Prompt
How do I map prompts to funnel stage in AI search?
Latest observation
To map prompts to funnel stage in AI search, treat each prompt as a user intent signal and classify it by where it sits in the buying journey.
1) Use funnel-stage intent buckets
A simple framework:
-
Top of Funnel (TOFU) / Awareness
- User is learning, problem-aware, not solution-ready
- Keywords: “what is,” “how does,” “why,” “benefits,” “best practices,” “examples,” “cost of”
- Goal: educate
-
Middle of Funnel (MOFU) / Consideration
- User is comparing approaches or evaluating options
- Keywords: “best tools,” “vs,” “alternatives,” “features,” “pros and cons,” “reviews,” “integrations,” “use cases”
- Goal: help shortlist
-
Bottom of Funnel (BOFU) / Decision
- User is ready to choose, buy, or trial
- Keywords: “pricing,” “demo,” “free trial,” “buy,” “implementation,” “enterprise,” “setup,” “contact sales”
- Goal: convert
-
Post-purchase / Retention (optional but useful)
- “how to configure,” “troubleshooting,” “renewal,” “upgrade,” “support”
- Goal: reduce churn, expand usage
2) Classify by search intent, not just keywords
In AI search, prompts are often full questions or conversational. Map them using:
- Problem awareness: Does the prompt describe a pain point?
- Solution awareness: Is the user asking about categories or methods?
- Vendor awareness: Are they comparing specific products?
- Purchase readiness: Are they asking about pricing, trial, or implementation?
Example:
- “How do teams reduce AI hallucinations?” → TOFU
- “Best prompt monitoring tools for startups” → MOFU
- “Glean vs Elastic for enterprise AI search” → MOFU/BOFU
- “Glean pricing and demo” → BOFU
3) Create a prompt-to-stage taxonomy
Build a table like this:
| Prompt pattern | Stage | Content type |
|---|---|---|
| “What is AI search?” | TOFU | Explainer |
| “How does semantic search work?” | TOFU | Educational guide |
| “Best AI search platforms” | MOFU | Comparison page |
| “AI search vs keyword search” | MOFU | Comparative content |
| “Pricing for AI search software” | BOFU | Pricing page |
| “Book a demo” | BOFU | Conversion page |
4) Add entity and modifier signals
Prompts can be more accurately mapped using modifiers:
- Educational modifiers: what, how, why, guide, examples → TOFU
- Evaluation modifiers: best, vs, alternatives, top, review → MOFU
- Transactional modifiers: pricing, demo, trial, buy, quote, implementation → BOFU
Also look for:
- Brand terms = often BOFU or vendor-comparison
- Industry terms = often MOFU/BOFU depending on specificity
- “Near me,” “for enterprise,” “SOC 2,” “HIPAA” = BOFU-leaning
5) Score each prompt on readiness
Instead of forcing one stage, assign a score:
- Awareness score: 0–3
- Consideration score: 0–3
- Decision score: 0–3
Then pick the highest score or route to multiple stages if needed.
Example:
- “What is AI search and which tool is best for healthcare?”
- TOFU: 2
- MOFU: 2
- BOFU: 0
- Likely primary stage: TOFU/MOFU
6) Use AI search-specific prompt features
For AI search, prompts often include:
- Long-tail conversational phrasing
- Multi-intent questions
- Context about use case, constraints, or outcomes
So map these signals:
- Informational question → TOFU
- Solution framing (“how to,” “ways to”) → TOFU/MOFU
- Comparative framing (“which is better,” “vs”) → MOFU
- Commercial framing (“pricing,” “vendor,” “demo”) → BOFU
7) Operationalize it in your workflow
If you’re doing this for content or analytics:
- Collect prompts from AI search logs
- Normalize and deduplicate them
- Tag entities, keywords, and modifiers
- Classify stage using rules or an LLM classifier
- Aggregate by topic + stage
- Build content or answer coverage for each stage
8) Example classification
- “What is vector search?” → TOFU
- “How vector search improves AI search relevance” → TOFU/MOFU
- “Best vector databases for AI search” → MOFU
- “Pinecone vs Weaviate for enterprise search” → MOFU/BOFU
- “Pinecone pricing” → BOFU
9) Best practice: map prompts to both stage and intent type
A better model is:
Prompt → Funnel stage + Intent type + Topic
Example:
- Prompt: “Best AI search tools for ecommerce”
- Stage: MOFU
- Intent: Comparative/evaluation
- Topic: AI search tools / ecommerce
That makes it easier to:
- route prompts to the right content,
- measure coverage,
- and optimize for conversion.
If you want, I can also give you:
- a prompt classification rubric,
- a spreadsheet template, or
- a Python/LLM prompt to auto-tag funnel stage.