This episode explores Generative Engine Optimization (GEO) and how LLMs are making an impact on the Public Relations industry. Fred, Michael, and Lindsay discuss three years of joint research on AI in communications, looking at why GEO matters now, how LLMs source information, and what communicators must do to monitor and shape their brand’s presence in AI tools.

Key Ideas:

1. Three Years of AI & PR Research

• Year 1 – “Fascinated and Frightened”
  • Conducted shortly after ChatGPT’s launch.
  • Communicators were both excited and unnerved by credible, fast AI-generated writing.
• Year 2 – Hands-On Adoption & Culture Impact
  • Teams heavily using AI were 90% more likely to feel valued in their work.
  • Clear culture mandate: integrating AI into workflows boosts engagement and perceived value.
• Year 3 – Focus on Geo (Generative Engine Optimization)
  • In a world of “ask an AI, not Google,” LLMs become a new gatekeeper.
  • Communicators must monitor LLM outputs the way they already monitor social and earned media.

2. What Is Geo and Why It Matters

• Geo = Generative Engine Optimization – optimizing how your brand appears in AI-generated responses.
• 62% of communicators say Geo is the biggest opportunity since search, but nearly half don’t know where to start.
• First step: education and monitoring
  • Ask LLMs about your brand, sector, and competitors.
  • Use customer language and repeat prompts to see answer variance.
  • Identify which sources are being cited most often.

3. LLMs as a New Corporate Spokesperson

• AI is increasingly the first and sometimes only “voice” people hear about a brand.
• Owned and earned content are now critical inputs:
  • Recent research (e.g., MuckRock) shows that for “recent” queries, nearly half of LLM sources are authoritative earned news.
  • Brand blogs, press releases, and wire-distributed news releases have renewed authority in AI responses.

4. Misconceptions About AI in PR

• Fear that “everything will be outsourced to bots” and humans become mere prompters or reviewers.
• The research suggests the opposite:
  • The “human in the loop” is more important than ever, especially for strategy, ethics, nuance, and relationship-building.
  • AI can surface a roadmap that still leads to deeply human interventions (e.g., local officials, in-person engagement).

5. How LLMs Source Information

• Frontier models now typically have access to the open internet, so their data sources overlap more.
• Copilot has strong adoption in enterprises due to integration with Microsoft 365 and perceived security.
• Outlets with good SEO and page rank often also have outsized authority in LLM answers.
• Licensing deals (e.g., with major publishers) can shift which outlets LLMs favor, changing which journalists and outlets matter most for PR.

6. SEO vs Geo

• SEO:
  • Traditionally more in the marketing domain.
  • Focused on ranking web pages in search results.
• Geo:
  • Deeply tied to communications, because LLMs lean on owned and earned content.
  • A major opportunity for PR teams to lead and own this new discipline.
• Important: Don’t abandon SEO. Search rank still influences AI summaries and zero-click results.

7. Zero-Click Search and AI Summaries

• Gartner forecast: search volume could drop ~25% by 2026 due to AI summaries.
• Users get answers from AI boxes at the top of search without clicking through:
  • Less traffic to owned and earned properties.
  • Higher stakes for what the AI summary says and which sources it uses.

8. Trust, Misinformation, and Hallucinations

• Hallucinations remain common; humans are ultimately responsible for accuracy and must fact-check.
• Surveyed communicators (majorities in high 50s/low 60s) are concerned that:
  • They have limited control over how AI represents them.
  • AI may misrepresent their brand or values or share outdated content.
  • AI could amplify negative narratives and misinformation.
  • Stakeholders may trust AI answers over official corporate communications.
• Younger communicators are more likely to view AI outputs as high-authority sources, comparable to or above trusted news outlets.

9. Correcting Errors in LLMs

• You can’t directly “edit” an LLM, but you can:
  • Trace which sources it is citing.
  • Correct those sources via traditional PR methods (contacting reporters, correcting articles).
  • Publish better owned content and ensure it’s discoverable (e.g., wires, authoritative formats).
• Because LLMs increasingly pull from live web content, correcting upstream sources can change downstream AI answers.

10. Crisis Communications in the Age of AI

• LLMs are another active voice in every crisis.
  
   
• Monitoring is crucial to know:
  
  
  • What AI is saying about the issue.
  • Which sources are influencing that narrative.
• DoorDash case study:
  
  
  • Viral Reddit post falsely claimed exploitative internal algorithms at a generic delivery company.
  • Even before the fraud was fully exposed, DoorDash’s CEO quickly said “this is not us,” creating distance from the allegation.
  • This is a shift from old crisis playbooks, where brands might wait until accusations were more specific or confirmed as relevant.
• 43% of communicators don’t know whose job it is to handle AI-related reputation and crisis issues:
  
  
  • Opportunity for comms teams to step up as conveners and leaders across legal, IT, marketing, and CX.

11. Tools and Measurement: AI Share of Voice

• Geo Compass (We Communications tool):
  • Daily queries to leading LLMs about a brand and its category.
  • Measures AI share of voice, tracks how often brands appear, and which sources shape those answers.
• Adds a new layer alongside traditional KPIs:
  • Not just “What coverage did we get?”
  • Also “How did this coverage and content change what LLMs now say about us?”

Production Credits

A production of the USC Annenberg Center for Public Relations at the University of Southern California.

Host: Fred Cook
Executive Producer: Ron Antonette

Season 7 Producers: Joe Carreon and Anvi Mahajan
Production: Camille Culbertson, Jack Gisler, Toma Battino
Editorial: Joey Cha, Ivan Feng, Natalie Lopez, Grace An
Social Content: Angelina Tran, Hailey Evans, Arushi Purkayastha, Chris Apy
Growth: Van Luu, Shaan Dhaliwal

Links

Follow the USC Center for PR (@usccenterforpr) on Instagram and LinkedIn. 
Follow Fred Cook on LinkedIn. 
Find all our reports at annenberg.usc.edu/cpr.

Download the 2025 Relevance Report at

 annenberg.usc.edu/relevance