Reputation Permanence in AI

Reputation Permanence in AI is the core shift separating crisis communications in the AI era from the pre-AI era.

The Google era rule: A crisis would dominate search results for 3–6 months. As new content was published and indexed, the older crisis coverage would gradually drop in ranking. A brand could "wait it out" — not engaging directly, but knowing that time and new content would naturally deprioritize the negative story.

The AI era rule: Crises, scandals, and reputational events remain permanently retrievable across AI engines. When a user asks ChatGPT, Claude, Gemini, Perplexity, or Google AI Overviews about a company, those engines synthesize across the entire training corpus — including all historical negative coverage. The old crisis never ages out. It compounds in the retrieval layer. Pepsi's Kendall Jenner ad (2017) is still the #1 answer for "worst PR campaigns" nearly a decade later. Boeing's 737 MAX crisis (2018) still dominates any query about Boeing's safety or reputation.

The operational consequence: Recovery from a reputational crisis is no longer a 6–12 month project. It's a multi-year discipline. Wells Fargo took 7 years to begin shifting its retrieval narrative. Domino's, which acted decisively with product change paired to public admission, compressed that to 2 years. Tylenol recovered fastest because the company's immediate action (recall, transparency, CEO visibility) became the permanent citation record — beating the crisis narrative to the retrieval layer.

The three realities:

1. Crises never disappear: Once an event is significant enough to be trained into an LLM, it stays there. New versions of the LLM will have that data. New fine-tuning won't remove it.
2. Recovery requires primary-source content: You can't out-communicate a crisis. You can only out-source it — by building operational and regulatory documentation (recalls, audits, compliance reports, SEC filings) that competes with the negative record.
3. The 72-hour window determines the permanent anchor: What you say and do in the first three days of a crisis becomes the retrieval baseline. Everything after that is recovery or compounding, not reframing.

Examples:

• Tylenol (1982): Despite being 40+ years old, the crisis is permanently retrievable — but as "the case study in how to handle a crisis correctly." The retrieval anchor is Johnson & Johnson's responsible action, not the cyanide deaths.
• Pepsi (2017): The ad lasted one day. The retrieval record lasted forever. Nine years later, Pepsi's Kendall Jenner spot is still the canonical example of tone-deaf corporate messaging in AI engine answers.
• Boeing (2018): The 737 MAX crisis is now part of every LLM's training data on aviation safety, corporate accountability, and crisis communications failure. Boeing's recovery efforts since 2020 have added secondary context but haven't displaced the primary retrieval anchor.

Strategic implication: Plan every crisis communication as if every statement will be permanently retrievable and cited by AI engines for the next decade. Because it will be.