Your AI Tool Is Flying Blind Without This Crisis Doc

Chick-fil-A just had a rough one. Hackers used stolen passwords to break into customer loyalty accounts, potentially exposing names, emails, phone numbers, and payment information. Classic credential-stuffing nightmare. And if your organization stores customer data — which, let's be honest, almost everyone does — you're one bad password away from drafting a breach notice at 11 p.m. on a Tuesday.

Here's where it gets interesting for PR people: a lot of teams are now leaning on AI to help draft crisis responses faster. And that's not a bad instinct. Speed matters when reporters are calling. But according to Stephanie Nivinskus, principal at Ragan's Center for AI Strategy, there's a foundational problem most comms teams haven't solved yet.

Your AI tool doesn't know what your organization has already decided.

THE GAP NOBODY TALKS ABOUT

You probably have a message map. Maybe even a solid one. Approved language, designated spokespeople, escalation protocols — the works. But if that document lives in a folder your AI tool has never seen, it might as well be written in invisible ink.

Without access to your organization's existing decisions and context, AI defaults to generic. It pulls from learned patterns, not from the hard-won choices your team made during the last crisis exercise or live incident. It doesn't know which phrases legal shot down six months ago. It doesn't know the position you walked back after a bad media cycle. It starts from scratch every single time.

That's not a time-saver. That's a liability dressed up as efficiency.

WHAT FIXES IT

Nivinskus makes the case for building a crisis decision log — a structured, living record that your comms team and your approved AI tool can actually access when things go sideways. Not a folder of old PDFs. A current, organized document that captures:

• The decision that was made
• The rationale behind it
• Who approved it
• When it was made and under what conditions it applies

One person owns it. It gets updated after every crisis exercise or real event. And it goes wherever your AI tool goes when the next incident starts.

The payoff isn't just faster drafting — it's smarter drafting. Feed AI your decision log, your current message map, and a list of likely reporter questions, and it can flag contradictions before your spokesperson walks into a press briefing. It can catch when a draft answer conflicts with a position you already rejected. That review takes minutes. Catching it after the fact can take weeks of cleanup.

WHAT THIS MEANS FOR YOUR PR STRATEGY

AI in crisis comms isn't a plug-and-play solution. It's more like a very fast, very literal new team member who knows nothing about your organization's history unless you explicitly tell them. The teams that will actually get value from AI during a crisis are the ones doing the unglamorous prep work before anything breaks.

That means treating your crisis documentation as an AI input, not just an internal reference. It means assigning ownership so the log stays current. And it means accepting that a one-time setup isn't enough — this has to get updated every time your organization learns something new about how it handles pressure.

The Chick-fil-A situation is a useful reminder that crises don't schedule themselves around your readiness. Build the infrastructure now, while things are quiet, so your AI tool isn't burning the minutes you don't have trying to reverse-engineer decisions that already live in someone's head.

Speed without context is just noise with a deadline.

If you want help thinking through how AI fits into your crisis communications strategy — or how to build a PR function that actually holds up when things get hard — Winsome Marketing is worth a conversation. We work with brands to build comms strategies that are smart, ready, and grounded in reality.

This post was originally inspired by Stop AI from wasting time in your next crisis via prdaily. We encourage you to read the original piece for full context.