Google AI Overviews dominate software review searches, new study says

A new NetPartners Marketing study found Google shows AI Overviews on nearly all software review and alternatives searches, while review traffic has dropped sharply over the past year. The findings suggest publishers and affiliate marketers now need original, source-worthy content to stay visible in AI-driven search results.

Why it matters: - Software research is shifting from traditional search results into AI answers, reducing the chance that users will click through to review sites. - The study found review searches fell by a median 50% year over year, a sign that demand is still there even as search visibility changes. - Publishers that rely on affiliate and review traffic may need deeper reporting, original data and dated numbers to get cited by AI systems.

What happened: - NetPartners Marketing released a study on Sept. 24, 2026 examining how Google, ChatGPT and Gemini surface software review and affiliate content. - Google showed an AI Overview on 98% of "[brand] review" searches for business software and on 100% of "[brand] alternatives" searches. - AI Overviews appeared on 89% of combined software review, alternatives and pricing searches. - The study used 1,761 live Google results pages in the U.S. collected on Sept. 23, 2026, plus 479 questions posed to ChatGPT and Gemini about software categories and affiliate programs.

The details: - Pricing searches had the lowest AI Overview coverage at 75%, because Google often sent users directly to the vendor. - Among 177 "[brand] review" keywords with at least 50 monthly searches, 86% declined year over year. - AI Overviews cited 8,503 sources across the sample. - Independent websites earned 36% of citations. - Forums and video accounted for 23% of citations. - Software vendors accounted for 19% of citations. - Review platforms such as G2 and Trustpilot accounted for 18% of citations. - YouTube was the most cited single domain, appearing in 53% of AI Overviews that listed sources. - ChatGPT and Gemini never named 48 of the 117 top-ranked recurring affiliate programs across 326 category answers. - HubSpot, GetResponse and ActiveCampaign were named most often. - Of 90 programs advertising "30% recurring" commissions or more, 29 did not hold up. - Twenty-seven of those programs paid less at the entry tier or stopped paying too early under their own terms pages. - Two programs published no terms at all. - The Atlas scores each program by what one referral is worth over 24 months, search demand and whether independent websites can realistically win traffic. - HighLevel, which pays 40% recurring for the life of the customer, is the benchmark index of 100. - SE Ranking, Skool, Systeme.io and Thinkific scored highest among the 116 other programs on the board. - The full methodology, program pages and underlying data are available free of charge in the Atlas AI search report.

Between the lines: - The study suggests Google’s AI Overviews are absorbing a growing share of top-of-funnel research, especially for comparison and review queries. - That could make page-one rankings less valuable if AI summaries already satisfy the searcher’s first question. - The citation mix shows AI systems still lean on a broad set of sources, but they favor content that looks current, specific and independently useful. - The affiliate-program findings point to a gap between marketing claims and actual commission terms.

What's next: - NetPartners Marketing says the Atlas and the AI search report will continue to publish ranking data and methodology for free. - Review publishers and affiliate marketers will likely track whether AI Overviews keep expanding across software-buying queries. - The company’s findings may push more publishers to optimize for citation-worthy original research instead of generic list posts.

The bottom line: - AI answers are now a major gatekeeper in software discovery, and the sites most likely to survive are the ones offering original data that AI systems can quote.**