Sponsored by

Search used to send users down a funnel. Now it gives them one answer. If it’s not yours, you’re invisible.

That shift isn’t theoretical—it started more than a year ago.

In 2023, Google rolled out its Search Generative Experience (SGE), and the ripple effects were immediate. Combined with the rise of AI-powered assistants like ChatGPT, Gemini, and Perplexity, traditional organic traffic has started to erode.

Industry averages showed a 6% decline in clicks due to SGE, according to data published less than a year later. But the impact wasn’t evenly distributed:

  • Travel sites saw a 9.8% drop

  • Local search experienced a 14.9% reduction in clicks

The implications are clear: AI doesn’t deliver a list of options—it delivers an answer. And if your content isn’t the answer, it’s not even part of the decision.

Some SEO experts and agencies have observed organic traffic drops of up to 40% for certain keywords, especially where SGE provides comprehensive answers directly on the search results page.

At Redfin, we obsessed over rankings—“homes for sale,” neighborhood pages trimmed for crawlability, bounce rates tracked to decimal points. At Movoto, we scaled that precision across thousands of ZIP codes. At UpNest, the game changed. Visibility still mattered, but trust mattered more.

Who should I hire?
What’s a fair commission?
How do I compare agents?

Each stop sharpened my SEO instincts. But that game is over.

Search Isn’t the Starting Point Anymore. Conversation Is.

People no longer search. They ask.

They query ChatGPT. Prompt Perplexity. Trust Siri or Gemini to cut through the noise and deliver a clean answer.

“What’s the best mortgage lender for first-time buyers in California?”
“Is now a good time to sell in San Mateo?”
“What are the hidden costs of fractional homeownership?”

No links. No scrolling. No click-through.

Just answers.

It used to work like this: you’d search “best mortgage lender in California.” You’d get 10 blue links, a few ads, maybe a map pack. You’d click, skim, compare rates, read reviews—the funnel in motion.

Now? Ask ChatGPT and you’ll get:

“For first-time homebuyers in California, Rocket Mortgage and Better.com are frequently rated highly for low down-payment options, digital-first experiences, and competitive closing costs. Local credit unions like Golden 1 may offer better terms for in-state residents.”

No site visit. No form fill. The user’s journey ends at the point of the answer.

Same with:

“Should I sell my house in Palo Alto right now?”
You get market trends, seasonality, and pricing—all summarized in seconds.

“What are the downsides of fractional homeownership?”
You get a clean list: liquidity issues, resale complexity, usage limits.

The funnel isn’t shrinking. It’s disappearing. If you read my recent article, I discussed in detail how Google influences and shapes consumer behavior.

And if your brand isn’t part of the answer, you don’t get a seat at the table.

What We Used to Optimize For: SEO and SEM

Back then, we had two levers:

  1. SEO: Long-form content. Crawlable structure. Intent-rich keywords.


“Real estate agents near me.”

“Closing cost calculator.”


The goal: earn Google’s favor.

  1. SEM: Bid smart. Capture demand in real time.
    Every click mapped. Every conversion is tracked.

**

Looking for unbiased, fact-based news? Join 1440 today.

Join over 4 million Americans who start their day with 1440 – your daily digest for unbiased, fact-centric news. From politics to sports, we cover it all by analyzing over 100 sources. Our concise, 5-minute read lands in your inbox each morning at no cost. Experience news without the noise; let 1440 help you make up your own mind. Sign up now and invite your friends and family to be part of the informed.

**

This builds pipelines.

Clicks became leads.

Leads became deals.

However, clicks no longer matter as much as they used to in the past.

Now We Optimize for AI: AEO and LLMO

Search is no longer about visibility. It’s about eligibility.
You’re either part of the answer or you're not in the conversation.

Answer Engine Optimization (AEO)

You’re not writing for people to read. You’re writing for machines to understand.

Siri. Alexa. Google’s AI Overviews. They don’t care about elegance. They care about clarity and structure.

What that looks like in practice:

Q&A blocks
Build your content around specific user questions.

Q: What are the closing costs for a home in Austin, TX?


A: Closing costs in Austin typically range from 2% to 5% of the purchase price. On a $500,000 home, that’s $10,000 to $25,000. Common fees include title insurance, appraisal, escrow, and lender charges.

Schema markup
Add structured data so AI can easily identify what each section contains.

Example, screenshot, 25V

Direct, complete answers
Don’t bury the value. Lead with it. The user shouldn't need to click through—your answer should be the end of their search.

Large Language Model Optimization (LLMO)

This is the more profound shift: you’re not writing just for people—you’re writing for the systems that speak for you.

To qualify:

Authoritative, source-backed content
LLMs favor citations. Quote data sources. If it’s your research, explain how you got it.

Real names and attribution
Content with bylines and bios performs better in AI-generated results. Expertise signals trust.

Contextual specificity


“Prices are rising” is not often quoted.

“Phoenix home prices rose 4.2% in April 2025,” might.

Machine-readable formatting
Use bullets, short paragraphs, and clear headers. LLMs love structure. Write like your content is meant to be quoted.

The Dual Mandate of SEO: Be Findable, Be Quotable

To go deeper, I reached out to a colleague of mine, Toshiyuki Yamoto, who led SEO strategy at Movoto.com—where we scaled one of the largest long-tail search architectures in PropTech. Toshi is not just an SEO expert for the last decade; he’s a systems thinker. He understands how search works today, and how it’s evolving to serve machine-driven discovery.

When I asked him how content strategy must adapt, he said:

“The best SEO content must now serve both users and machines—be found in search and trusted by AI.”

Toshiyuki Yamoto, Architect of Search Visibility Systems, RedFrasco (Japan)

That quote gets to the heart of the new landscape. Yes, large language models may still access indexed web content through search engines—but they only surface it when they deem it credible, authoritative, and contextually relevant.

It’s no longer enough for content to be technically optimized. It must also be structurally sound, semantically clear, and supported by evidence. This is the dual requirement: discoverability and trust.

One gatekeeper is Google’s crawler. The other is the LLM’s internal judgment of truth and utility.

If your content doesn’t satisfy both, it won’t rank or be cited.

I decided to keep digging.

From Visibility to Eligibility: AEO as a Feedback Loop

I recently spoke with Sam Hogan, founder of Split.dev, after his post—SEO is about ranking. AEO is about being the answer—sparked conversation across the SEO and AI communities. In it, Sam draws a clear contrast between how we’ve traditionally approached content visibility and how that mindset needs to evolve.

“SEO is about ranking. AEO is about being the answer.

Sam Hogan, Founder at Split.dev

As Sam puts it, SEO has always been about getting to the top of a results page. AEO, on the other hand, is about making sure your content is the result—ready to be used, not just discovered.

Sam is critical of how most current tools approach the problem. “They’ll tell you, ‘You’re at 22% AI visibility,’” he said, “and that’s where the insight ends.” But knowing the number doesn’t move the needle.

The real shift, he argues, is from reporting to execution—closing the gap between where your content stands today and what it needs to become to surface in LLM-generated answers tomorrow.

“The tools that matter won’t just diagnose,” Sam said. They’ll act.


They’ll tell you, ‘You’re at 22%’.

‘Here’s how we get you to 45%’,


and then they’ll do the work to get you there.

That’s precisely what he and his team are building at Split.dev—a system designed to function as a continuous feedback loop, not a static report. AEO, in Sam’s view, isn’t a dashboard. It’s a living process: analyze, adapt, and improve. Learn what qualifies. Identify what disqualifies. Iterate.

In his words: “In a world where LLMs are the interface, every sentence you publish either increases your eligibility—or quietly disqualifies you.”

And the real takeaway? “The winners in this space,” Sam told me, “won’t be the ones reporting the score. They’ll be the ones raising it.”

From what I am seeing, he’s right. Visibility alone no longer creates leverage. In a system governed by AI, what matters is whether your content can be used. AEO isn’t just the next evolution of SEO—it’s a shift from publishing for clicks to publishing for inclusion in the answer itself.

25V Research, 2025

Where This Is Going

Every company I’ve been at—from Redfin to UpNest—was built to be found.

But the future isn’t about being found.


It’s about being used…..

Used by AI. Quoted by machines. Trusted by both humans and algorithms.

We have to stop optimizing for traffic.
We must optimize for truth, trust, and transmissibility.

Because in the next era of PropTech and FinTech, the winners won’t be the ones with the most visits.

They’ll be the ones with the most citations.

xoxo,

Maximillian Diez,

GP, Twenty Five Ventures

P.S. Stay with me on this journey. 

Keep Reading

No posts found