How many prompts should I track for AI search?
- Also asked as:
- What's the ideal number of prompts to track?
- How many AI search prompts should I start with?
Scrunch recommends estimating prompt tracking needs using the following formula: X [# of core topics] x Y [5-8 questions related to each topic] = Z [# of AI search prompts to track]. The primary goal is to get a representative sampling of data across all customer journey stages via a mix of branded and non-branded prompts.
Additional context: As a general rule of thumb, brands should track at least 15-25 varied prompts across a minimum of 6-12 topics related to core business themes and customer needs to acquire representative AI search performance data.
Example
For example, a Scrunch user at an enterprise SaaS company determining their prompt tracking needs would follow this workflow:
Identify core topics
Start with SEO keywords, paid search keywords, and major product and content areas of the website. Filter to the top 20-30% of topics most relevant to brand and most likely to drive customer demand.
In this case, the user filtered a list of 3,000 keywords and product/content areas to the top 25%, resulting in 750 core topics.
Identify related questions
Aim for 5-8 questions for each topic that map to the different stages of the funnel (i.e., advice, awareness, evaluation, and comparison), include branded and non-branded variations, and are informed by questions that commonly appear in SEO research, PPC queries, and communications to sales, success, and support teams.
In this case, the user created six questions per topic.
Calculate total prompts
Multiply the number of core topics by the number of questions.
In this case, the user multiplied 750 core topics by six questions, resulting in 4,500 prompts.
If tracking a large number of prompts: Organize prompts by topic, persona, funnel stage, and other filters to obtain clearer reporting and analysis.
Follow-up question: How do I know if my prompt strategy reflects real user behavior?
Scrunch recommends tracking user agent traffic and LLM-referral traffic patterns. Compare the most-crawled and visited URLs with existing prompts to ensure prompt coverage and identify gaps where high-traffic pages lack corresponding prompts.
Brands should also refresh their prompt lists over time by:
- Reviewing AI search trend data on a monthly basis to identify new topics and related prompts
- Adding new prompts on a quarterly basis to reflect product launches, campaigns, and competitive landscape changes
- Archiving prompts that no longer matter or replacing them with updated variants
Related FAQs
What benchmarks or baselines are useful when evaluating AI search performance?
Scrunch recommends tracking brand presence, citations, referral traffic, AI agent traffic, and share of voice versus competitors as key performance indicators.
How can I see if my visibility in AI search is improving or declining over time?
Scrunch recommends monitoring AI search trend data like brand mentions and citations consistently over 2-3 week periods to identify real trends versus one-off changes.