# Event Recap: AI Search Summit 2026

Three hours of real data, live demos, and hard-won lessons from Chris Donnelly (Searchable), Dan Slagen (Zapier), and Ethan Smith (Graphite) on how brands are winning visibility in AI search right now.

**Published:** January 22, 2026
**Author:** Chris Donnelly

---

**Chris Donnelly (Searchable), Dan Slagen (Zapier), Ethan Smith (Graphite) — January 22, 2026**

AI search is not coming. It has already arrived. In this three-hour summit, Chris Donnelly, Dan Slagen from Zapier, and Ethan Smith from Graphite broke down exactly how brands are winning visibility in AI search right now, from real citation data to live demos showing what is actually driving results.

### Watch the full recording

<div style={{ position: "relative", paddingBottom: "56.25%", height: 0, overflow: "hidden", borderRadius: "12px" }}>
  <iframe src="https://www.youtube.com/embed/X6Ki5yJgxuc" title="AI Search Summit 2026 — Full Recording" frameBorder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowFullScreen style={{ position: "absolute", top: 0, left: 0, width: "100%", height: "100%", border: 0 }} />
</div>

&nbsp;

## Top 5 takeaways

1. **AI search is already shaping purchase decisions before anyone clicks.** Buyers research on ChatGPT, form a shortlist, then Google the brand they already chose. The decision happens in the conversation, not the click.

2. **One in four Zapier signups is now influenced by LLMs.** That number is growing and Zapier treats GEO as a core acquisition channel, not a side project.

3. **50% of Zapier's content effort is going offsite.** Reddit, YouTube, PR, and community are now just as important as the company website for AI citations.

4. **Everything that works in SEO works in AEO.** FAQs, question-format headers, and original data all drive higher AI citations. The incremental work is offsite.

5. **Search data beats prompt volume data.** The prompts people use in AI are longer and more specific than keywords, but search data is still the most reliable starting point for understanding what your customers are asking.

## The search landscape has already changed

People are no longer choosing between Google and AI. They use both, but for different things. AI handles the research and recommendation phase. Google handles the final click. That split means brands are losing attribution on the most important part of the journey.

58.5% of US searches and 59.7% of European searches now end without a click. Meanwhile ChatGPT has 800 million weekly active users, bigger than Twitter. The adoption curve of AI has moved faster than the internet, the iPhone, and the app store combined. This is not a trend to watch. It is a behaviour to optimise for now.

## Zapier: treating GEO as a core acquisition channel

Dan Slagen, SVP Marketing at Zapier, was direct. One in four Zapier signups is now influenced by LLM recommendations. The team has moved AI search from a side project into a clearly owned programme with its own strategy, measurement framework, and dedicated resources.

The shift in mindset is as important as the tactics. Rather than asking why a page is not ranking, Zapier now asks how a specific model is learning about them and from whom. That question leads to very different answers.

## Why offsite matters more than most brands realise

Zapier is putting 50% of its content effort offsite. Reddit, YouTube, PR, earned media, social platforms, and review sites are all being cited by LLMs. The company website alone is not enough.

- Original proprietary data signals to LLMs that a brand has something no competitor can copy
- PR is no longer just top of funnel. It is training data. Where you get coverage matters
- Community and customer content build the kind of consensus that LLMs use to form recommendations
- Reddit ads work well for threads that are already talking positively about your brand

The key distinction Dan made: do not spray and pray on PR. Know the specific publications being cited for your category and overinvest there.

## Ethan Smith: live demos on what actually drives citations

Ethan Smith from Graphite ran the session with no slides and live demonstrations only. Three areas stood out.

**Reddit:** engage when people are talking about you, not before. Showing up to threads unprompted and promoting your brand will get you banned. Ethan demonstrated this with a live example. The better approach is to engage authentically when your brand is already being discussed, even in negative threads. His own AMA in a subreddit that initially criticised him became the most liked post in that community's history.

**Video:** the least entertaining categories are the best opportunity. Nobody watches videos about payroll software on a Friday night, which means the category is completely uncrowded. Ethan showed a Rippling video on identity access management with 4,300 views that still shows up consistently in AI search results. Take your top SEO landing pages, make a video summarising the content, put it on YouTube, embed it on the page, and it shows up in video search, AI search, and gets views from your existing page traffic simultaneously.

**On-site:** FAQs and question-format headers work. Ethan tested FAQs across 250 Webflow template pages and saw an 8% traffic increase from LLMs. Changing headers from "PNG versus JPEG" to "What are PNG and JPEG image formats?" also produced measurable lifts. The structure does not need to be a formal FAQ schema. It just needs to cover the subject comprehensively.

## Information gain: the content that actually separates you

The reason AI generated content usually does not work is that it is a derivative of a derivative. Everyone is rewriting each other's content, and the most typical version wins. That is not useful.

What works is deriving from a unique data set. Ethan showed a medical site called Sermo outranking WebMD despite far lower domain authority, because it had AI-summarised reviews from real doctors sharing actual patient outcomes. That content existed nowhere else. The same principle applies to any brand with proprietary data, real customer stories, or internal expertise that a competitor cannot copy.

## How to measure AI search impact

Both Chris and Dan were honest that measurement is still being figured out. The most useful signals right now:

- Track branded and non-branded mentions and citations separately across ChatGPT, Perplexity, Gemini, and Claude
- Monitor sentiment: how the AI describes your brand, not just whether it mentions you
- Run incrementality tests when you launch campaigns to separate AI-driven lift from organic growth
- Ask customers directly where they heard about you and include LLMs as an option
- Do not rely on last-click attribution to tell the full story. The decision often happens in a conversation that never shows up in your data

The traditional visibility score that most tools offer is, in Chris's words, like marking your own homework. Tracking the right prompts matters far more than tracking a lot of prompts.

## What is next

The next webinar goes deeper on competitor gap analysis: how to find every prompt where rivals are being cited and you are not, and exactly how to close those gaps.

**[April 1st, 4pm BST — Save your spot](https://www.searchable.com/events/competitor-gaps)**

**[Want to see where your competitors are showing up and you are not? Start your free trial](https://www.searchable.com/signup)**

---

[Back to Blog](https://www.searchable.com/blog) | [Searchable Homepage](https://www.searchable.com)