Hi there,

How to influence LLMs and win in Search. That's the topic of my deep dive today.

I also discuss the usefulness of AI tracking tools, and show you an excellent resource on thinking like a world-class marketer.

As always, thanks for being here.

I'll be back on February 19.
Nicole

P.S. If you'd like more of these deep dives, please hit reply with a 'YES.'

I want to share your:

Big news, latest resource, or outstanding results.
I'd love to support you by featuring you here. It's FREE!
​
Smash the button and show me what you've got.

Insights

​

How To Influence LLMs and Win in AI Search

Last week, I posted a slightly provocative statement on LinkedIn: "There is currently no sure way to influence LLM results, which means that LLM tracking tools are basically useless."

As expected, that statement resulted in a wave of replies from people who disagreed—and several thoughtful comments from others.
​

As someone pointed out, "it's way more nuanced."

They're right. The problem isn't the tools. It's how marketers expect deterministic answers from probabilistic systems. While there are no guarantees, there are structural advantages that materially shift probability through scale, authority, and repetition.

Let's examine how brands create repeatable advantage patterns that compound over time.

How brands build AI visibility

1. Digital PR is now the primary upstream input

Brands with larger budgets place themselves in more visible, more frequently indexed environments through "best of" lists, cross-collaborations, and sponsored posts on high-authority publications.

This matters because repeated exposure in credible publications is how brands enter the set of entities AI systems consider plausible in the first place.

Digital PR is now a primary upstream input into AI visibility.

If an LLM already has significant exposure to a brand because it appears consistently in credible contexts, that brand is more likely to be mentioned when relevant. LLMs don't discover brands in the moment—they remember them from accumulated exposure and associations built over time.

Digital PR drives brand mentions, co-occurrence, and credibility. It feeds the environments LLMs already trust and retrieve from, expanding the probability space in which a brand is even eligible to appear.

Larger brands don't just have more visibility—they occupy more of the training and retrieval surface area. This creates a compounding advantage smaller brands must counter with creativity, narrative hooks, and earned media.

​
​2. Source authority matters

AI systems don't treat all sources equally.

Some domains are more trusted, more frequently referenced, and more likely to be retrieved. An article in the New York Times carries more weight than a forum post, though community content still plays a supporting role for niche or experiential queries.

​
3. AI-specific content: Proceed with caution

(Note on the image: I predict that one of Google's next updates will address these AI-generated article dumps).

When AI systems have limited prior exposure to a brand, some companies experiment with AI-specific pages written primarily for ChatGPT and similar tools. Results are inconsistent and highly dependent on the system being tested.

Models evolve, retrieval methods change, and filters tighten. Practices that look clever today often become tomorrow's cautionary screenshots. History suggests platforms eventually adjust to tactics that prioritize manipulation over usefulness.

What actually drives AI visibility is not any single tactic, but a compounding loop: Credible mentions increase recognition. Recognition expands eligibility. Eligibility raises the likelihood of citation, and repeated citation strengthens future retrieval.

​
4. Authoritative on-site content is non-optional

LLMs perform best with sources they can easily parse, summarize, and trust: clean site structure, plain language, clear definitions, and content that answers questions directly.

This is also what good writing looks like for humans.

On-site content is not just "good for humans." Authoritative on-site content acts as the canonical reference layer AI systems summarize, compare, and validate against. Without it, brand mentions lack context and AI systems default to safer, better-explained competitors.

Digital PR creates awareness. On-site content converts awareness into comprehension and trust.
​

5. Freshness helps, but isn't everything

Fresh content can help in some AI retrieval contexts, particularly when recency is relevant to the query. However, freshness alone is not a universal ranking factor. Evergreen, authoritative content often performs just as well for foundational topics.

Some AI systems supplement their knowledge with live or recently indexed content, while others rely more heavily on existing training data. Regularly updating high-quality content helps ensure accuracy and relevance for both humans and machines.
​

The relationship between prompts and AI visibility

​SparkToro's research highlights how inconsistent AI systems are when recommending brands or products. That inconsistency isn't a flaw to be fixed. It's a feature of probabilistic systems working with incomplete information.

This is why there's no "sure way" to appear in AI-generated answers. There are no guarantees, but there are structural advantages that materially shift prior probability.

The growing list of acronyms—GEO, AEO, etc.—may suggest something fundamentally new, but they share the same foundations. Improving visibility in traditional search still increases the likelihood of being surfaced in AI-generated responses, because many AI systems are built on the same discovery and trust signals.

This aligns with what Google has said for years. John Mueller has noted that there is no meaningful optimization for AI without solid SEO fundamentals.

How prompting works

If a user asks a vague question, they get a generic answer with safe, well-known names. If they ask a tightly scoped, context-rich question, they'll often surface more specific brands, niche players, or newer entrants.

Prompting influences outcomes, but it doesn't override a model's internal confidence boundaries. You cannot reliably prompt an LLM into citing a brand it doesn't already recognize as relevant or credible.

What prompting does change is which part of the model's knowledge gets activated. Constraints like "specialist," "B2B-focused," or "less obvious" widen the citation set, but only within the limits of brands the system already considers plausible.

The implication is important: Brands aren't competing for a single AI ranking. They're competing across millions of slightly different prompts written by real people with wildly different levels of clarity and intent.

The indirect influence on prompts

While prompting is ultimately user-driven, marketers still influence the shape of prompts indirectly by shaping category language, normalizing comparisons, and publishing "how to choose" frameworks that buyers reuse.

This is different from manipulation, but it does affect how questions get asked. This is why user education matters alongside brand optimization. When users know how to express their needs more precisely, the answer space expands.
​

Are LLM tracking tools useful?

AI tracking tools can offer directional insight, but only if you're clear about what they can and cannot measure.

The mental model needs to shift from "Where do we rank?" to:

• How often do we appear?
• In what contexts?
• Alongside which competitors?

SparkToro's research suggests that Visibility %—the frequency with which a brand appears across dozens or hundreds of model runs—is a reasonable proxy for prominence. That's very different from a ranking, and it only works when the testing methodology is consistent, repeatable, and transparent.

If you're evaluating AI tracking tools, the real question is whether the tool can reliably generate this kind of aggregate visibility signal. Many cannot. Small sample sizes, opaque prompts, and single-run outputs produce noise that looks like insight.

Be cautious of any platform that claims to show definitive "AI rankings." AI systems don't operate on a single, stable ranking system, and treating them as if they do creates a false sense of precision.

Instead, pair any AI visibility data with leading indicators you already trust: branded search demand, direct traffic, repeat visits, and whether prospects mention discovering you through AI tools in sales calls or surveys.

Used longitudinally, with consistent prompts and paired with PR signals, tracking tools can still inform strategy. The problem isn't the tools themselves, but the expectation of deterministic answers from probabilistic systems.

Imperfect signals are far more useful than precise-looking numbers built on shaky assumptions.

​

Related: Check this section from an article by Gaetano DiNardi: SEO ROI is now the downstream outcome of brand traction.

Do you follow me on LinkedIn? I share regular tips and stories I don't have room for here. Come and join me​.

Information

✶ Mt. AI

This image is a perfect example of what I highlighted in my last newsletter: The Overlooked Risks in AI Content Plans.

In that piece, I reacted to a CMO Investment Report by Conductor showing that marketing leaders are prioritizing "AI content generation at scale:"

"CMOs betting on AI velocity are assuming that future algorithm updates will be neutral or forgiving. History suggests the opposite. Google tends to let a behavior spread before they aggressively correct it."

​Glenn Gabe coined this "Mt. AI." That's clever. It's also a reminder that content intended to manipulate search rankings will eventually be punished.

✶ What if user satisfaction is the most important factor in SEO?

Marie Haynes invites us to reconsider a central idea in SEO: what if user satisfaction really is the most important factor Google considers when ranking pages?

She bases much of her thinking on insights revealed during the US Department of Justice vs. Google trial, where internal details about Google’s ranking systems and how they use user data were discussed.

Google uses real user interaction data to teach AI systems what helpful content looks like. That means your success in SEO increasingly depends on being the page that genuinely satisfies the searcher’s intent.

Inspiration

Think Like a World-Class Marketer

​
Rory Sutherland has spent almost four decades at Ogilvy studying why people behave the way they do and how to change that behavior. The lessons from his interview—tiny and big—will improve your marketing in 2026.

2 big lessons

Rationality isn’t enough.

To really win, you have to go beyond logic. If you only do what makes sense, you’re competing with everyone else who has the same data. The magic happens when you do the things that look wrong on a spreadsheet.

Fix the feeling, not the tech.

The best way to solve a problem is often to change how it feels, not how it works. A wait feels shorter if you’re entertained, even if the time is the same. Psychology is cheaper than engineering.

​
That's all for today. Thanks for being here!
I'll see you again on February 19.

In the meantime, feel free to ask me a question, send an interesting link, or tell me what's on your mind. I read all your emails!

Dates to watch

March Monthly Observances

• Women’s History Month
• Nutrition Month
• Music in Our Schools Month
• National Craft Month
• American Red Cross Month
• Irish-American Heritage Month
• Ramadan (projected to end on March 18-19)

Weekly Observances

• March 9-15 – Girl Scout Week
• March 9-15 – National Sleep Awareness Week
• March 18-24 – National Agriculture Week
• March 23-29 – National Cleaning Week

Days

• March 1 – Zero Discrimination Day
• March 3 – World Wildlife Day
• March 3 – National Anthem Day
• March 4 – Holi Festival
• March 5 – World Book Day
• March 6 – Global Unplugging Day
• March 7 – Employee Appreciation Day
• March 8 – International Women’s Day
• March 8 – Daylight Saving Time
• March 13 – Purim
• March 13 – World Sleep Day
• March 14 – Pi Day
• March 15 – The Ides of March
• March 15 – 98^th Academy Awards Ceremony
• March 17 – St. Patrick’s Day
• March 18 – Global Recycling Day
• March 18-19 (expected) – Ramadan ends
• March 19-20 (expected) – Eid Al-Fitr
• March 20 – Nowruz
• March 20 – Spring Equinox
• March 22 – World Water Day
• March 26 – Epilepsy Awareness Day
• March 27 – World Theatre Day
• March 27 – MLB Opening Day
• March 29 – Palm Sunday.​

​

Keep in touch