Traffic down, panic up: Why AI is what SEO needs in 2025

If you’re reading this, congratulations.

It means you’ve survived approximately 47 Google algorithm updates, 12 existential crises about “the death of SEO,” and at least three moments where you seriously considered pivoting to llama farming.

But here’s the twist: 2025 isn’t the year SEO dies. It’s the year it finally grows up.

The truth is that we’ve heard this tune before when:

• In 2011 after Google’s Panda update penalized sites with low-quality content that was designed to manipulate rankings.
• In 2012 after Google’s Penguin update that targeted shady backlink practices.
• In 2013 after Google’s Hummingbird update that drastically boosted their ability to understand complex search queries, meaning you couldn’t just keyword your way into high rankings.
• In 2018 after Google’s Medic update popularized the now ubiquitous phrase Your Money Your Life (YMYL).
• In 2022 after Google’s Helpful Content update that obliterated sites with tons of shitty content and thin pages.

And now that we’re in the age of explaining to executive leadership that losing traffic = good, it’s time to break down why this is the best time for SEO to thrive.

First, search is alive and well

Let’s start with the big picture: search volume is exploding, not shrinking.

As of 2025, Google processes over 14 billion searches daily, a 22% increase from 2024. To put that in perspective, that’s roughly 2,000 searches per second.

AI overviews are only cannibalizing “quick fix” informational searches, while helping to create new ones that are more elaborate. They’re kind of like the potato chip of information: nobody stops at one. Users dive deeper, ask follow-up questions, and yes, still click through to websites.

And the queries are getting wilder. Thanks to AI’s ability to handle complex, multi-part questions, people are searching for things like:

• “How to sound smart at a dinner party using 17th-century naval metaphors”
• “Explain quantum physics using only emojis”
• “Why does my cat judge me?” (Spoiler: Because it can.)

So, you can put to bed the idea that users aren’t searching as much.

AI overviews are actually good... even though they steal your traffic

Overall organic traffic from Google is down for many sites. AI Overviews, snippets, and a smorgasbord of SERP features have made the competition for clicks fiercer than ever.

But!

Click-through rates (CTR) for top organic results are actually up while AI overviews are actually increasing traffic and clicks to homepages.

In fact, in 2025, the #1 organic result on Google now enjoys a 39.8% CTR, up from 39.6% last year. The second position saw a similar bump, from 18.4% to 18.7%.

And when an AI Overview is present, the links within those summaries are pulling CTRs that rival the top 3 - 5 organic results. In other words, the clicks that do happen are more intentional, and therefore, valuable.

Why does this matter? Because while casual, low-intent traffic may be down, the users who do click through are more likely to be genuinely interested, engaged, and ready to convert. It’s quality over quantity and for brands focused on meaningful engagement (and not just vanity metrics), that’s a major upgrade.

AI content has lowered the bar – now jump over it

AI content creation has basically ensured that the entirety of the internet is enshrouded in a fog of baseline mediocrity.

In other words, it’s like they all hired the same shitty writer who only knows dated corporate speak, emdashes, and labored metaphors.

That means the bar is low. With 83% of marketers now using AI for content, the internet is drowning in perfectly serviceable, yet utterly forgettable articles. The secret is prompt engineering (which is what any good editorial director was already doing).

Except now, prompt engineering is like the new SEO: Tools like Gemini want content that’s:

• Structured
• Authoritative
• Conversational

On top of that, LLMs prioritize content that is;

• Deep: Offer insights that AI can’t synthesize from existing sources.
• Human: Inject personality, anecdotes, or Hot Takes™ that algorithms can’t replicate
• Structured: Use clear headers, data tables, and FAQs so AI can easily crawl your content

So, if your medical advice isn’t backed by peer-reviewed studies or your financial tips are sourced by “trust me, bro”, you’re going to feel the hit even more.

AI content attribution and content ethics

Want your content to train an LLM and never get credit for it? Didn’t think so. Protocols like llms.txt are early steps in giving content teams a voice in what gets scraped.

Remember, you’re still optimizing for the user

Going back to the days of getting absolutely hammered on your lunch break, marketing executives had one goal in mind: get as many people talking about their brand in as many relevant places as possible.

And guess what? That’s still the goal (not the getting hammered part).

The more brands get talked about, the more they’re mentioned, and the stronger their reputation, the more they’ll show up in LLM search and AI overviews because these are all strong intent signals of a brand that provides a meaningful user experience.

If this is a problem you’re also looking to solve, start by focusing on:

• Experience, Expertise, Authoritativeness, Trustworthiness (EEAT): Google (and even LLMs) now determine if you’re AI overview material based on brand reputation and trust worthiness.
• Brand presence: Brand reputation is more important than ever: You no longer need backlinks to prove authority. If you’re being mentioned in authoritative sources, AI will pick up those references and the likelihood of you showing up is that much higher.
• Technical SEO:
  • Core Web Vitals 2.0 ensured that slow sites = fast penalties.
  • AI systems are dependant on structured data implementation using schema markup to generate summaries. Google, in no uncertain terms, is suggesting users to provide explicit clues about the meaning of a page by including structured data.
  • AI overviews favor detailed, question-focused content over transactional phrases

Here’s your 2025 AI+SEO checklist:

• Track intent-rich traffic over session volume
• Optimize headlines and intros for LLM skimming
• Prioritize brand mentions and PR-style visibility
• Let AI handle the first draft—humanize the final one

The bottom line: Get clicks or die trying

The SEO apocalypse is already here. Except now, there’s no black box of Google’s algorithm to hide behind.

And that’s a gift – because it means clarity is finally winning. Because now, we’re prioritizing the metrics that actually matter; the people that actually went to your site.

And here’s the deal: Every Google update, from Panda to AI Overviews, has rewarded the same thing: content that serves humans first, algorithms second. So, if you were already...

• Using AI tools to find content gaps and technical issues, then fixing them with human ingenuity...
• Letting AI handle drafts, but infusing final copy with personal wit, stories, and vulnerability...
• And structuring content for AI parsing, but packing it with insights only a human could provide...

You’re not behind. You’re built for what’s next.

The marketers who win in 2025 won’t be the ones who chase traffic. They'll be the ones who earn trust.

Because clicks are optional. But credibility? That’s everything.