1 Million Keywords Reveal Where AI Search Impact Actually Landed in 2026

Fractl and Search Engine Land just published the largest keyword-level study of AI's impact on search: 1,010,848 high-volume keywords across 379 brands in eight verticals. The headline number is that 29% of search demand is declining. The data underneath says something more useful: demand didn't vanish. It redistributed to wherever AI decides who gets named.

What 1 Million Keywords Show About Demand Redistribution

The Fractl study measured every keyword with 10,000+ monthly searches across FinTech, HealthTech, Wellness, Insurance, SaaS, Lifestyle, Travel, and more. 29% of high-volume search demand is in measurable decline, 4 percentage points above Gartner's original 25% forecast from February 2024.

Here is where most coverage stops and the useful part starts. 20.1% of keywords are growing by the same threshold. The 285,489 declining keywords add up to 10.29 billion monthly searches. The 140,835 growing keywords add up to 10.31 billion. Net change: +16.8 million searches per month. Fewer keywords are growing, but each growing keyword carries more volume. Demand relocated. It did not contract.

Why Non-Branded Queries Move First

Across the Fractl dataset, 90% of all tracked search volume is non-branded. That number climbs to 99.6% in HealthTech and 98.5% in Wellness. When a query does not include a brand name, there is no specific site the user has to reach. AI answers the question, the session ends, and the user never visits a search result page.

This is the structural fact most operators are still not measuring against: 93% of AI search sessions end without a single website click. The query did not disappear. The user got an answer from ChatGPT, Perplexity, or Google AI Mode. If your brand was named in that answer, you won the impression. If not, the query still happened. You just were not in it.

I have watched this exact pattern across our client base at AuthorityTech for the past two years. The non-branded queries are the first to move because AI handles them without sending the user anywhere else. The branded queries follow only for the brands the AI named in the non-branded answer.

The Verticals Where Demand Is Growing, Not Shrinking

The Fractl data breaks the "AI is killing search" narrative by vertical. FinTech saw the largest decline at -37.7%. Lifestyle had the smallest at -15.2%. But the growth-to-decline ratios tell the real story:

Lifestyle leads at 2.6x: 40% of keywords growing versus 15% declining. SaaS sits at 2.5x: 48% growing versus 19% declining. HealthTech inverts at 0.4x, the most disrupted vertical in the set.

The pattern is structural. Verticals where AI provides a complete answer (drug interactions, deductible explanations, fund comparisons) lose search volume permanently. Verticals where AI recommendations drive a downstream transaction (software purchases, retail comparisons, travel bookings) gain volume because the AI answer creates a new branded search.

I coined Machine Relations because this pattern was already visible in 2024. When the AI names your brand in a non-branded answer, the buyer searches your name next. That downstream search is new volume. It only exists if you earned the citation.

AI Referral Traffic Converts at 2.5x Google Organic

Axis Intelligence compiled conversion data from Similarweb's April-May 2026 clickstream panel: AI referral traffic converts at 7.1%. That is second only to paid search at 7.8% and 2.5 times higher than Google organic at 2.8%. Claude generates the highest conversion rate of any AI platform at 16.8%, per Digital Bloom's February 2026 analysis.

The implication for operators: you do not need the same traffic volume from AI that you had from Google organic. You need a fraction of it, because it converts at more than double the rate. A single citation in a ChatGPT response to a buyer query is worth more downstream revenue than a top-three organic ranking on the same keyword. The buyer arriving from the AI answer has already been told by the machine that your brand is the one worth considering.

This is why citation architecture replaces keyword rankings as the measurement that matters. The metric is not whether you rank. The metric is whether the AI cites you when the buyer asks.

Six AI Surfaces Now Decide Brand Visibility

The search market fragmented faster than most operators realize. PPL Studio's mid-2026 benchmark maps the six surfaces that now control brand visibility:

ChatGPT Search processes roughly 3 billion AI-search queries per month across 700 million weekly active users, with a median brand citation rate of 1.2% on commercial queries and a top quartile of 6.8% or higher.

Perplexity runs 85 million monthly queries and surfaces 5 to 12 cited sources per answer. DTC brands report 3 to 8 times higher conversion rates from Perplexity referrals compared to Google organic.

Google AI Overviews now appear on 73% of US informational queries, up from 47% at the start of the year. Brands that fail to appear in AI Overviews on their top 50 commercial queries see informational traffic decline 30% to 55% year over year.

Microsoft Copilot handles 1.6 billion monthly queries across Windows, Edge, and Microsoft 365. PPL Studio found that citation share inside Copilot is the single highest-correlated metric with B2B SaaS pipeline lift, at r = 0.71 across a 40-brand cohort.

Amazon Rufus reaches 110 million weekly active users. Listings that fail to surface in Rufus on category queries lose an average of 22% of organic glance views inside 90 days.

Claude handles roughly 80 million weekly active users with the highest per-session conversion rate of any platform.

StatCounter's April 2026 data shows ChatGPT's referral share fell to 76.85%, its lowest on record, while Gemini reached 9.0%, Perplexity 7.73%, Copilot 3.76%, and Claude 2.66%. The market is fragmenting and every surface has its own retrieval logic. A single-platform AI strategy built in 2025 is already obsolete.

What Operators Should Measure Instead of Rankings

I built AuthorityTech on the conviction that earned media in trusted publications is the raw material AI engines use to decide who gets cited. The 1-million-keyword study confirms the structural thesis: search demand is not dying. It is moving to surfaces where machines decide which brands get named.

The Axis Intelligence AI Search Fragmentation Index dropped 50.5% in 12 months, confirming rapid fragmentation. But the conversion premium rose simultaneously, from 3.1x to 4.8x Google organic. The market is splitting across more surfaces while delivering higher-quality traffic per session.

For operators, the audit is straightforward. Pull your tracked keywords and filter by year-over-year volume change. The keywords losing volume are the ones where AI already answers the query. For those keywords, your measurement target shifts from ranking to citation. Are you the brand ChatGPT, Perplexity, and Google AI Mode name when the buyer asks? That is Machine Relations: the discipline of earning AI engine citations through trusted third-party sources. The brands doing it now are collecting the 7.1% conversion rate on traffic that used to be a click on a search result page.

FAQ

What does the Search Engine Land 1 million keywords study show about AI search impact in 2026?

The Fractl and Search Engine Land study analyzed 1,010,848 high-volume keywords across 379 brands in eight verticals. 29% of search demand is declining, but 20.1% is growing. The net volume change is nearly zero because demand redistributed rather than contracted. Information-heavy verticals like FinTech lost the most volume while transaction-driven verticals like SaaS gained 48% in growing keywords.

How much better does AI referral traffic convert compared to Google organic?

AI referral traffic converts at 7.1%, based on Similarweb's April-May 2026 clickstream data. That is 2.5 times higher than Google organic search at 2.8%. Claude generates the highest conversion rate of any AI platform at 16.8%. The higher conversion rate means operators need less absolute traffic volume from AI citations to match or exceed the pipeline value of organic search.

Which AI search platforms matter most for brand visibility in 2026?

Six platforms now control brand visibility: ChatGPT Search (3 billion queries per month), Google AI Overviews (73% of US informational queries), Microsoft Copilot (1.6 billion queries per month), Amazon Rufus (110 million weekly users), Perplexity (85 million queries per month), and Claude (80 million weekly users). Each has its own retrieval logic and citation pattern, per PPL Studio's mid-2026 benchmarks.

What is Machine Relations and why does it matter for AI search impact?

Machine Relations is the discipline of earning AI engine citations through trusted third-party sources. I coined the term in 2024 after documenting the link between earned media placement and AI citation eligibility. For the 29% of search queries where AI now provides the answer directly, traditional SEO rankings no longer determine brand visibility. The metric that predicts pipeline impact is citation architecture: whether AI engines name your brand when buyers ask.

Who is Jaxon Parrott and what is AuthorityTech?

Jaxon Parrott is the founder and CEO of AuthorityTech, the first AI-native Machine Relations agency. He built the company on 1,500+ direct editorial relationships and measures client success by citation architecture rather than clip counts. AuthorityTech charges brands only when placements land in publications AI engines actually cite, closing the gap between "placement secured" and "citation eligible" that defines the AI search era.

Related Reading

• Identity Security Companies Sell Trust but Are Invisible to the AI Engines That Sell Them
• How AI Security Companies Build Earned Media and AI Search Citations in 2026