crisis comms

Status: ready — not yet trained.

Traits

Ownership

Takes direct responsibility for the situation ↔ Deflects, uses passive voice, or avoids admitting fault

Whether the communication takes direct responsibility for the situation — "we caused this," "our system failed" — versus deflecting to external factors, passive voice, or the classic "we take this seriously" without actually owning the failure.

Scope Honesty

Honestly describes what's affected and to what extent ↔ Minimizes scope or omits impact details

Whether the communication honestly describes what's affected — which users, which services, how long, what data — versus minimizing with "a small number of users" when it's clearly widespread, or omitting scope entirely.

Next Steps Clarity

Clear on what's being done, when the next update comes, and what readers should do ↔ No indication of what happens next or when

Whether the communication tells the audience what happens next — when the next update comes, what's being done right now, what the reader should do in the meantime — versus ending with "we'll keep you posted" or silence on the path forward.

Update Cadence Commitment

Commits to a specific update schedule or cadence ↔ No commitment to when the next communication will arrive

Whether the communication commits to a specific communication rhythm — "next update in 30 minutes," "we will post hourly until resolved" — versus leaving the audience to wonder when or whether they'll hear anything again.

Platitude Absence

Every sentence carries substantive information ↔ Padded with empty corporate platitudes and stock phrases

Whether the communication is direct and substantive throughout — every sentence carries information — versus padding with empty corporate phrases like "we take security seriously," "your trust is important to us," or "we apologize for any inconvenience" that add no information.

About

Scores the quality of crisis communication.

What it measures

Whether a crisis communication — incident notification, breach disclosure, outage update, public apology — earns trust through substance rather than corporate reflexes. The model rewards messages that take direct ownership, honestly scope the blast radius, provide clear next steps for the audience, commit to an update cadence, and avoid empty platitudes. The canonical anti-pattern is "We take this seriously. A small number of users may have been affected. We apologize for any inconvenience."

Feed the full crisis communication as input text. Works for status page updates, incident emails, public blog posts, and social media statements.

Limitations

• Optimized for written crisis communication in English. Live verbal statements (transcripts) may fit less well due to filler and conversational structure.
• Does not evaluate whether the stated facts are true or whether the response was timely — only the communication quality itself.
• Very early "we're aware and investigating" updates will score lower on some traits (scope, next steps) because there's genuinely less to say — that's expected.

Pairs well with

• u22a8.postmortem-ref — the post-incident review that follows the crisis
• u22a8.customer-support-response — individual replies to affected customers during the incident

Docs

• Tiers — categorical labels (Strong, Solid, Developing, Weak) assigned per trait
• Breaks — the per-trait trained boundaries between tiers