AI Prompt Tracking
Tier: Pro | Type: User-Facing UI
!!! tip "Key Insight" Discovers which AI search prompts already mention your brand or cite your pages — across ChatGPT, Perplexity, Gemini, Google AI Overview, and AI Mode. No setup required: SE Ranking's prompt database does the discovery. Uses data credits (100/scan, 24h cache). Follows the same credit-gating pattern as Link Profile.
Overview
AI Prompt Tracking answers the question: "When people ask AI engines about my industry, do they mention me?" It scans SE Ranking's AI search prompt database across five engines and surfaces every prompt where the brand appears or a page is cited.
What It Shows
| Panel | Data |
|---|---|
| Summary Cards | Total prompts found, brand mentions, page citations, number of AI engines where you appear |
| Per-Engine Breakdown | Prompt count, brand mention count, and citation count for each engine (ChatGPT, Perplexity, Gemini, Google AI Overview, AI Mode) |
| Discovered Prompts | The actual prompt text, which engine, type (Brand Mention / Page Citation / Both), and cited URL if applicable |
| Cited Pages | Which of your pages are being cited by AI engines, with citation count and engine badges |
How It Works
- User clicks Scan Prompts (or Refresh if cached data exists)
- Confirmation modal shows credit cost (100), current balance, and projected balance
- Backend loops through 5 AI engines, making 2 SE Ranking API calls per engine:
/ai-search/prompts-by-brand— prompts where the brand name is mentioned in AI answers/ai-search/prompts-by-target— prompts where the domain's pages are cited/linked
- Results are merged, de-duplicated, and normalized into per-engine and per-page summaries
- Saved as a snapshot in
seranking_snapshots(24h cache) - Cached snapshots load for free on subsequent visits
Smart Brand Matching
The SE Ranking API requires a brand name that matches how AI engines reference the business — which often differs from the domain name. The system tries multiple brand name variations:
- Site name from the database (e.g., "Divorce Amicably")
- Domain stem as fallback (e.g., "divorceamicably")
It tries each variant per engine and stops at the first one that returns results. This prevents zero-result false negatives for sites where the brand name doesn't match the domain.
Citation Attribution
Each AI engine response includes an answer.links array — the URLs cited in the AI-generated answer. The system extracts URLs matching the target domain from both prompts-by-brand and prompts-by-target responses and maps them back to specific pages.
This powers two UI elements:
- Cited Page column in the Discovered Prompts table — shows exactly which page was cited for each prompt
- Your Pages Being Cited panel — aggregates all citations by page URL, showing citation count and which engines cite each page
Prompts that both mention the brand AND cite a page show a "Brand + Citation" badge.
Empty State Handling
The feature distinguishes between three states:
- Never scanned — shows CTA to run first scan
- Scanned, no results — shows actionable guidance with link to AI Search Readiness for improvement recommendations
- Scanning in progress — hides empty state, shows spinner
Why It Matters
- AI search is growing — an increasing share of search queries are answered by AI engines instead of traditional results
- Visibility you can't see otherwise — there's no way to manually check if ChatGPT mentions your business
- Actionable intelligence — knowing which prompts cite you (and which don't) tells you exactly what content to create or optimize
- Progress tracking — re-scanning after improvements shows whether your AI visibility is growing
Requirements
- Required: SE Ranking API key (
SERANKING_API_KEY) - Depends On: Site Profile (site name used for brand matching)
Cost Per Scan
| Component | Cost |
|---|---|
| SE Ranking API (10 calls: 2 endpoints x 5 engines) | ~50 vendor credits |
| AlmaSEO data credits | 100 credits |
| Cache window | 24 hours |
Backend Modules
| Module | Purpose |
|---|---|
integrations/seranking_client.py |
get_ai_prompt_visibility() — API calls, merging, normalization |
integrations/seranking_credit_guard.py |
ACTION_REGISTRY["seranking.ai_prompt_tracking"] — cost and cache config |
routes/ai_prompt_routes.py |
Blueprint with /balance, /data, /refresh endpoints |
templates/client_profile/ai_prompt_tracking.html |
Full UI template with credit gating |
Demo
Demo deep link: /demo/go/ai-prompt-tracking
All three demo sites have pre-populated prompt data:
| Site | Prompts | Brand Mentions | Citations | Engines |
|---|---|---|---|---|
| Rapid Flow Plumbing | 23 | 14 | 9 | 4 |
| Carter & Associates Law | 18 | 10 | 7 | 4 |
| Glow Skincare Studio | 12 | 6 | 4 | 3 |