The New Physics of Digital PR: How B2B Brands Get Cited in the AI Search Era

September 08, 2026

By Lee Odden Note from editor: Today’s post covers an increasingly important topic in the B2B marketing mix: the role of PR in driving AI search visibility.  I am happy to welcome this guest post from Dakota Shane Nunley, Director of Content Strategy at Product.ai who is on the front line of turning earned media into AI search visibility and shares some important lessons for all B2B marketers under pressure to deliver AI search performance. I am also happy to say that it lines up well with our Best Answer Marketing framework. 
For B2B marketing teams, digital PR spent the past decade perpetually on the chopping block, and as with everything in tech, we’re now hearing the same words uttered every time we open LinkedIn: AI changed it. This time it’s true. Answer engines are grounded on exactly the assets digital PR has always produced, the authority publications, the expert quotes, the third-party mentions, the original data.

I run the Authority Program (AEO/SEO and digital PR) inside Product.ai (formerly Demand.io). Here are the new physics of digital PR, how to get cited in the age of AI, and the most common myths I see circulating your feeds today.
What Are the New Physics of Digital PR in the Age of AI?
What actually changed is the physics of the machines we’re optimizing for. For years, digital PR was about driving authority via traditional search engines, namely Google, which is an army of crawlers that ranks pages. Today, AEO is optimizing for LLMs, pattern-recognition machines that cite passages.
So why would that matter? Well, SEO has the luxury of pages and links to distinguish entities, LLMs don’t. An answer engine runs on an accumulated sense of which claims keep showing up next to your name, and if it can’t resolve you into a real, distinct entity, you’re out of the answer, and no ranking report will ever tell you it happened.
How Does My B2B Brand Get Cited in AI?
Chase the Canonical Stat
As a B2B marketer, you should be chasing what I’ve started referring to as the “canonical stat,” a piece of original research, usually in the form of a fact or figure, adopted as canon in your category. LLMs cite it, journalists quote it, and industry roundups are built on it. It’s the holy grail of these new physics, because the companies that own the canon own the LLM answers.
So, how do you get there? Start by productizing your data. Somewhere in your company is the survey nobody has fielded, the experiment nobody has run, the dataset nobody outside your walls has seen. Then design the research backwards from the answer gap.
Before we fielded ours at Product.ai, I audited which studies the models already treated as canonical, found the questions they had no source for, and built our survey to address those gaps while fulfilling our curiosity and questions we wanted answers to. Since then, that AI shopping study has become a marquee citation inside the answer engines faster than any other activation we’ve run, with over a hundred …read more

Source:: Top Rank Blog