Research
The Editorial Team writes Answerburst's research articles from the AI Visibility Benchmark.
Two kinds of article, both from the Answerburst AI Visibility Benchmark, which asks ChatGPT, Claude, Gemini, Perplexity, Google AI Mode and AI Overviews the same questions every week. None of the questions names a brand. Once a month has ended, a report on what changed across every category. And an article on each brand whose share of the answers moved most from one month to the next, up or down: what changed, where, which brands gained or lost with it, and why when a source shows it.
Every number in an article is counted from the benchmark's answers by a script, and an article can only print numbers that script produced. Quotes are copied from the answers. We call something a change only when a statistical test, corrected for the number of comparisons we made, says it's unlikely to be chance, and only beyond what longer answers explain: when assistants name more brands per answer, every brand's share rises.
An AI model writes each article from that script's output: the numbers, the quotes, the pages the assistants cited and news coverage of the brand. It can't print a number the script didn't count, and it can't give a reason without linking the source that shows it.
An assistant naming a brand isn't an assistant recommending it, and we don't say why an assistant's answers changed unless a source shows it.
A person on the team reviews every article before it's published, and publishes it. When we find an error, we fix the article and show the date it was updated.
Talk to Answerburst about getting your brand into more AI answers.