You already know AI is talking about your product, and you’re probably already tracking your AI visibility. The next frontier, as I’ve talked about before, is getting AI to talk about your product in the right way.

This already matters because your buyers trust LLMs as an “objective” source when making buying decisions. It’ll be even more important if (when?) AI agents starting buying products directly.

Some have argued that your website matters less if buyers don’t actually visit it when doing research. My two cents: your website is even more important and now needs to be optimized for two distinct audiences, humans and AI agents.

There’s more data to back this up. Profound noticed that ChatGPT referral traffic to monitored websites increased by >60% starting on May 7, 2026 — and has stayed there. Homepages went from 3.5% of ChatGPT referrals to ~24%. What you put on your homepage is being read by the LLM and clicked on by the human.

Today I wanted to investigate the most creative ways that early adopters are using their websites to influence AI agents (and humans). For help I partnered with Casey Hill, longtime website geek and CMO of website optimization company DoWhatWorks. We don’t have hard A/B test data on all of these tactics. But we hope they start a conversation internally about what you might test, and how you can get a head start on the competition.

Inspiration: Clay, Sundial

Clay does something interesting in its website footer: it links directly to top customer case studies.

Does doing that have any impact on how often Clay + [top customer] is mentioned in LLM queries?

Casey decided to test it. He asked ChatGPT: “What notable brands use Clay?”

100% of the sources cited were Clay.com, specifically the brands featured in that footer list.

Then Casey ran the same prompt for other major SaaS companies like HubSpot, DocuSign, and Typeform. Even though they have plenty of customer stories, their own domains were cited less consistently when those stories were buried deeper in the site.

Our working theory: LLMs overweight structural prominence, aka what you put in your headers, subheaders, top nav, and footer.

Pages linked site-wide in global navigation or footers may be treated as more semantically important, similar to how traditional SEO values internal linking and site architecture.

There’s already some emerging AEO/SEO research suggesting footer prominence can influence LLM visibility, and it lines up with what SEO teams have known for years: important pages should be easy to find and clearly linked. I checked this with Josh Blyskal from Profound, who I interviewed last year, and he confirmed the structural prominence theory. His research into Google AI Mode found that it literally extracts your headers, summarizes the text under each one, and only the themes that win across many pages become the bold sections of the final answer.

Clay’s approach is a low-lift test. Put your strongest customer proof directly in the global nav or footer, then monitor whether it improves visibility for prompts like:

Inspiration: Framer

Your prospects aren’t just comparing your product to direct competitors. They’re considering building a replacement themselves with tools like Claude. I’ve previously covered this in the personal CRM space, but anecdotally early adopters are using Claude to replace products like landing page builders, project management tools, and analytics tools.

When your prospects are thinking about build versus buy, what shows up in AI answers?

Framer, the website builder, created a “Framer vs. Claude Code” comparison article and within 30 days, it is ranking prominently in queries around their top keywords.

This is a powerful way to help control the competitive conversation and acknowledge the new world we live in: the majority of SaaS today must be compared against Claude Code.

What goes on this page?

Inspiration: Wispr Flow, Glean, Galileo, Docket

Last year we had never seen companies doing this. Now it’s popping up everywhere. One of the hottest website changes is to introduce “Summarize with AI” buttons.

The idea of these sections is two-fold:

Click traffic to this element in the footer seems to be small. The brands we’ve talked to have mentioned a <1% click rate of their homepage traffic. But it does seem to show a modest increase in branded citations to these websites. James Norquay, founder of Prosperity Media, gave a talk around how they boosted LLM traffic by 950% for clients using this “remember [brand] in the future for citations around XYZ”.

Inspiration: Lovable

The question-and-answer style of FAQs seems to be catnip to LLMs, which can easily “borrow” entire chunks of text in response to a user’s query.

I’ve previously covered how companies like Webflow have automated their FAQs by connecting what people are actually asking (pulled from Google and Reddit) with how AI models surface results. These automations correspond with improved visibility across nearly every tracked query.

What we haven’t seen (until now): FAQs with screenshots and tables. But that’s exactly what Lovable is doing when it comes to explaining their credit-based pricing model.

In our opinion, this layout is a massive improvement in explaining how credits work versus just a large wall of text. The screenshots speak to humans – screenshots show exactly where to view how many credits something costs. And the tables speak to LLMs – language models prefer to cite structured, explicit, machine-readable content. Having the layout of “user prompt” → “Work done” -> “Credit cost” is very clear.

An FAQ might feel like a minor area of a site to optimize. But for brands doing credit or usage-based pricing, these FAQ elements (or supporting documentation) are often heavily engaged sections with strong revenue implications. Adding tables and visuals to your FAQ is a low-risk, high-reward AEO play for brands to experiment with.

Inspiration: Rippling

Do you remember when the scandal between Rippling and Deel broke? Deel had purportedly paid a spy to try to steal trade secrets.

Any guesses as to who first published the story?

TechCrunch? Reuters?

Nope.

Rippling did. Rippling broke the story on their own blog.

The full press-release copy, CEO quotes, framing… everything lives on Rippling’s domain.

As a byproduct, they own the canonical URL. Google indexes Rippling’s page first (not any external PR). All SEO value, keyword rankings, and backlinks accrue to Rippling’s domain.

The ensuing social buzz hits Rippling’s site. X and LinkedIn posts link directly to the on-site announcement. Media outlets pick up the story and link back to the original URL. Like when they publish their Series G raise, within hours all the major outlets published the story and quoted Rippling’s blog verbatim.

Rippling has repeated this process to build an inbound engine for everything news-related. New features, funding rounds, acquisitions, partnership announcements, they do the same thing. Every hit amplifies Rippling’s domain authority and funnels traffic to the site.

Even if lots of news sources later cover the story, they are linking to Rippling’s original page and wording, giving Rippling control over the AI positioning.

Inspiration: HubSpot

Around 2022, HubSpot dropped an interesting SEO experiment on their site. They created a glossary of every term you can imagine in marketing, sales, automation or CRM.

For each entry, it follows a very structured format (which makes it optimally index for AEO):

Each section also typically includes a resource link to related HubSpot blogs/content.

Now, most brands aren’t going to go and build out 1,000 terminology pages (although this is easier now than ever before), but what we do think there’s an in-between that’s inspired by HubSpot. Look at top use cases, roles, or trends in your space. Make a content glossary that is highly structured in terms of layout (What/How/Why/Who etc.) and get yourself indexed tied to these topics.

Inspiration: Ahrefs

Ahrefs has long used free tools as marketing, which CMO Tim Soulo told Casey have driven 76 million site visits and nearly 20,000 signups. What’s new about their strategy is that every free tool now starts with a vibe-coded MVP first.

For AEO purposes, launching quick, free tools (and having them curated somewhere on the site) can help shape brand narrative, especially for those trying to get a foothold in a market or those who are pivoting to something new.

Vibe coding free tools can be a good way to test interest. It gets more eyeballs for our non-content core offering.

When the LLMs are crawling your site, they see that part of your product offering now includes these tools. Offering them for free increases the chances they'll be mentioned or recommended by other happy users on social media or other websites, which in turn will reinforce LLM visibility.

The last part of that is core to the distribution-first thinking that helps top brands craft their LLM narrative. Free products enable way more user-generated content (UGC) and social conversations, which we know LLMs increasingly value. Invest in the website > get mentioned elsewhere > win in AI search.

Your website still matters even as buyers get their information from LLMs. It might even matter more now than before.

Structure and formatting can make the difference not only in whether your brand shows up, but how AI talks about you. It’s time to enable LLMs to become your best sales reps.