I just hosted a webinar with Chris Long, co-founder of Nectiv, where we walked through five recent studies that every marketing leader should know about right now. Chris has been in SEO for over a decade and posts some of the most tactical AI Search content on LinkedIn, so I wanted to compare notes on what the data actually says about how brands are showing up in AI Search, and what to do about it this quarter.

Before the studies: Google I/O and the slow march toward AI Mode

We kicked off with a quick read on Google I/O. Google is reconfiguring its search box to push more users into AI Mode, and AI Overviews are now blended directly into the AI Mode experience. Click an Overview, expand it, and you scroll straight into a conversational prompt without ever hitting the traditional ten blue links.

Chris's take: AI Mode is almost certainly going to become the default search experience inside the next year. Google is stair-stepping users there deliberately because it has to protect its ad revenue while migrating an entire ecosystem.

I come from a paid background, so I look at every Google change through the monetization lens. Every new prompt is another chance for Google to serve a contextually relevant ad against a much richer query than a keyword. The slow rollout is the monetization test, not hesitation.

If you're planning for the next twelve months, assume AI Mode is the default and work backward from there.

Study #1: $750B of consumer spend will run through AI search by 2028

McKinsey projects that around $750 billion in consumer spend will flow through AI-powered search by 2028, with 20 to 50 percent of existing traffic at risk for many brands. About half of consumers already use AI-powered search today.

What sticks out to me is that 71 percent of users said they start their purchase journey in AI Search. That makes discovery the single highest-leverage stage to optimize for.

If a buyer asks AI for product recommendations and gets back four or five options, that's the shortlist. If you're not in the awareness bucket, you don't get to compete further down the funnel.

Study #2: 84% of CMOs now use LLMs to find vendors

Wynter surveyed roughly 110 CMOs at $50M+ ARR companies. The numbers at a glance:

  • 84% use LLMs for vendor discovery, up from 24% in 2025. Roughly a 4x jump in twelve months.
  • 80% arrive at sales calls already moderately or very familiar with the vendor.
  • 34% now invest in GEO or AEO software as a distinct line item.

The behavior pattern matters as much as the headline. Buyers use AI to build the shortlist, then bounce to Google to look for reviews, G2 ratings, and red flags. If you show up in the LLM but your reviews are thin, you won't survive verification.

Study #3: What actually influences AI citations

Zippy aggregated more than 50 studies on AI citation ranking factors. Three patterns:

  1. Traditional SEO still matters a lot. URL accessibility, traditional search rankings, on-page relevance. Good AEO starts with good SEO.
  2. Topic clusters beat keywords. As queries get longer and more conversational, you can't optimize for an exact phrase. You optimize for a topic and cover it across use cases, feature pages, and documentation.
  3. Content structure is doing heavy lifting. Tables, bulleted lists, numbered steps, explicit phrasing, dense answers near the top. AI rewards content that's easy to parse.

Quick plug here: we just released a new version of the Profound Index, a free report that lets you plug in your domain and see how your brand compares to competitors across a series of prompts in your industry, including topic-level breakdowns.

Study #4: Fan-out queries are the language AI actually searches in

Nectiv analyzed roughly 60,000 Google fan-out queries. A fan-out query is the set of sub-queries an AI generates from a single prompt before aggregating an answer.

The numbers are wild:

  • The average prompt fans out into about 9 sub-queries.
  • 24% of fan-out queries are between 12 and 19 words. The longest Chris's team measured was 28 words.

If you're tracking 100 priority prompts, the AI is actually searching across roughly 900 sub-queries to answer them. And those sub-queries explicitly look for accreditations, reviews (G2, Capterra), and freshness signals. The year is often in the query, which means content last updated in 2023 or 2024 isn't getting ingested.

What I'd take from this if I were a marketing leader: ask your team for a fan-out report this quarter. Which authorities is AI invoking in your category, what trust signals does it look for, and where are your content gaps. This shapes content strategy more directly than keyword research does now.

Study #5: Social is an AEO surface

Our own data shows social platforms account for a meaningful share of AI citations: 15.3% for AI Overviews, 14.5% for AI Mode, down to 3.6% for Gemini. The mix varies sharply by language and geography. YouTube is cited 38% of the time in English-language AI Overviews in Australia and the UK, but 65% of the time in Portuguese in Brazil.

Reddit dominates many categories. So does YouTube. But for B2B specifically, LinkedIn often outweighs both. At Profound, LinkedIn is in our top five cited domains, and it's almost always posts from our co-founders talking about product releases.

The lesson is to track first, generalize second. Look at your top cited domains and let citation volume direct your social strategy. If LinkedIn is in your top five and your company doesn't post on LinkedIn, that's a gap your competitors are already filling.

What I'd do this quarter

A few actions came up repeatedly across these studies:

  • Add a self-reported attribution field to your sales form. Required, open text, five-character minimum so people can't dodge it. This is the cheapest way to prove AI Search is driving pipeline.
  • Track fan-out queries, not just prompts. If you're tracking 100 prompts, you're missing 800 sub-queries that decide whether AI cites you.
  • Diversify your reporting metrics. Measure conversions, share of voice, Agent traffic from server logs, and citation rank together. Organic traffic alone won't tell the story anymore. We just launched Benchmarking inside Profound that compares your agent traffic to industry aggregates, which allows you to both know what good looks like and if your work is moving the needle.
  • Break out of the SEO silo. The teams winning are working in pods across analytics, comms, and affiliates toward shared AEO goals.

The brands moving fastest are the ones that stopped treating AI Search as a future planning exercise and started treating it as a present-quarter performance channel.

If you want to understand and take control of how your brand is showing up AI, reach out to our team.