$ ~/ym8 --define query-bank
Query Bank
definition
A Query Bank is the measurement instrument of AEO analytics. It is a carefully curated set of queries that represent the questions your target audience would ask AI engines about your product category, industry, or brand. By running these queries through multiple AI engines and analysing the responses, you can calculate Share of Model, Citation Rate, and other key AEO metrics.
Building an effective Query Bank requires understanding your audience's intent categories: informational queries ("what is a CRM"), comparative queries ("best CRM for small business"), transactional queries ("CRM pricing comparison"), and brand-specific queries ("is Salesforce good for startups"). Each category reveals different aspects of your AI visibility.
Query Banks should be refreshed regularly to account for evolving language patterns. AI users tend to phrase queries more conversationally than traditional search users, so Query Banks should include natural language variations alongside keyword-focused queries. For example, "CRM small business" might be a traditional search query, while "what CRM would you recommend for a 10-person startup" reflects how users actually talk to AI engines.
The size and composition of a Query Bank directly affects the accuracy of your AEO metrics. Too few queries produce unreliable data; too many become impractical to run regularly. Most effective Query Banks contain 30-100 queries per product category, covering the full spectrum of relevant intents.
why_it_matters
Without a well-constructed Query Bank, AEO metrics are unmeasurable. It provides the systematic basis for tracking your brand's AI visibility over time, benchmarking against competitors, and measuring the impact of your optimisation efforts. A Query Bank transforms AEO from guesswork into data-driven strategy.
examples
- A SaaS company maintaining a 50-query bank covering product category, feature-specific, and competitive comparison queries
- A fintech firm building separate Query Banks for different audience segments (consumers, businesses, regulators)
- An agency using a 100-query bank to benchmark client visibility across 8 AI engines monthly
faq
How many queries should a Query Bank contain?
Most effective Query Banks contain 30-100 queries per product category. Fewer than 30 may produce unreliable data, while more than 100 per category becomes impractical for regular monitoring. Start with 50 queries and adjust based on your measurement needs.
How often should I update my Query Bank?
Review and update your Query Bank quarterly. AI users' language patterns evolve, new competitors emerge, and product categories shift. Adding new queries and retiring obsolete ones keeps your metrics relevant and actionable.
Related Terms
Share of Model
Share of Model (SoM) measures how frequently a brand is mentioned or recommended by AI engines in response to relevant queries. It is the AI-era equivalent of Share of Voice, quantifying your brand's presence across ChatGPT, Perplexity, Gemini, Claude, and other answer engines.
Competitor Visibility
Competitor Visibility in AEO measures how often and how favourably your competitors appear in AI engine responses compared to your brand. It provides the competitive context necessary to understand whether your AI visibility position is strong, weak, or at risk.
AI Search Optimization
AI Search Optimization is the broad practice of optimising digital content and brand presence to perform well across all AI-powered search interfaces, including conversational AI (ChatGPT, Claude), AI-native search (Perplexity), and AI-enhanced traditional search (AI Overviews, AI Mode).
Related Engines
Monitor Your AI Visibility
See how your brand appears with the default core pair. Start with ChatGPT and Claude by default. Expand monitoring only when the workflow needs it.