Podqi Test Finds Risky Sites in 15% of Perplexity Answers

A Podqi shopping test found risky sites in 15% of Perplexity answers, 8% of ChatGPT answers and 4% of Gemini answers, raising concerns for consumers using chatbots to find sellers.

The rates do not show how often a chatbot endorsed a site. Podqi counted recommendations, links shown without comment and warnings about a site. A warning therefore counted toward the same measure as a shopping suggestion.

Podqi, an intellectual-property protection company, tested 6,795 answers about 151 brands across two testing periods in September 2026. Its prompts asked where a brand’s products could be bought, whether it shipped internationally and whether it had regional websites. Podqi wrote the questions itself rather than using real shopper conversations.

The test’s risk category included fake sites impersonating retailers, sellers of knockoffs, old brand domains registered by someone else and inactive sites that another party could buy and reuse.

Queries about smaller brands surfaced risky sites about twice as often as queries about larger companies. Questions about regional websites also produced risky sites more often, including purported regional sites for some brands that did not have them. ChatGPT warned users in 24% of the instances in which it surfaced a risky site.

Podqi CEO and co-founder Trevor West said it is “always best to verify the source” before entering personal or credit-card information. Google spokesperson Jessica Levis Boord said, “We take information quality seriously.” OpenAI spokesperson Taya Christianson said ChatGPT should connect people with “reliable places to shop.” Perplexity spokesperson Jesse Dwyer said, “We never claim to be 100% accurate.”

The test did not report whether people clicked the sites, bought anything or suffered losses. Similar risks have appeared in crypto-related searches, including fake websites targeting Ledger users’ recovery phrases.