Salesforce AI Research put 21 expert users in front of Perplexity, Bing Chat, and You.com. The output: 16 failure modes, 8 measurable metrics, and the Answer Engine Evaluation benchmark — published in ACM FAccT 2025.
Reddit is the second-most-cited consumer source in ChatGPT, Claude, Gemini, and Perplexity — behind only Wikipedia. No Fortune 500 comms team has a named Reddit lead. The most-quoted source on your brand is the one nobody on your payroll owns.
Six academic studies that define how ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews cite the web. Authors named, methods explained, every number sourced. The reference document for practitioners, researchers, and journalists in AI visibility.
Five proprietary models. Five distinct citation logics. The category-level reference on who runs AI-mediated discovery — ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews — and the linked playbook for each.
The University of Toronto's definitive audit of GPT-4o, Claude, Perplexity, and Gemini versus Google Search across 1,516 queries. Domain overlap, source typology, content freshness, and the controlled experiment on pre-training versus retrieval bias.
AI search engines are creating challenges in legal contexts by oversimplifying complex legal distinctions. Examples show how AI flattens critical differences in employment law, eviction procedures, and sentencing exposure, leading to potentially inaccurate or misleading information. The dominant sources cited by AI include high-level legal bodies like the Supreme Court and popular platforms like Wikipedia, often overlooking crucial state-level and specialized legal authorities. This trend suggests legal institutions must publish structured primary sources to maintain their authority in the age of AI-driven synthesis.
Answer engines — ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews — now answer most early-stage buyer research. The pillar on how they decide which brands get cited.
ChatGPT is the most widely used AI product ever built — and the default starting point for product research, reputation queries, and category comparisons. The full entity profile.
Two and a half years into the NYT v. OpenAI lawsuit, the Times's archive remains largely outside ChatGPT and the major answer engines. Why the bet is rational — and what happens to the publisher class regardless of how the case resolves.
The structural shift is complete. What actually changed, what it means for every brand and communications team, and what the next 18 months look like.
