Insights
Teaching AI: Brand Building and Digital PR
Mark Williams-Cook covered a bunch of topics during his appearance on the Market Movers podcast.
You all know I'm obsessed with brand building and Digital PR, so I paid attention to Mark's thoughts on those.
Here's where Mark and I agree:
- Brand perception is shifting - from what companies say about themselves (1950s), to what customers say online (2000s), to how AI models understand and represent your brand (2020s).
- Digital PR works on two levels: influencing the model's long-term "memory" (training data) and short-term retrieval (web searches the model performs).
- Quality mentions matter more. Being discussed on authoritative sites that can't be bought influences how models weight your relevance to topics.
- Brand-building beats manipulation. Create genuine value rather than chasing visibility tricks, asking "would this activity provide value if search didn't exist?"
My final thought
Digital PR is a priority because it allows you to leave imprints in the training data that AI models consume. When journalists and authoritative sites discuss your brand in context, you're teaching AI systems who you are and why you matter.
You'll enjoy more of Mark's observations in this 45-minute video: Forget GEO Hype.
Why every AI search study tells a different story
Why does almost every study about AI search seem to contradict the last one?
Each has data to back up their claims, but the stories they tell are different depending on who’s doing the analysis and what they measure.
People are hunting for a definitive answer when there probably isn't one.
How does AI search affect traffic and conversions?
The answer is... it depends (I know).
The impact of AI search depends entirely on your specific situation, which includes your industry, your audience, and your content strategy. What's true for an e-commerce site selling headsets might be completely wrong for a B2B software company.
Instead of looking at data produced by those with some self interest, we need to measure what's happening on our clients' (and our own) sites.
Track the AI referral traffic and monitor conversion rates. See if the quality matches the volume changes you're seeing.
The practical approach is to hold your conclusions lightly. The changes in Search are happening fast and rigid thinking based on someone else's study won't help you stay ahead.
Keep testing multiple approaches and stay curious about what the data tells you specifically.
What's one thing you've noticed with AI search traffic on your sites? Hit reply and tell me what's going on.
Related: You could turn some of these tips into a checklist: 7 success criteria for organic search.
GEO clean-ups
Last month, I wrote a short piece on DIY SEO.
Let me refresh your memory:
"We're about to see a wave of "DIY AI SEO gone wrong" stories.
Our expertise isn't going anywhere, and might be used to help brands harness AI tools without wrecking their website—or to clean up the mess when they do. Have you thought about this for your 2026 service offerings?"
Soon after, John Mueller commented on a Bluesky post in which he joked about SEOs offering GEO-Detox services:
This thread wrote itself!
Have you done any GEO clean-up work yet?
Reply and tell me what you had to fix. I'll mention a few examples in my next newsletter.
We may have passed "peak AIO"
Tom Capper suggests that the upward march toward more prominent AI Overviews (AIOs) is leveling off, or possibly in decline.
An analysis of 40,000 keywords tracked in STAT, shows 15% of AI Overviews (desktop) or 13% (mobile) appearing in position two or lower.
I noted this happening 6 months ago, and showed how several of my Google searches resulted in an AIO appearing below a website link—examples here and here.
We can only guess what Google is up to, but it's clear they continue to experiment with how results are displayed in Search.
Any decline in AIOs is somewhat good news for SEOs and marketers because it indicates that high-quality, well-optimized content still has a strong role to play.
I’d double down on delivering value and clarity, especially for complex B2B topics that AI is less likely to handle cleanly.
See all the numbers on "Whiteboard Friday."
Do you follow me on LinkedIn? I share regular tips and stories I don't have room for here. Come and join me.
