89% of AI Search Demand Still Unsettled: Kevin Indig’s Data Shows a Narrow Window for Brands

By Coruja News Desk

A new analysis of ChatGPT’s answer patterns across 1,094 U.S. categories has produced a clear message for brands and SEO professionals: the fight for AI search visibility is still in its early innings, and the window to claim a dominant position may close within the next few quarters.

Kevin Indig, a well-known growth strategist, obtained access to six months of ChatGPT answer data from Semrush covering January to June 2026. The dataset spans five prompts per category, more than 50,000 brands, and over 600,000 citations. Indig published his findings on his Growth Memo Substack on July 20, 2026, and shared the headline results on LinkedIn the same day. The analysis was later covered by Search Engine Journal, which highlighted the strategic implications for digital marketers.

The data behind the finding

Indig classified each category into one of three buckets: clear owner (one brand appears in at least four of five test prompts and leads the runner-up by five percentage points or more), emerging leader (a brand leads at least three prompts but does not meet the five-point gap), and unsettled (no brand leads even three of five prompts). The results show that only 15.2% of categories have a clear owner. More than half (53.7%) are unsettled, with multiple credible contenders and no brand close to dominance.

The picture becomes sharper when broken down by AI search volume. The higher-volume half of categories accounts for 98% of all AI search demand in the sample, yet it has an even lower owner rate of 11.3% compared to 19% for the lower-volume half. In total, 89.3% of estimated AI search demand sits inside categories that still have no clear owner.

What it means for brands

Indig’s data settles a long-running debate about whether topical authority carries over into AI search. The answer is yes, but only for brands that show up consistently across all five prompt types: definition, comparison, alternatives, use case, and buying decision. The analysis shows that once a brand earns an outsized share of mentions in a category, it tends to hold onto it. But the vast majority of categories have not yet reached that stage.

The categories worth the most money to occupy are currently the least decided. Brands that treat category focus as optional are effectively giving away large, high-value segments to any competitor willing to invest in consistent presence.

Why citations are not enough

One of the most counterintuitive findings is that citations and brand mentions are only weakly related. Indig found a slightly negative correlation of -0.229 between the two. The most frequently cited domain in a category matched the most frequently mentioned brand only 20.8% of the time. However, the most mentioned brand was cited at least once 69.9% of the time — meaning that being cited at all is a door, but not the room.

Owners did share some common traits compared to their runners-up: higher branded search volume in 55.7% of pairs, higher organic traffic in 48.4%, and a higher Semrush Authority Score in 52.5%. Indig stresses these are correlations, not a formula. Yet they suggest that a combination of brand strength and consistent topical relevance is more important than raw citation volume.

Key Data

Metric	Value
Categories with clear owner	2%
Categories unsettled	7%
Share of AI search demand in unsettled categories	3%
Correlation between citations and brand mentions	-0.229

What practitioners should do next

Indig also tracked month-over-month leadership changes. A clear owner held first place in 90.4% of comparisons. The switches that did occur were concentrated in categories where the leader’s margin was thin — a median lead of 1.3 percentage points going into the switch, compared to 2.9 points for categories where the leader held on. This means that a one- or two-point lead is not a safe lead, and brands should assume they are still in a fight.

The GEO (Generative Engine Optimization) conversation over the past year has fixated on citation counts because they are easy to measure. Indig’s data suggests that practitioners who optimise only for citations while a competitor optimises for being the name ChatGPT actually says out loud are likely to lose categories they thought they were winning.

The window Indig describes will not remain open. With 89% of demand still unsettled, the next few quarters represent an unusually generous head start. Brands that move now — focusing on deep topical content across all prompt types, building branded search volume, and monitoring their share of mentions rather than just citation counts — can still become the default answer when similar studies are run in 2027.

Source: Search Engine Journal, “89% Of AI Search Demand Has No Clear Owner: Use This Before The Window Closes” by Greg Jarboe, July 29, 2026. https://www.searchenginejournal.com/89-of-ai-search-demand-has-no-clear-owner-use-this-before-the-window-closes/583255/