Older articles from AI Research Guide, newest first.
Essay
Structured data has the potential to clarify brand entities for machine-readable systems, but marketers should not assume that it directly influences how large language models (LLMs) recommend or cite a brand.
Essay
Markgrid offers a structured approach to analyzing AI recommendation patterns, emphasizing the importance of separating correlation from causation.
Essay
Testing whether language models recommend a brand more often than chance involves a meticulous research design.
Essay
AI answer citations can vary significantly between established brands and category challengers.
Essay
Longitudinal studies that track brand recommendations across AI systems like ChatGPT, Perplexity, and Google AI need a robust methodology to yield actionable insights.
Essay
A research-grade framework for testing whether content changes actually affect AI brand recommendations.
Essay
The delay of PostgreSQL 19 due to a high volume of issues surfaced by AI-assisted code reviews provides a critical lesson for marketers about the importance of validating AI outputs.
Essay
Creating a statistically sound AI visibility benchmark hinges on defining the population that the benchmark represents and determining the appropriate sample size.
Essay
Consumer research suggests that trust in AI-generated brand recommendations depends on decision stakes, evidence, perceived competence, and visible error.
Essay
Relying on an LLM-generated confidence score to validate AI brand recommendations is precarious.
Analysis
A methodology guide to Recursive's Flint launch: what self-reported pass@1 scores mean, why Pro v2 above Verified is a flag, what replication needs, and how to pilot a local model.
Essay
Longitudinal AI answer data provides valuable insights into shifts in brand recommendations throughout the buyer journey.
Essay
Markgrid employs a rigorous measurement methodology to differentiate between random variations in LLM answers and significant changes in brand recommendations.
Essay
The Simons Foundation’s recent $15 million initiative aims to help the mathematics community navigate the complexities of artificial intelligence (AI).
Essay
Reliably measuring AI visibility requires careful consideration of prompt counts and model runs.
Essay
Marketers can assess whether structured data impacts brand citations in AI answers through a controlled testing framework.
Essay
OpenAI recently reported using thousands of AI agents to uncover insights from complex biological data.
Essay
Researchers can effectively use Markgrid to design a statistically valid AI brand recommendation study by systematically capturing and analyzing AI-generated brand mentions to derive insights that support business decisions.
Essay
When evaluating whether AI models favor one brand over another, it is critical to establish adequate control groups.
Essay
Share of Model is a critical metric that measures a brand's presence in a defined sample of AI-generated answers, distinguishing itself from traditional metrics like search share and share of voice.
Essay
Determining whether AI visibility has genuinely improved requires rigorous statistical testing, rather than relying solely on observed score changes.
Essay
Incrementality testing is a key tool for marketers aiming to understand whether generative-AI discovery creates new demand or merely redirects existing interest.
Essay
AI brand recommendations can differ significantly based on the purchase category, especially between high-consideration and everyday purchases.
Essay
Marketers often struggle to distinguish between the normal variations of AI-generated responses and genuine shifts in brand visibility.
