Status: in development with design partners. This capability is part of the EchoDepth Signal narrative analysis programme. We are testing it with defence and government communications teams against one benefit: clearer communication and fewer misunderstandings. Ask about joining.
Why Crisis Statements Get Misread
Crisis statements are written quickly, by people who know the full picture, and checked by colleagues who know it too. Nobody in that room reads the draft the way a worried family member, a local community, an international partner or a hostile commentator will. Misreadings usually surface after publication, when correcting them costs time and credibility.
In defence and government, the cost is higher. An ambiguous line can raise anxiety, read as evasive, or hand a hostile actor a phrase to reframe. The same tactics EchoDepth flags in hostile messaging (urgency, scapegoating, implied threat) can appear unintentionally in your own wording.
How It Works
Load the draft
A holding statement, press release, briefing lines or Q&A. Several versions can be compared side by side.
Choose the audiences
Your team selects or defines simulated audience panels, for example service families, a local community, international partners, specialist press or hostile commentators.
Review the simulated responses
Each panel produces plausible readings, questions and objections. Every response links to the passage that prompted it, with a stated level of uncertainty.
Revise and compare
Edit the wording, run it again, and compare versions to see which issues have been resolved and whether any new ones have appeared.
What It Shows
- Likely misreadings: passages that different audiences may interpret differently from the intended meaning
- Trust signals: phrasing that may read as evasive, defensive or as shifting blame
- Unanswered questions: what each audience is likely to ask next that the statement does not address
- Reframing risk: lines a hostile actor could lift out of context or turn into a counter-narrative
- Version comparison: how each revision changes the issues raised
What It Is Not
Simulated audiences are a structured review aid, not a poll and not a forecast of public opinion. They help a communications team see its draft from outside the room. They do not replace audience research, legal review or the judgement of the people signing off the statement.
Panels are synthetic. They are not built from personal data about real individuals, and the tool does not monitor or profile members of the public. Like the rest of EchoDepth, it is analyst-led: it surfaces evidence and uncertainty, and people decide.
Who It Is For
- Defence and government press offices and strategic communications teams
- Military public affairs and media operations
- Emergency planning and resilience teams
- Critical national infrastructure operators and defence suppliers with crisis communications responsibilities
Deployment and Data Handling
Draft crisis communications are sensitive before release. Crisis communications testing is designed to run in the same deployment models as the rest of the platform, including fully on-premise with no external data transmission and UK data residency as standard. See AI governance and data processing for how text inputs are handled.
Related Capabilities
- Narrative and influence analysis: linguistic tactics and narrative change in hostile messaging
- EchoDepth Signal: forensic statement and linguistic analysis of written statements
- Analyst-led AI: evidence, uncertainty and human decision-making
Help shape crisis communications testing
We are working with a small number of defence and government communications teams to test this capability against real workflows before release. NDA available.