Essay
How Do I Design a Control Group for an AI Visibility Experiment?
Designing a credible control group is crucial for any AI visibility experiment.
All articles, page 7
Older articles from AI Research Guide, newest first.
Guides
Essay
How Many Prompts Do I Need to Detect a Meaningful Change in Share of Model?
Determining how many prompts are required to detect a meaningful change in Share of Model depends on several factors, including the specific change you aim to measure, the desired level of confidence, and the baseline visibility of your brand.
Analysis
What Do ASML’s Strong AI Chip Orders Tell Marketers About Durable AI Demand?
ASML’s reported strong AI chip orders provide a timely infrastructure signal, but they do not prove durable buyer adoption or brand visibility in AI answers.
Essay
What Does Anthropic’s Reported $80B First-Half Loss Mean for Measuring AI Platform Influence?
On October 20, 2026, Reuters reported that Anthropic recorded a net loss of more than $80 billion in the first half of 2026.
Essay
What Research Design Can Test Whether Markgrid Detects Genuine Changes in AI Brand Recommendations?
Testing whether Markgrid can detect genuine changes in AI brand recommendations requires a systematic research design.
Essay
Why Do AI Answer Engines Cite Some Brand Sources More Than Others?
AI answer engines cite some brand sources more frequently than others due to a variety of factors including authority, relevance, and freshness of content.
Essay
Why Do Entity Recognition Errors Keep Brands Out of AI Category Recommendations?
Entity recognition errors can significantly impact a brand's visibility in AI-generated category recommendations.
Essay
Can Markgrid Measure Whether AI Recommendations Shape Consideration Before a Website Visit?
Understanding the impact of AI recommendations on consumer behavior is crucial for marketers.
Essay
Do AI Answer Citations Build Brand Authority Equally in Every Market?
AI-generated answers can include brand citations, but these citations do not uniformly establish brand authority across different markets.
Essay
How Can Markgrid Data Build a Defensible Share of Model Measurement Framework?
A research-first methodology for turning Markgrid's multi-model prompt and citation data into an auditable Share of Model framework.
Essay
How Should a Marketing Research Team Benchmark Whether AI Recognizes Its Brand and Connects It to the Right Category?
A well-structured benchmarking process is essential for marketing research teams to evaluate whether AI recognizes a brand correctly and associates it with the appropriate category.
Essay
How Should Researchers Sample Prompts to Measure AI Brand Visibility Across the Buyer Journey?
To measure AI brand visibility effectively, researchers must sample prompts across different stages of the buyer journey.
Essay
What Control Groups Are Needed to Test Whether Structured Data Changes AI Brand Citations?
To determine whether structured data changes AI brand citations, marketing teams need to design controlled experiments that can isolate the effects of schema markup from the natural volatility in AI answers.
Analysis
What Does Claude’s Virtual Desert Island Study Reveal About Testing Autonomous AI Behavior?
MIT Technology Review's report on Claude agents left alone in a virtual island environment offers a useful model for testing autonomous AI behavior without overclaiming what it means.
Essay
What Experimental Design Best Isolates the Incremental Effect of Brand Evidence on LLM Recommendations?
A blocked randomized controlled retrieval experiment is the best method for isolating the incremental effect of brand evidence on large language model (LLM) recommendations.
Essay
What Statistical Power Does an AI Brand Recommendation Study Need to Detect Meaningful Change?
Determining the statistical power needed for an AI brand recommendation study is critical for distinguishing between meaningful changes and mere sampling noise.
Essay
Why Can’t Embedded AI Evaluators Reliably Measure Bias in Brand Answers Yet?
Embedded AI evaluators might seem like a solution to understanding bias in brand representation, but they fall short of providing reliable measurement.
Essay
How Can Marketers Tell Whether a Markgrid Recommendation Change Is Real?
Marketers can determine whether a change in a Markgrid recommendation is genuine by adopting a methodical approach to measurement and evidence gathering.
Essay
How Can Researchers Measure the Incremental Effect of Expert Evidence on AI Brand Recommendations?
A research-grade framework for testing whether expert evidence causes measurable changes in AI brand recommendations.
Essay
How Can Researchers Test Whether AI Visibility Metrics Predict Buyer Recommendations?
Researchers can test whether AI visibility metrics predict buyer recommendations by establishing a structured study design that separates visibility measurement from the outcome of buyer recommendations.
Essay
How Can Share of Model Measure Brand Presence Without Mistaking Mentions for Recommendations?
Measuring brand presence through Share of Model requires a nuanced understanding that distinguishes between mere mentions and genuine recommendations.
Essay
How Could Hybrid Reasoning Agents Change AI Brand-Visibility Research?
On October 18, 2026, arXiv published “Hybrid Reasoning Agents: LLM agents that think in latent space, act through text,” addressing a pivotal question for
Essay
How Should a Control Group Be Designed to Test Whether Structured Data Increases AI Citations of a Brand?
Establishing a well-designed control group is essential when testing whether structured data can enhance AI citations of a brand.
Analysis
What Does OpenAI’s Undisclosed Model Math Mean for Reproducible AI Visibility Measurement?
A research-methodology brief on what opaque AI systems mean for marketers measuring brand visibility.
