How to spot outdated information before your audience does.
If communicators want to know how AI is shaping their organization’s reputation, they should be asking the same questions as their audiences.
Alex Sevigny, associate professor at McMaster University and an adviser at Ragan’s Center for AI Strategy, recommended frequent checks on how different generative AI platforms describe your organization, then compare those answers with what you know to be accurate.
“Make sure the model holds the most fundamental facts,” Sevigny said, during Ragan’s AI Certificate Course. “Who is your organization and what are its main business segments?”
Here are the five questions Sevigny says every comms team should test in LLMs.
- Who are we and what do we do?
Start with the basics. Ask ChatGPT to summarize your organization, its major business areas, products, leadership and headquarters, Sevigny said.
This will show whether AI has current information or if it’s relying on outdated sources, he said.
- What are our biggest strengths and weaknesses?
Ask the model to compare your organization with one or two competitors.
Sevigny suggested prompting the tool to “compare your org’s competitive strengths and weaknesses” with others in your industry.
See what differentiators AI associates with your brand and whether those match what you know to be true, he said.
- What complaints do customers have about us?
Then ask for the most common customer complaints and how the company addresses them.
This can help show any reputation issues communicators may underestimate or narratives that continue to circulate even after the company has addressed them.
- What controversies or scandals are associated with us?
Test “risky narratives and hallucinations” by asking what scandals are associated with the organization and what evidence is behind those claims.
“Look for outdated stories, incorrect information or claims that lack context,” Sevigny said.
- Why would someone want to work here?
Finally, ask why working at the organization would advance someone’s career.
Job seekers as well as customers use AI tools for their research, so this information should be clear, Sevigny said.
Then run the same questions across several AI platforms and write down where the answers differ.
With every question, Sevigny said teams should consider: Does this information reflect how the organization actually wants to be understood?
“Identify inconsistencies…then, document your findings and develop recommendations,” Sevigny said. “Document specific examples of misalignment. Include actionable recommendations.
The post Is AI getting your brand right? Ask these 5 questions. appeared first on PR Daily.
