ChatGPT and Perplexity have entered Decodo's ranking of the most-scraped websites, in third and fifth place, signalling rising demand for data from AI answer platforms.
TikTok retained the top spot for the second year running, while Google stayed in second place in Decodo's 2026 ranking. Naver ranked fourth, followed by Perplexity, Make.com, Target, Amazon, GitHub and Stack Overflow.
The ranking is based on anonymised usage data from Decodo's customer base and tracks which websites received the highest volume of scraping requests during the year. This year's list saw the biggest change in the report's three-year history, with seven of the previous top 10 dropping out.
Only TikTok, Google and Amazon remained from the previous year's list. YouTube, Walmart, Coupang, eBay, ScienceDirect, Crunchbase and Airbnb all fell out of the top 10.
Vaidotas Juknys, Chief Executive Officer at Decodo, said the purpose of web scraping was changing as users and companies shifted towards AI-generated answers.
"Two years ago, businesses scraped the web to see what people were searching for. Now they scrape AI platforms to see what people are being told. The use case for data collection is changing," Juknys said.
Businesses are increasingly collecting public data from AI platforms to assess how brands and products are presented, what sources are cited, and how their positioning compares with rivals. Decodo linked that activity to the growth of generative engine optimisation, or GEO, as marketers try to understand visibility in AI responses rather than only in conventional search rankings.
Gabriele Vitke, Product Marketing Lead at Decodo, said the appearance of AI answer engines in the ranking reflected a broader shift in where customer decisions are influenced.
"For years, AI companies were the ones collecting the web, crawling pages to train their models and, in Perplexity's case, pulling live sources into every answer. Now those same platforms have become some of the web's most valuable data sources. ChatGPT and Perplexity together accounted for 24% of scraping requests in our top 10 this year, more than Google on its own.
"The bigger point is that AI answers have become a new layer between brands and customers. When someone asks ChatGPT which project management tool to use or which running shoe is best for flat feet, the answer shapes the decision before they ever reach a website. If a brand isn't mentioned, or is described incorrectly, it may be losing customers it never knew it had. Collecting data from these platforms is the only way to see that layer at scale," Vitke said.
Vitke said scraping of AI platforms tended to focus on three areas: whether a company is mentioned in response to a question, whether details such as pricing and policies are described correctly, and which sources are cited in the answer.
"Most of the data we see being collected falls into three main categories. First, brand visibility: whether a company is mentioned for a given question, how often, and how it ranks against named competitors. Second, accuracy: how the AI describes a company's pricing, features or policies, because an outdated or wrong answer can cost a sale. Finally, citations: which sources ChatGPT and Perplexity rely on. That list becomes a guide to where a brand needs to be published or reviewed to appear in AI answers," she said.
New entrants
Developer and automation platforms also appeared in the ranking, signalling a broader change in the type of data being collected. Make.com reached sixth place, while GitHub and Stack Overflow entered in ninth and 10th.
"GitHub [#9] and Stack Overflow [#10] mark a new kind of entry for this list. Both sites hold code and technical answers instead of search results or product listings. Teams pull that content straight into their own AI models," Vitke said.
Search and retail platforms remained prominent. Google held second place, while Naver entered in fourth, reflecting demand for public data from regional search engines as businesses look beyond Google in overseas markets.
In retail, Target entered in seventh and Amazon remained in the top 10 in eighth, despite falling five places from the previous year. Retailers continue to gather pricing, product, inventory, marketplace and review data to track competitors and shifts in demand.
Vitke said changes in AI usage were one reason for the increase in scraping activity aimed at answer engines. She pointed to OpenAI's reported 900 million weekly active users in February 2026 and said marketers were following audiences that once relied more heavily on search.
"OpenAI reported 900 million weekly active users in February 2026, more than double the year before. Questions that used to go to Google now go to an AI agent, and marketers are following their customers. This practice has a name - generative engine optimization, or GEO. And it's growing the way SEO did years ago.
"There's also a practical reason this has to be done through scraping: AI answers aren't fixed like a search results page. The same question can get a different answer depending on how it's phrased, where the user is or when they ask," Vitke said.
"Every year this report gets less about websites and more about where decisions get made," Juknys said. "In 2025 that was search. In 2026 it's AI answers and automation. Next year the agents might be doing the deciding instead of just reading the answers."