AI Lessons Learned: Starting Your AI Optimization Journey

There’s no question that large language models (LLMs) are reshaping digital marketing and PR. Industry forecasts suggest visitors from AI search could surpass traditional search traffic by 2028. Just when we thought we had SEO figured out, the rules shifted.

The good news? The foundational elements of PR matter more than ever. As with traditional SEO, generative AI engines prioritize authority, consistency, relevance and clear, quotable insights.

The bad news? AI search algorithms are evolving so quickly that there’s no single formula to “win” AI optimization.

Experimentation in the Wild West of AI

AI optimization is a moving target. Semrush’s AI Visibility Index found that over three months, ChatGPT’s model evolved to use 80% more sources — underscoring how dynamic LLM algorithms are and how critical diverse citations have become. The same report shows that while models like ChatGPT and Google AI Mode agree on 67% of mentioned brands, only 30% of their sources align. Even when the goal is similar, each model plays by different rules. There is no silver bullet for answer engine optimization (AEO) or generative engine optimization (GEO).

While we don’t have all the answers, we also don’t have time to wait. Brands have less than two years to establish authority in AI search results. Now is the time for PR pros to experiment with applying A/B testing principles to AI visibility. In fact, 63% of PR pros say their workplace encourages generative AI experimentation. Organizations willing to test and adapt will move faster than those waiting for certainty.

BEYOND CONVERSIONS: BROADENING THE MARKETING FUNNEL

The first step in prioritizing AI optimization is a mindset shift. As PR pros, we’ve been trained to see the marketing funnel and PR conversions attributed to PR as the holy grail.

While measuring the ROI of PR efforts remains critical, we need to broaden our view beyond the traditional three layers of the marketing funnel. In the new, AI-driven model, a lot happens before a click. AI becomes a stakeholder, synthesizing the entire gamut of your content into a snap judgment about how to characterize your brand. That means there’s an added layer of interpretation before your audience can even reach the “awareness” stage of the funnel. So, before we focus on outcomes, we need to focus on perception.

THREE WAYS TO START EXPERIMENTING WITH AI OPTIMIZATION

1. Establish a baseline. As with any experiment, you need a starting point. We’ve begun pulling referral traffic from LLMs in Google Analytics for clients — first as a benchmark, then as a quarterly KPI. Within a quarter or two, we’ve seen measurable increases in LLM referrals, signaling that integrated marketing efforts can influence AI visibility.

• • Reverse engineering AI-cited content: Map pitch angles to the queries with the top AI search volume in each category. If a top HR query is, “How can organizations reduce bias in leadership selection and promotion?”, we’d conduct media outreach on this topic and aim to place a contributed article with a similar headline.
    
    
  • Boosting narrative control: Particularly for companies that put out consistent proprietary data, we’ve seen a high volume of AI citations, but often these citations are not branded and don’t contribute to brand sentiment. In this case, we could recommend a paid council membership, like Forbes Business Council. This gives the client greater narrative control, with ongoing, long-form content on priority topics on a high-ranking domain that feeds into LLMs.

To sum it up, I won’t pretend to know every secret to AI optimization. But we won’t win by maintaining the status quo. PR has a real opportunity to lead in AEO and GEO — and the only way to know what works is to experiment. Now is the time to test, learn and adapt.