A year ago, we told you the story of the best new thing in corporate crisis management. It began as a casual challenge salted with heavy skepticism, and lead to the combination of guidance from Dr. Timothy Coombs (the inventor of crisis response theory) and some amazing programming to create an AI-driven crisis prediction tool.

A year later, Professor Coombs and I were back at the 2021 International Public Relations Research Conference discussing the results of a year’s worth of testing and refinement that has resulted in a fully functioning product, already in use by some organizations.


Training the algorithm


The company that developed the tool, CapeStart, created three separate but crucial databases to feed data to its AI algorithms:


  1. Global Events Database: All significant events referenced within 40 years of media data, including non-crisis events.
  2. Media Crisis Database: All PR crisis-related events within the Global Events Database.
  3. Crisis Ontology: All crisis-related topics, categories, entities in the lexicon, and metadata labels.


After rigorous testing, human review, further testing, and further refinement, the tool now accurately:


  • Performs early detection of crisis signals,
  • Diagnoses the type of crisis you’re facing,
  • Offers suggestions for the best way to respond, and
  • Automatically calculates the time it will take to return to neutral (as opposed to negative) coverage depending on which response you chose, essentially recommending the right strategy to diffuse the crisis. 

The tool gains improved insights through ongoing monitoring of stakeholder commentary. It features continuous learning by the AI algorithms to adjust itself based on the facts on the ground.


Your Crisis Situation Room questions answered—in real time


As a result, the tool can now accurately answer questions that communicators always face when they enter a Crisis Situation Room. Including:


  • Are we in a crisis?
  • What type of crisis is it?
    • A Victim Crisis—beyond the immediate control of the crisis victim, including natural disasters, workplace violence, and product tampering.
    • An Accidental Crisis—borne out of an accident, such as technical errors.
    • An Intentional Crisis—created by organizational misdeeds, management misconduct, and similar incidents.
  • What is the optimum response? The tool is programmed to predict impact from four categories of common responses:
    • Deny—Includes attacking the accuser, denying the crisis, and finding a scapegoat.
    • Diminish—Includes finding a justification or excuse for the crisis.
    • Rebuild—Includes providing compensation to mitigate the crisis.
    • Bolster—Includes presenting the company as a victim and ingratiating itself with stakeholders.
  • What effects will different responses have, and how much longer will the crisis last given that response?


It predicts how long your corporate crisis will last!


That last point is the most powerful element of the software, particularly useful when you are arguing with lawyers. It can show you how much longer the crisis will continue depending on what response you choose. So, when the legal beagles are advocating for the “no comment” route, you can show that it would prolong your crisis for X weeks, depending on your situation, and based on historical present and data.


The system also displays various metrics and predicted metrics based on the recommended response options. This allows the user to, for example, compare the expected difference between an aggressive Deny-style response or a more laid-back Rebuild approach. 



An example of what your AI crisis dashboard might look like, courtesy CapeStart.


Find out more about this exciting new technology here. ∞


Photo by Raychel Sanner on Unsplash.