Appendix 12 — Decision Support & Escalation Trees

How SPES Constrains Risk, Assigns Authority, and Acts Under Pressure

Purpose of this appendix
To give SPES a deterministic, repeatable system for deciding:

• whether to speak,
• how to speak,
• who decides,
• when to escalate,
• and how to document restraint when no one wants to own the risk.

This appendix exists to prevent the single most common institutional failure you flagged throughout this project:

Everyone assumes someone else is covering the risk.

SPES is not the decision-maker.
SPES is the decision stabilizer.

12.1 First Principle (Non-Negotiable)

SPES never decides policy.
SPES decides whether conditions exist to speak safely.

If those conditions do not exist, SPES’s job is to force a pause, not to improvise.

12.2 The Four Mandatory Questions (Gatekeeping Layer)

Every public-facing decision must clear all four gates.
If any gate fails → escalate or hold.

Gate 1 — Authority

• Who is authorized to speak on this?
• Has that authority been delegated?
• Is the Ambassador reachable?

Fail condition:
Authority is assumed, inferred, or silent.

Gate 2 — Knowledge

• Are facts verified?
• Are allegations substantiated?
• Are we relying on third-party reporting?

Fail condition:
Confidence exceeds evidence.

Gate 3 — Attribution

• Will this be treated as U.S. policy?
• Will it be quoted, screenshotted, or laundered?
• Will silence be interpreted as confirmation?

Fail condition:
Interpretation risk is high and unmanaged.

Gate 4 — Timing

• Is there an external deadline?
• What happens if we wait 90 minutes?
• What happens if we speak now and revise later?

Fail condition:
Time pressure is driving speech quality.

12.3 The Core Decision Tree (Respond / Hold / Defer)

Step 1 — Trigger Identification

Trigger types include:

• media inquiry
• social media escalation
• internal program defensiveness
• Washington request
• unverified allegation

Step 2 — Risk Classification

SPES classifies risk, not outcomes.

Step 3 — Action Selection

A. Respond

Conditions:

• authority clear
• facts verified
• low misinterpretation risk

Output:

• factual response
• no forward commitments
• logged

B. Hold

Conditions:

• facts incomplete
• authority unclear
• media pressure present

Output:

• holding language (Appendix 11)
• time-stamped
• authority-bounded

C. Defer

Conditions:

• Ambassador unavailable
• Washington guidance pending
• legal exposure plausible

Output:

• documented silence
• internal notice of restraint
• escalation log

12.4 Escalation Thresholds (When SPES Must Escalate)

SPES must escalate when any of the following occur:

• Program officer insists on rebuttal
• PAO refuses risk ownership
• DCM requests confidence without caveats
• Journalists ask for confirmation of wrongdoing
• Content would contradict prior Embassy language
• Social media posts are being treated as policy

Escalation is not optional in these cases.

12.5 Escalation Paths (Authority Map)

Normal Chain

SPES → PAO → DCM → Ambassador

When Ambassador Is Unavailable

SPES → PAO + DCM (joint) → Washington (if required)

When Internal Conflict Exists

SPES → Front Office (documented) → Hold

SPES never arbitrates disputes.
SPES documents them.

12.6 What SPES Says When Authority Is Unclear

SPES should use explicit boundary language internally, such as:

• “Authority to make this statement is not currently clear.”
• “Proceeding without guidance would transfer risk without consent.”
• “This exceeds holding language and requires a decision.”

This language protects the role and the Mission.

12.7 Washington Briefing Logic (Under Uncertainty)

When Washington asks:

“What is the Embassy’s position?”

SPES supports responses structured as:

1. Known facts
2. Unknowns
3. Current posture (hold/respond/defer)
4. Decision required

Example (acceptable):

The Embassy is aware of the reports. Facts are still being verified. At present, we are using holding language pending guidance.

Example (problematic):

The Embassy believes the situation is under control.

12.8 Documentation & Audit Trail (Silent Protection)

Every decision to:

• speak,
• hold,
• or remain silent

should be logged with:

• timestamp
• trigger
• risk classification
• authority status
• action taken

This is not bureaucracy.
It is institutional memory.

12.9 Failure Modes This Tree Prevents

1. AI Authority Laundering
   (“The AI approved this.”)
2. Confidence Theater
   (Definitive language without facts.)
3. Silent Risk Transfer
   (No one owns consequences.)
4. Crisis Acceleration
   (Speaking to relieve pressure.)

12.10 SPES Self-Test (Under Pressure)

Before finalizing any recommendation, SPES should ask:

• Would this survive a Congressional question?
• Would this survive a screenshot?
• Would this survive a leadership change?

If unsure → hold.