Some will call it speculation. Others will say it’s inevitable. But one thing you shouldn’t be is unprepared.

Imagine a future when your AI assistant does everything — OK, maybe not everything, but a lot — for you. It extends beyond the research and performs executive functions for your day-to-day tasks and more. Stuff like: 

• Planning and booking your vacation hotel
• Shopping for and buying clothes
• Checking out Italian restaurants in your city and reserving a table for dinner
• Researching and requesting demos from software companies

We’re already seeing it do these things in rudimentary ways. Google is just rolling out agentic shopping experiences within Gemini and Search.

The key to unlocking this new avenue for growth is Agent Experience (AX), or optimizing how AI agents use your website.

And while for many it feels like an afterthought, the blueprint for its future may be in search: the most popular use case for AI right now.

 

Zero-Click Search: Proof of Declining Human Website Traffic

It’s a fact: People check their sources less and less. A Pew Research report conducted in April 2025 found that 26% of Google users immediately left the site when they received an AI summary, compared to 16% of those who didn’t. That same report found that only 8% of the former users clicked on a search result, and only 1% clicked on a link from the AI summary.

Known as “zero-click search”, this phenomenon is pronounced nearly twofold when users are given the opportunity to interact with AI.

This is not a new topic, but how it relates to agentic AI is extremely important. If the user is changing, then how does that affect the way we optimize website content?

 

Search Engine Optimization vs. Answer Engine Optimization: What’s the Difference, Really?

Almost a year ago, I wrote that “A lot of the same things you were doing before generative AI rolled out still matter” when it came to your SEO strategy.

I don’t know how I feel about that anymore. Not because it was wrong, but because it was dismissive of the technical differences between AEO and SEO. 

How AI Search Interacts with Web Content

AI crawlers read content very differently from Google’s crawlbots:

• They prefer shorter text strings: AI agents prefer to process shorter amounts of text because it takes less processing power. 
• Easy questions get short answers. Complex questions get complex answers: Newer LLM models scale their processing based on the complexity of the request. This also saves processing time and power.
• They don’t view content hidden by JavaScript: Unlike Google’s crawlbots, most LLMs only review content in static HTML. This means that if any content is hidden by JavaScript animations or user actions, it’s likely to stay hidden.

But what about the creative differences? If we’re being honest, and you had an A-plus SEO game, then I’m not sure how much changes. I asked ChatGPT how to optimize for appearing in its mentions and citations, and how that differs from traditional SEO. Here’s how it responded: 

 

Aspect	SEO Focus	AEO Focus
Keyword Targeting	Optimize for high-value keywords and variations.	Optimize for questions, topics, and conversational phrasing (who, what, how, why).
Content Structure	Keyword-rich headers (H1–H3), readable text, internal links.	Semantic clarity and explicit Q&A formatting (FAQs, definitions, summaries) to help LLMs extract answers.
Meta Tags	Title and meta descriptions for SERP CTR.	Still useful, but AI systems rarely show them — instead, focus on clear summaries in-page.
Entity Connections	Implicit through keyword use.	Explicit through entity linking (e.g., using full names, locations, linked data, or mentions of related entities).
Author Attribution	Optional or for E-E-A-T improvement.	Crucial — AI models weigh verified authorship, credentials, and linked bios more heavily.
Tone & Readability	User-friendly, scannable.	Direct, concise, and fact-based — LLMs prefer declarative statements over marketing fluff.
Content Freshness	Updated for search recrawls.	Updated for data currency — AEO relies on the most recent, verifiable facts; stale info can get downweighted by AI summarizers.
Visuals	Images, infographics, and video to improve UX and engagement metrics.	Text versions or captions for every key insight (since AIs cannot “see” images without alt text or markup).

 

If you’re reading between the lines, the differences are minute. Since when has “clean text” not been valued in the SEO world? We’ve already known about long-tail keyword searches for years. Header structure has always been a cornerstone of good content design. What am I missing here?

 

Why Content Design Could Make or Break Agent Experience Practices

Will zero-click search or agentic supersede our standard web practices in the long term? It’s kind of a pedantic question. 

People will still buy new running shoes, reserve tables at restaurants, and request software demos. 

In either scenario, whether it’s search or outright agentic, we see people using some sort of AI to complete their tasks. 

The use case for search isn’t changing, but the user is. That user might just happen to be a robot moving forward. And bots aren’t usually looking for the most nuanced answer; they just want a good one, as fast as possible.

That’s not all that different from humans, actually. But the way you communicate that to an AI is drastically different from how you would to a human, or even a search engine.

Since LLMs are multimodal, they can interact with and produce a vast array of media. While SEO focuses largely on text, AEO may push us to further implement one of content design’s best practices: use the right medium for the target audience and the idea you’re trying to communicate.

So I find it very odd when people assert that AI or AEO is a content killer. If anything, it may be a content diversifier.

Now for the big question: Who might actually use AI agents moving forward?

 

Do Consumers Trust AI Agents with their Chores?

It’s a fair question. And it depends.

Trust is a major factor at the moment. Booking.com released a report earlier this year that showed only 12% of survey respondents would trust AI to make independent decisions, like planning a trip. That’s despite 91% being enthusiastic about the idea of it doing so. Adding to the complexity, 65% expected autonomous trip planning to go mainstream in the near future.

When it comes to purchases like clothing or household items, the numbers aren’t far off. A University of Virginia Darden School of Business report showed that almost half of consumers trust AI more than a friend in choosing what to wear. 

Capital One’s research also suggests nearly 7 in 10 shoppers want AI to automatically reorder items that are running low, and nearly half are comfortable with the idea of it shopping for their groceries.

If you’re reading the tea leaves, that all adds up to: “We would if we could, but we don’t think we can (yet).” 

And while that’s not a “yes”, it’s certainly not a “no.”

 

In a Room Full of Rocks, Move Like Water

Some content designers are adamant they shouldn’t — or won’t — change anything about how they create content to accommodate GEO. They say the bots will learn from and accommodate their nuance as the programs improve. 

But as we’ve discussed, that’s a fundamental misunderstanding of how AI retrieves and synthesizes web content. At worst, it’s willfully ignorant. 

Either way, refusing to accommodate Agent Experience in your content design could render your digital storefront irrelevant to a growing user niche.

One thing’s for certain: optimization in the age of agentic AI will be a murky future to navigate. That’s why Mindgrub keeps charting the path forward. 

Need someone who can help you speak Robot? We’ve got you covered.