Prompt Mapping by Funnel Stage: The Content Gaps AI Visibility Scores Won't Show You

A 40% AI visibility score sounds respectable until you crack it open by funnel stage. Then you find your brand shows up in 70% of branded queries and 10% of problem-discovery prompts. The aggregate number was hiding a gap shaped exactly like the top of your funnel.

That gap matters more now than it did six months ago. A January 2026 HubSpot survey of 3,000+ decision-makers found AI search was the single strongest predictor of CRM purchase intent, beating demos, review sites, and sales calls. Meanwhile, 70–80% of B2B buying research happens before a buyer contacts sales, and 32% of those buyers use generative AI as much as or more than traditional search for vendor research. If your brand is invisible during the prompts that shape the shortlist, you're losing deals you'll never know existed.

The Funnel Compressed Into a Conversation

G2 has documented what many of us suspected: awareness, consideration, and shortlist formation can now happen inside a single AI conversation. A buyer who used to browse five tabs, read three G2 reviews, and attend a demo might now type progressively specific prompts into ChatGPT or Perplexity over fifteen minutes, land on two vendors, and reach out to validate a choice they've already made.

Buyers average 6.3 AI prompts before purchase. Only 14% convert in three prompts or fewer; 22% need ten or more. Each prompt carries different intent, which means prompt mapping is a diagnostic: where does your brand enter the AI conversation, where does it hold, and where does it vanish?

What to Track at Each Stage

Organize your prompt set around buyer intent, not keyword volume. Forrester and McKinsey both recommend optimizing for scenario-based context rather than search terms. Strong B2B prompts combine a task verb, a category, buyer identity, a buying trigger, and constraints like budget, tech stack, compliance, or industry.

Problem-aware (awareness). Prompts center on symptoms: "Why is our SaaS churn increasing even though CSAT scores are stable?" Your brand probably won't be named here. You're looking for whether AI platforms associate your category with the problem. If you sell churn prediction software and don't surface when buyers ask about identifying at-risk accounts, that's your first content gap.

Solution exploration (consideration). Prompts introduce capabilities: "What tools help predict customer churn before renewal?" Check whether AI connects your brand to the solution category. Category pages, use-case pages, and integration pages are the content types most likely to earn citations here.

Comparison and shortlist (evaluation). Prompts narrow: "Best churn prediction tools for enterprise SaaS that integrate with Salesforce." Watch how visibility changes as constraints tighten. You might appear for the generic version and vanish once the buyer adds an industry or integration requirement. Comparison and alternatives pages do the heavy lifting.

Validation (decision). Prompts turn branded: "Is [your brand] worth it?" Inclusion rate matters less; accuracy matters more. Is the pricing current? Are key capabilities represented correctly? Proof pages, case studies, and customer evidence shape what AI returns.

How to Build This

Start with 15–25 prompts covering category discovery, problem-aware, solution exploration, comparison, alternatives, integration, and validation clusters. Mature programs expand to 40–80, but starting too wide produces noisy data and no clear priorities.

Don't brainstorm prompts in a conference room. Pull real buyer language from sales calls, support tickets, on-site search, Reddit threads, and People Also Ask results. The questions prospects ask during discovery calls translate directly into consideration and decision-stage prompts. Turn those into natural-language prompts: "Best CRM software" becomes "What CRM is best for a 50-person B2B sales team using HubSpot for marketing that needs better pipeline reporting?" Change the company size, tech stack, or pain point and you have another prompt worth tracking.

Tag each prompt by funnel stage, topic, persona, and whether it's branded or non-branded. Then test across multiple AI systems: ChatGPT, Perplexity, Gemini, Google AI Mode. The same prompt can produce different brand mentions on different platforms.

The Measurement Problem Nobody's Solved Yet

Attribution for AI-influenced pipeline is still rough. AI exposure can shape a buyer's shortlist through repeated interactions that never produce a click. Standard analytics and last-click attribution miss this entirely. A Digiday report noted CMOs are struggling to connect AI visibility to sales outcomes, and that gap won't close quickly.

Don't wait for perfect measurement. Use prompt tracking as a leading indicator. If your brand appears consistently in problem-aware and comparison prompts across multiple AI systems, that's a directional signal your content architecture is working. If it doesn't, you've found the content sprint that matters most.

An estimated 80–100 million B2B-related prompts hit major AI tools daily in 2026. Somewhere in that volume, a buyer just asked an AI to shortlist vendors for your exact category, with your exact buyer's constraints. Whether your brand appeared depends on work you either did or didn't do last quarter. The prompt map is how you stop guessing which one it was.