For the last 20 years, the internet’s incentive structure was simple.
You published content. Google ranked it. People clicked. Your site did the persuading.
That chain is breaking.
Today, a growing share of “research” is happening within AI responses, not on your website. Prospects ask a question, get an opinionated synthesis, and only sometimes follow the links. In many categories, the first impression is now formed before a visitor ever sees your homepage.
So, if you’re still treating visibility as “rankings + traffic,” you’re measuring the wrong outcome.
The new question is: When someone asks an AI tool about your category, do you show up, and are you described correctly? (I first saw this framed sharply in Stoica’s writing, but the implication goes well beyond any single article.)
This is not “SEO with a new coat of paint.” It’s a shift from optimizing for discovery to optimizing for selection.
And selection is driven by a different set of mechanics.
From traffic to trust, explaining the new funnel
In the classic funnel, your website is the moment of truth. The click is the gateway to conversion.
In the AI-mediated funnel, your website becomes supporting evidence. The “moment of truth” happens earlier, inside the answer itself.
That means brands now compete on two planes:
- Search visibility: can people find you?
- Answer visibility: Will AI systems include you in the response?
These are not the same contest.
Search visibility rewards pages that are relevant, authoritative, and technically sound.
Answer visibility rewards brands that are:
- easy to understand (clear positioning)
- easy to verify (explicit claims and proof)
- easy to extract (structured, quotable content)
- easy to corroborate (credible mentions beyond your site)
If you want your brand to be present in the “answer layer,” you need to build for all four.
Why do AI mentions feel random, and how do I make them less random?
Most teams start here: they type a few prompts into an AI tool, see whether they appear, and draw conclusions.
That’s like checking your SEO by Googling yourself once.
AI outputs vary due to phrasing, context windows, retrieval sources, and model behaviors. Your visibility can swing dramatically even when nothing on your site changes.
The goal is not to win one prompt. It’s to become consistently “retrievable.”
To do that, you need to understand what AI systems are actually doing in the background:
- They try to identify the topic and the user’s intent.
- They pull from sources they can access, parse, and trust.
- They favor explicit, comparative content because it reduces ambiguity.
- They triangulate against external signals (mentions, references, reputation footprints).
That last point is where most “content-only” strategies fail.
The four assets that make you citable
If you want thought leadership and visibility, stop thinking in blog posts and start thinking in assets. Specifically, build assets that AI systems can confidently summarize.
A Category Explanation Page That Does Not Sell
Most websites confuse “positioning” with “promotion.”
A category page should answer, in plain language:
- What is this thing?
- Who is it for?
- When does it fail?
- What should a buyer look for?
- How do you evaluate options?
This is leadership: teaching the market how to buy.
When your site becomes the best explanation of the category, you earn citations by reducing uncertainty.
What to do this week
- Create a single page titled like a buyer would search it: “What is [Category]?” or “How to choose [Category].”
- Put a direct, two-paragraph answer at the top.
- Add a decision framework below it (a checklist, a scoring guide, or an “if/then” tree).
- Add a “last updated” date, and update it quarterly.
Comparison Pages That Tell the Truth
If you only publish “why us” content, AI tools have nothing to work with besides your claims.
Comparison content is different. It forces specificity.
High-performing comparison pages:
- define the decision context (“If you’re a mid-market team with X constraint…”)
- lay out tradeoffs
- include a simple matrix
- include “best for” scenarios
The fastest way to become visible in AI answers is to publish the pages people are already asking for:
- “[Tool A] vs [Tool B] for [use case].”
- “Alternatives to [Tool] if you need [constraint].”
- “Best [category] for [industry].”
Important: The pages must be useful even if the reader doesn’t choose you. That is what makes them citable.
Proof Pages That Are Built for Extraction
Most case studies are written like magazine features: narrative, vague, and heavy on adjectives.
AI systems do better with proof that is structured:
- starting state
- intervention
- measurable outcome
- time window
- constraints
- what didn’t work
- what you learned
A case study should feel like a field report, not a press release.
Upgrade your next case study
Include a box at the top:
- Industry
- Use case
- Time-to-value
- Inputs required
- Outcome metrics
- Implementation risks
Then, tell the story underneath.
An off-site reputation footprint you can’t fake
This is where most teams misunderstand “LLM optimization.”
You can publish great content and still be invisible in AI answers if no credible source mentions you.
AI systems usually lean on:
- industry publications and newsletters
- community discussions
- review platforms
- podcasts and video transcripts
- documentation and technical references
- public Q&A and forum threads
You don’t control these spaces, which is why they are trusted.
Thought leadership is not only what you publish. It’s what the market repeats.
What to do this quarter
- Identify 10 places your buyers already consult for advice.
- Earn 1 meaningful mention per month through:
- contributed commentary
- guest interviews
- data-driven mini studies
- honest participation in discussions
- partner co-marketing that includes specifics
This is the part marketing teams call “PR.” It now directly impacts AI visibility.
The most overlooked risk: AI can misstate your brand
There’s a safety dimension here that leadership teams often miss.
If AI answers can become the first touchpoint, then they can also become the first point of confusion:
- incorrect pricing
- wrong positioning (“they’re an agency” vs “they’re software”)
- outdated features
- inaccurate compliance claims
- wrong contact details
So yes, we care about being included.
But we also care about being included correctly.
Simple hygiene that reduces misrepresentation
- Maintain a canonical “About” page with a crisp definition of what you do
- Maintain a canonical “Pricing / Packages” page, even if pricing is a range
- Keep your “Contact” details consistent across every platform
- publish a “Last updated” footer on key pages (pricing, comparisons, compliance)
This is not busywork. It’s brand protection.
Measurement: “Share of Answer,” Not “Share of Search”
If you want this to be an operating system, you need a measurement method that doesn’t depend on vibes.
Here’s a practical approach:
- Build a list of 50 prompts your buyers would ask.
- category basics
- “best” queries
- comparisons
- implementation questions
- “Is it worth it?” questions
- Run them monthly across the AI experiences your audience uses.
- Track:
- Are we mentioned? (yes/no)
- Are we cited? (yes/no)
- How are we framed? (accurate/neutral/wrong)
- Which competitors appear with us?
- Use the findings to decide what to build next:
- missing comparison pages?
- unclear positioning?
- too little proof?
- weak off-site footprint?
This makes “AI visibility” a real backlog, not just a buzzword.
What thought leadership looks like now
Thought leadership used to mean: publish a hot take, get likes, drive traffic.
Thought leadership now means: become the reference point the answer layer relies on.
That requires a different kind of discipline:
- clarity over cleverness
- proof over adjectives
- structure over storytelling (at least at the top)
- external credibility over internal confidence
If you do this well, you’ll notice a side effect: your SEO improves too. Because the same content that helps AI systems extract meaning also helps humans make decisions.
The endgame is simple:
When your buyer turns to the internet for guidance, your brand is part of that guidance.