assumption
object of control (already)
decision optimized (already)
artifact(s)
constraint + default
misuse warnings + mitigations
adoption path

Facilitator prompts to keep it narrow:

“What is the smallest artifact that makes the decision inspectable?”
“What constraint forces the tradeoff?”
“What default triggers when we avoid it?”

B2) Minimal Viable System builder (60–70)¶

Write the MVS in one block:

Decision:
Artifact (fields):
Constraint:
Default:
Misuse:
Mitigation:

If you can’t state it in 10 lines, it’s not minimal.

B3) First-run plan (70–80)¶

Write:

“We will run this on _ (date) with _ (people).”
“The artifact will live at ____.”
“The decider is ____.”
“The default triggers when ____ happens.”
“We expect to see _ change within _ cycles.”

6) Constraint + default (mandatory checkpoint) (end of either path)¶

Facilitator asks:

“How will people avoid this decision?”
“What rule blocks avoidance?”
“What happens automatically if they still avoid it?”

Write:

constraint
default
who enforces it

Stop rule: if there is no enforceable constraint, you do not have a system.

7) Misuse rehearsal (mandatory checkpoint)¶

Run three rehearsals (5 minutes total):

Busy team: “How will teams weaken this to save time?”
Leadership: “How will this be turned into reporting/control?”
Politics: “How will someone use this as a shield/veto/weapon?”

For each misuse, write one mitigation by changing:

artifact, or
constraint/default, or
authority boundary

8) Close: commit to a real run¶

End with a decision output:

“Keep / modify / subordinate / remove” (if reviewing)
or “Run the MVS once” (if inventing)

Write the commitment:

Date/time:
Participants:
Artifact link/location:
Success criterion (observable change after 1–2 cycles):

Final rule: if there is no scheduled first run, the session produced theory, not a system.

Optional 90-minute extensions¶

Use only if needed.

Extension 1 — Landscape collisions (add 10 minutes)¶

Prompt:

“Which existing system’s artifact will this override or replace?”
“What’s the precedence rule?”

Extension 2 — Sunset clause (add 5 minutes)¶

Prompt:

“When will we review this system?”
“What evidence renews it?”
“What kills it?”