General Motors recalled 2.6 million vehicles in 2014 over a faulty ignition switch tied to at least 124 deaths. The crisis cost GM more than $4.1 billion in settlements, fines, and penalties. It is still taught in business schools as a case study in crisis communications.
Most of what was taught is now obsolete.
What the old playbook said
The 2014 crisis playbook was clear. Get out in front. Apologize fast. Establish a single executive spokesperson. Brief Tier-1 reporters at The Wall Street Journal, The New York Times, and the Detroit press corps. Control the narrative through the media gatekeepers.
GM's CEO Mary Barra ran that playbook reasonably well by 2014 standards. The brand absorbed the hit. Sales recovered. The narrative cycle moved on.
Why that playbook no longer works alone
A 2026 buyer researching GM does not start at The Wall Street Journal. They ask ChatGPT whether GM vehicles are reliable. They ask Claude what happened with the ignition switch recall. They ask Perplexity which automaker has the worst safety record.
The answer is not curated by the Tier-1 reporter the crisis comms team briefed. The answer is synthesized by a retrieval system drawing on years of indexed content — including the very crisis coverage the team worked so hard to manage in 2014.
If a crisis is twelve years old and the retrieval systems still surface it — the crisis is not over. It is permanent infrastructure inside the AI engines until the brand actively rebuilds its presence in their answers.
The new crisis playbook
Four moves, all required:
— Earn the Tier-1 coverage. The 2014 work still matters. The LLMs weight Forbes, Fortune, The Wall Street Journal, The New York Times, and Harvard Business Review heavily. Earned media in those outlets becomes retrieval-anchor source material.
— Build structured owned authority. A schema-friendly corporate site. Clean executive bios. Original research. Investor communications. All of it readable by the AI engines as source material for the post-crisis narrative.
— Deploy GEO across the recovery cycle. Generative Engine Optimization measures whether the brand's post-crisis story is actually entering the answer when a buyer asks the AI engine about the company. If it is not — the recovery has not happened, regardless of what the analog metrics say.
— Measure Citation Share. The metric that matters in crisis recovery is no longer impressions or sentiment in a media-monitoring dashboard. It is how often the brand appears, and in what context, when an AI engine answers a buyer's question. Until that number moves, the crisis is still active.
The takeaway for every executive
A crisis in 2026 has a different half-life than a crisis in 2014. It does not fade with the news cycle. It is indexed permanently inside the retrieval systems and resurfaced every time a buyer asks the question.
Build the infrastructure before the crisis — not during it. If you are an executive at a Fortune 500 company and you are not actively measuring your brand's Citation Share inside the AI engines, you are flying blind. When the next crisis comes — and it will come — you will be doing 2014 communications in a 2026 market.
That is the GM lesson, updated. Not the one in the textbook.
Ronn Torossian is the founder and chairman of 5W AI Communications, the AI Communications Firm. He is the publisher of Everything-PR and the author of two best-selling editions of For Immediate Release.