Jeremy Greenberg
Jeremy Greenberg

Every PR professional faces the challenge of delivering the right message to the right audience at the right time. During a crisis, the right time is often “yesterday” or “last week.” For PR teams that often means facing a slew of unattractive choices: trust your gut with messaging, borrow messages from the last crisis, or lose the news cycle in order to test your response. All this may be changing, however, thanks to the promise of AI agents––intelligent, simulated personas that can offer valuable feedback on messaging in real time. These tools can show how PR teams approach crisis management and message testing.

The Rise of AI-Generated Personas

AI-generated personas, powered by advancements in natural language processing, are providing new possibilities for organizations seeking to refine messaging with speed and precision. By simulating diverse viewpoints and reactions, these agents provide a safe environment to test ideas before they go public. For example, political affairs teams can gauge how a skeptical policymaker might respond to a proposed statement or assess the concerns of a constituent group impacted by a policy decision. AI agents offer insights that complement traditional methods, helping refine messaging for maximum effectiveness.

“Being able to test messages in real-time with all sorts of different audiences can be a gamechanger for public affairs messaging, especially during a crisis situation," says Steve Rabin, former speechwriter for NASA and communications advisor in the Obama Administration. “No one wants to have to rely solely on trusting their gut when the stakes are high.”

Why AI Agents are Gaining Ground in PR

Traditional approaches like focus groups and polling have long been staples of message testing. However, these methods face growing challenges:

  • Confidentiality: Sensitive or crisis-related messaging can’t always be shared widely for feedback.
  • Timeliness: In today’s rapid news cycles, delays in obtaining feedback can hinder effective response.
  • Audience Complexity: Messages must resonate across increasingly diverse demographics and psychographics. Screening for these specific audiences is both time consuming and expensive.

AI agents address these challenges by offering an alternative that is faster, more secure, and capable of reflecting varied perspectives. By simulating audience reactions within minutes, PR teams can adapt their strategies in real-time, ensuring their messages are both effective and appropriate for the moment.

What about accuracy?

Ensuring the accuracy and reliability of AI-generated personas is crucial for their effective application in crisis communication. Recent research has highlighted both the potential and limitations of current AI models in this domain.

A study published in Nature Communications examined the performance and biases of large language models (LLMs) in public opinion simulations, revealing that while LLMs can simulate human-like responses, their accuracy varies significantly across different cultural and linguistic contexts.

Further research from the Massachusetts Institute of Technology (MIT) investigated the ability of LLMs to generate consistent personas. The findings indicated that while models like GPT-3.5 and GPT-4 can tailor their responses to match assigned personalities, there are notable variations in the quality and consistency of these generated personas.

Despite these advancements, experts caution against relying solely on standard LLMs for generating custom personas.

Due to tight context windows and older training data standard LLMs like ChatGPT are not equipped out of the box to handle the nuanced requirements of generating a high volume of custom personas.

However, with targeted adaptation and specialized training, AI systems can provide improved accuracy and context-aware feedback.

The Bigger Picture

The adoption of AI tools for PR and public affairs aligns with broader shifts in the industry. As audiences become more fragmented and communication channels proliferate, PR professionals are turning to technology to manage complexity. The ability to quickly test and iterate on messages ensures that teams are better equipped to navigate crises with clarity and confidence.

Looking Ahead

As the use of AI agents becomes more widespread, it’s clear that these tools are not just a passing trend but a meaningful evolution in the field of PR. By combining human expertise with machine intelligence, PR professionals are better equipped to meet the challenges of today’s fast-moving communication landscape.

While AI will never be able to truly replace talking to humans, PR professionals are seeing the value in leveraging this technology in scenarios where cost, speed, and confidentiality are essential.

***

Jeremy Greenberg is the founder of market research firm Crowdwave.