The Maze: Google's AI-written search summaries can borrow information from a rival even when someone asks about a named software supplier. A newly published study found competitor-owned links in 21.7% of sampled answers about buying business software. That measures sources in answers, not customers switching suppliers.
Source: Foundation / Writesonic. Rival links by question type in this selected software study. B2B SaaS means subscription software sold to businesses; BOFU means questions close to a buying decision.
The Details:
The sample tracks questions, not purchases. Marketing agency Foundation and search-monitoring software provider Writesonic analysed 1,235 questions naming one software brand plus a buying term. They repeated them weekly from June 29 to September 21, generating 15,716 Google AI Overviews over 13 runs. An Overview is Google's AI-written summary with links to supporting websites. A competitor citation means a link to another supplier in the same software category. In 11.9% of answers, a rival had a source link while the named supplier did not. The named brand could still appear in the answer's text.
Asking for alternatives changes the result. The sample contained 541 alternatives questions, compared with 298 about pricing and 377 about reviews. Rival sites appeared in 35.8% of alternatives answers, 13.4% of pricing answers and 7.8% of reviews answers. That difference has a straightforward explanation: someone asking for alternatives wants other options. Rival inclusion is not automatically an intrusion. The named supplier's own site appeared in 75.7% of pricing answers but only 32.9% of alternatives answers. A company's ability to supply the right facts depends on what the buyer actually asks.
Existing comparison pages can become source material. Foundation says a Gorgias guide to Kustomer's pricing appeared in 12 of 13 weekly runs for one annual-contract question. Both companies sell customer-service software. The Gorgias page puts a pricing summary near the top and promotes its own product in a comparison table. It dates back to 2020 and shows a May 2026 update. This illustrates a potential influence point before a buyer visits the named supplier. It does not prove that the page's layout caused its citations, that every price is correct, or that anyone changed their purchase.
MarketMaze Take: A brand name is a starting point, not ownership of the answer. The useful finding is the large difference between question types. Suppliers can miss competitive exposure if they check only their company name. But a citation count alone cannot tell them whether that exposure cost a customer.
Separate the questions before buying the cure. Suppliers in other markets can test the pricing, reviews and alternatives questions their own buyers use. The study's weekly rival-citation rate moved between 12.7% and 30.2%, so one check can mislead. Foundation sells visibility services, and Writesonic supplies the tracking. Their data supports a better diagnostic; it does not establish that buying more comparison content guarantees citations or revenue.
Keep the comparison honest. Ahrefs found one rival site in 57 source lists from a separate panel of bare brand names. That is a different population, not proof that adding a buying word multiplies risk twelvefold. Foundation's selected English-language software sample also leaves search location and device conditions unspecified. Its percentage cannot stand in for all countries, industries or buyers.
Sources: Foundation, Gorgias, Ahrefs
