Crisis Management Decision Tree

GEO Crisis Management Decision Tree | Three Streams Methodology

Decision Progress Step 1 of 5

Source

Step 1: Crisis Source

Where did the problem originate?

AI Platforms

AI Misinformation Detected

AI generating incorrect or harmful info about your brand.

Examples: Hallucinated warnings, fabricated lawsuits, incorrect pricing

Community

Negative Community Sentiment

Viral criticism or hostile discussions on community platforms.

Examples: Reddit pile-on, review cascade, influencer conflict

Customers

Customer Confusion from AI

Customers acting on incorrect AI-provided information.

Examples: Support tickets citing AI, sales objections from AI claims

Internal

Actual Organizational Incident

Real product, leadership, or organizational issue.

Examples: Product recall, data breach, executive misconduct

📋

Step 2: Crisis Category

Classify the crisis type

Category 1

AI-Generated Crisis

AI systems generating false information about your brand.

Auto-assigned: Victim Cluster (you're the victim of AI misinformation)

Category 2

Community Crisis

Crises originating in community platforms (Reddit, social, reviews).

⚠️ GEO Impact: Reddit = 40.1% of LLM citations

Category 3

Product Crisis

Actual product problems—quality, safety, or performance.

Key question: Was issue previously known?

Category 4

Reputation Crisis

Ethics, leadership, or broader brand perception issues.

Auto-assigned: Preventable Cluster (organizational responsibility)

❓

Step 3: Clarifying Question

One question to determine response strategy

📊

Step 4: Severity & Risk Factors

Determine urgency and escalation factors

⚖️

SCCT Cluster: —

—

What is the current impact level?

Level 1ROUTINE — 72-hour window

Minor inaccuracies with no customer or business impact.

Level 2MINOR — 24-hour window

Incorrect claims that could mislead customers.

Level 3CUSTOMER IMPACT — 4-hour window

Safety claims, damage allegations, or health concerns.

Level 4BUSINESS IMPACT — 1-hour window

Fabricated lawsuits, regulatory actions, investor confidence.

Level 5EXISTENTIAL — IMMEDIATE

Widespread multi-platform crisis threatening business viability.

🎯 Crisis Response Plan

Recommended Response Strategies

🔧

Platform Correction Procedures

GPT

ChatGPT

Thumbs-down on inaccurate responses
Provide structured correction with sources
Level 3+: Submit at help.openai.com

Gemini / AI Overviews

Use in-product feedback
Update Google Business Profile
Submit to Wikipedia/Wikidata

Perplexity

Flag inaccurate responses
Key: Correct cited sources directly
Level 3+: [email protected]

Claude

Use feedback mechanisms
Update authoritative sources
Level 4+: Enterprise support

👥

Crisis Response Team

📋

RACI Matrix

R Responsible A Accountable C Consulted I Informed

📅

30-Day Remediation Protocol

For persistent Level 3+ misinformation or source contamination.

Week 1: Audit
Document all AI brand mentions, create baseline, identify cited sources

Week 2: Update Sources
Update owned properties, schema markup, Wikipedia/Wikidata

Week 3: Content Creation
Publish authoritative content, FAQs, press release if needed

Week 4: Outreach & Re-Audit
Contact sites with incorrect info, compare to Day 1 baseline

📝

Post-Incident Review

Complete within 5 business days of resolution. Per ISO 31000 & COSO ERM.

PIR Sections

Incident Summary → Timeline → Root Cause Analysis → Response Evaluation → Lessons Learned → Sign-Off

📁

Storage Requirements

Incident Log — Real-time during crisis
PIR Repository — Searchable by category/severity
Risk Register — Quarterly updates from PIR findings
Retention: Minimum 3 years

Educational tool only — not professional advice. Consult qualified experts before acting.

✍️ Strategy & Content

⚙️ Technical & Implementation

📊 Measurement & Monitoring