Parse
IndexPricing
Sign inCheck your brand
  1. Research
Research
Does AI keep the same recommendation when its sources change?

Does AI keep the same recommendation when its sources change?

Published Aug 7, 2026Updated Oct 12, 2026

On this page
  • The finding
  • How we measured
  • Sources
  • More like this
  • The finding
  • How we measured
  • Sources
  • More like this

Dimitry ApollonskyFounder, Parse

DataWhat reviewed means

Changing citations can make an answer look substantially different even when its recommendation stays the same. We wanted to compare stability in the first brand with stability in the cited pages.

We examined back-to-back answer pairs on ChatGPT Search and Google AI Mode from May 24 to July 16, 2026. We compared top-brand retention across different levels of source overlap, including pairs where none of the cited pages repeated.

An entirely new source list did not always change the top recommendation. The top brand stayed the same in 47.3% of consecutive answers where no cited page repeated.

The same top brand remained first after complete cited-page turnover in 47.3% of pairs

The top recommendation stayed the same in 12,243 of 25,883 back-to-back answer pairs where no cited page repeated, or 47.3%. It changed in the other 13,640 pairs, or 52.7%. A back-to-back pair contains two consecutive answers from the same engine to the same organic prompt.

A completely different set of cited pages did not always change the top brand. Track the recommendation outcome and the cited pages as separate fields. SparkToro found less than a one-in-100 chance that repeated ChatGPT or Google AI runs would produce the same complete brand list twice. This study narrows that question to one first recommendation and the pairs where every cited page changed.

Takeaway

Record the first recommendation and the cleaned set of cited pages separately for every repeated prompt.

Every cited page changed in 21.1% of back-to-back answer pairs

No cleaned cited-page address repeated in 25,883 of 122,882 back-to-back answer pairs, or 21.1%. These complete-turnover pairs covered 10,077 organic prompts.

Complete page turnover occurred often enough to audit as a recurring condition. The rate describes this observed prompt corpus and should not be treated as a probability for every future answer.

of back-to-back pairs repeated no cited page
21.1%

25,883 of 122,882

pairs with complete cited-page turnover
25,883
organic prompts represented
10,077

Two repeated pages raised top-brand stability to 62.8%

The same top brand remained first in 12,243 of 25,883 pairs with no shared page, or 47.3%; 14,597 of 27,424 pairs with one shared page, or 53.2%; and 43,707 of 69,575 pairs with two or more shared pages, or 62.8%. The last group was 15.5 percentage points above the no-shared-page group.

Page continuity was associated with higher recommendation stability. It does not show that a repeated page caused the engine to keep the brand first. Ahrefs similarly found that Google AI Mode and Google AI Overviews could reach similar conclusions while citing different pages, but its comparison was across engines rather than across back-to-back runs.

Same top-brand rate by repeated cited pages

  • No shared page47.3%12,243 of 25,883
  • One shared page53.2%14,597 of 27,424
  • Two or more shared pages62.8%43,707 of 69,575
  • 0%20%40%60%80%
Back-to-back answer pairs on the same prompt and engine.

Takeaway

Use page continuity to segment recommendation stability, not to infer causation.

ChatGPT Search held the same top brand more often than Google AI Mode

When no cited page repeated, ChatGPT Search kept the same top brand in 9,515 of 18,970 pairs, or 50.2%. Google AI Mode did so in 2,728 of 6,913 pairs, or 39.5%. The difference was 10.7 percentage points.

A combined rate hides a material engine difference. Run the same source-change audit for each engine instead of applying one aggregate benchmark to both.

Same top-brand rate with no repeated cited page

  • ChatGPT Search50.2%9,515 of 18,970
  • Google AI Mode39.5%2,728 of 6,913
  • 0%20%40%60%

Takeaway

Review recommendation stability separately for ChatGPT Search and Google AI Mode.

The time between answers did not create a steady pattern

With no repeated cited page, the same top brand remained first in 2,879 of 6,400 one-day pairs, or 45.0%; 8,705 of 17,994 pairs two to seven days apart, or 48.4%; and 659 of 1,489 pairs eight or more days apart, or 44.3%.

A longer time between answers was not associated with a steady rise or fall in this window. Do not turn one interval into a general decay rule. The intervals also contain different prompt mixes.

Same top-brand rate by time between answers

  • One day45.0%2,879 of 6,400
  • Two to seven days48.4%8,705 of 17,994
  • Eight or more days44.3%659 of 1,489
  • 0%20%40%60%
Pairs where no cited page repeated.

Industry stability ranged from 40.0% to 61.8%

Among displayed industries with at least 200 complete-turnover pairs, Consumer Electronics kept the same top brand in 134 of 217 pairs, or 61.8%. Information Technology reached 57.8%, Software 53.6%, Financial Services 45.0%, Commerce and Shopping 43.4%, Internet Services 42.4%, Food and Beverage 41.9%, and Professional Services 40.0%.

Compare a brand with its category baseline before treating its result as unusual. These rates describe the observed prompt mix and do not show that industry caused recommendation stability.

Selected industry same top-brand rates

  • Consumer Electronics61.8%134 of 217
  • Information Technology57.8%660 of 1,141
  • Software53.6%736 of 1,373
  • Financial Services45.0%712 of 1,582
  • Commerce and Shopping43.4%430 of 990
  • Internet Services42.4%179 of 422
  • Food and Beverage41.9%178 of 425
  • Professional Services40.0%221 of 553
  • 0%20%40%60%80%
Each displayed industry has at least 200 complete-turnover pairs.

LinkedIn stayed first in 86.1% of its complete-turnover pairs

LinkedIn remained first in 290 of 337 complete-turnover pairs across 66 prompts, or 86.1%. Linear reached 72.3%, Bright Data 70.6%, HubSpot 68.7%, Datadog 67.3%, Stripe 64.2%, Microsoft 54.1%, and Semrush 37.5%.

Named-brand rates show where to investigate repeated prompt histories. They reflect each brand's prompt mix and do not form a universal brand-quality ranking.

Selected brand stability after complete page turnover

Stayed-first rate · 8 results

  • LinkedIn86.1%
  • Linear72.3%
  • Bright Data70.6%
  • HubSpot68.7%
  • Datadog67.3%
  • Stripe64.2%
  • Microsoft54.1%
  • Semrush37.5%
  • 0%50%100%
Explore the data (8 rows)
LinkedIn33729086.1%66
Linear20214672.3%31
Bright Data17012070.6%26
HubSpot1157968.7%68
Datadog33322467.3%74
Stripe16210464.2%63
Microsoft29415954.1%196
Semrush722737.5%45
Cleaned root brands with at least 20 complete-turnover pairs.

  • Current HubSpot measurements

The result held under stricter page and prompt-history rules

The main result, using cleaned page addresses, was 12,243 of 25,883, or 47.3%. The original, uncleaned page addresses produced 12,508 of 26,377, or 47.4%. Requiring at least two cited pages in each answer produced 9,610 of 21,283, or 45.2%. Retaining only the first complete-turnover pair per prompt and engine produced 5,813 of 12,630, or 46.0%.

We excluded ambiguous first recommendations, answers without a cited page, and prompt, engine, and timestamp combinations that contained more than one answer. Page addresses were counted once after fragments, trailing slashes, and eight standard tracking parameters were removed. The study does not measure factual support, recommendation quality, user demand, or causation.

main result, cleaned page addresses
47.3%

12,243 of 25,883

original page addresses
47.4%

12,508 of 26,377

at least two pages in each answer
45.2%

9,610 of 21,283

first turnover pair per prompt and engine
46.0%

5,813 of 12,630

What marketers should do

The top brand stayed the same in 47.3% of complete-turnover pairs. Stability rose to 62.8% with at least two repeated cited pages, and the no-shared-page rate differed by 10.7 percentage points between the two engines.

Record the first recommendation and the cleaned set of cited pages for every repeated priority prompt. Review each engine separately. Inspect the answer before changing content because this study does not show that a source change caused the brand outcome. Rerun the fixed-window method next quarter before calling a difference movement.

Takeaway

Track the recommendation and cited pages together, then inspect engine-specific changes before acting.
  • Is the AI visibility drop real or just noise? A measurement guide

How we measured

We analyzed 122,882 back-to-back answer pairs across 159,114 AI answers, 14,881 organic prompts, 21,578 top brands, 835,775 cited pages, and 206,622 cited websites on ChatGPT Search and Google AI Mode from May 24 through July 16, 2026.

of pairs with no repeated cited page kept the same top brand
47.3%

12,243 of 25,883

of back-to-back answer pairs repeated no cited page
21.1%

25,883 of 122,882

separated pairs with two repeated pages from pairs with none
15.5 pp

62.8% versus 47.3%

separated the two engines when no cited page repeated
10.7 pp

50.2% versus 39.5%

Get the data

Dataset CSVThe metrics behind every figure in this report.

Sources

  1. SparkToro: AI brand and product recommendation inconsistency (opens in a new tab) · accessed August 7, 2026
  2. Ahrefs: AI Mode and AI Overviews (opens in a new tab) · accessed August 7, 2026
  3. BrightEdge: Different sources but similar brands (opens in a new tab) · accessed August 7, 2026
  4. Semrush: Ghost citations and brand mentions (opens in a new tab) · accessed August 7, 2026

More like this

Does AI recommend the same brand when you ask again?
Asking the same engine the same question again kept the first recommendation in only about six in ten comparisons.
How long an AI citation lasts
About half of a question’s cited websites disappeared on the next run, while a smaller set persisted for weeks.
Do shared sources lead to the same AI recommendation?
Citing the same page did not guarantee agreement: the engines still chose different top brands in over half of matched answers.
AI citation volatility: why one missing source is not a trend
Repeated answers often cite different sources. A missing citation alone does not establish a trend.
Does question wording change which brands AI names?
Rephrasing a buyer question changed the brands named more than repeating the identical wording did in our comparison.
Which AI engine changes its answers the most?
Neither engine was consistently steadier. Both changed much of their brand list when asked the same question again.
ChatGPT Search: do answers with sources name the same brands?
Answers with and without recorded sources usually named different first brands in matched comparisons.
Does AI change its tone when cited sources turn negative?
AI's descriptions stayed largely positive even after its cited sources changed from positive to negative.
How long does AI take to notice a website change?
Changed prices often took time to appear: most checked changes on cited pages were not reflected in answers within two months.

Track first recommendations and cited pages as separate measures.

Parse

Index

  • Brands
  • Markets
  • Topics
  • How we measure

For agents

  • Docs
  • Connect
  • Workflows
  • llms.txt

Parse

  • Pricing
  • Research
  • Blog
  • Compare

Company

  • About
  • Contact
  • Corrections

© 2026 Parse, Inc. All rights reserved.

Privacy PolicyTerms of Service