Home AI Search Monitoring Setting Up Prompt Monitoring

Setting Up Prompt Monitoring

Last updated on Mar 19, 2026

Setting Up Prompt Monitoring

Prompt monitoring is the core of AI Rankia. You choose prompts (search queries your customers ask AI), select which AI models to track, and AI Rankia monitors them on a recurring schedule.

How to add prompts

  1. Go to Prompt Selection in the sidebar
  2. Type your prompt — a conversational query your customers would ask
  3. Select which AI models to monitor
  4. Set your target country (AI responses vary by region)
  5. Choose frequency: daily or weekly
  6. Click to add the prompt to your monitoring list

Real prompts being monitored on the platform

Prompt Models Credits/run Location Frequency
"What is the best AI SEO platform I can subscribe to today for full site analysis?" GPT 5, Gemini, Claude, Perplexity, AI Mode, Grok 9 United States Weekly
"best ai seo tools" AI Overview, GPT 5, Gemini, Claude, Perplexity, AI Mode, Grok 11 Global Weekly
"cuales son las mejores herramientas para geo seo?" AI Overview, GPT 5, Gemini, Claude, Perplexity, AI Mode, Grok 11 Germany Weekly
"top indie musician in India" GPT 5 Search 1 Global Daily
"best biryani places in Kolkata" GPT 5, Perplexity, Claude 5 Global Weekly

Notice the variety — prompts can be in any language, target any country, and use any combination of models.

Choosing the right prompts

Think like your customer, not like a marketer:

  • Conversational questions: "What's the best tool for tracking SEO performance?"
  • Comparison queries: "Mailchimp vs ConvertKit for small businesses"
  • Problem-solving: "How do I get more organic traffic to my e-commerce site?"
  • Recommendation: "Best alternatives to [competitor name]"
  • In your target language: "cuales son las mejores herramientas para geo seo?"

Use AI Prompt Discovery (sidebar) to automatically generate relevant prompt suggestions. Costs 3 credits per discovery run.

Selecting AI models

AI Rankia offers two tiers:

Explorer — native models with web search

These simulate a real user searching the web through an AI engine. They produce the most realistic results.

Model Credit cost
GPT 5 Web Search 1
Google AI Overview 1
Google AI Mode 1
Perplexity Sonar 1
Grok 4.1 Web Search 1
Gemini Search Latest 2
Claude Search Latest 3

Voyager — non-native API models

These query the model directly via API without web search. Useful for testing how LLMs respond from training data alone. All cost 1 credit each.

Models: DeepSeek v3, DeepSeek R1, Llama 4 Maverick, Claude Sonnet 4, GPT 5, Grok 4, Ernie 4.5 (Baidu), Qwen 3, Mistral Medium, Kimi K2

Recommendation: Start with 5 cheapest native models (GPT 5, AI Overview, AI Mode, Perplexity, Grok) = 5 credits per run. This covers the engines most people actually use.

Monitoring cost calculator

Prompts Models Credits/run Weekly (×4) Monthly total
10 5 native (all 1-credit) 50 200 200 credits
10 7 native (incl. Gemini+Claude) 90 360 360 credits
20 5 native 100 400 400 credits
36 5 native 180 720 720 credits (= Starter plan)

Reading your results

After monitoring runs, go to Prompt Monitoring (main dashboard) to see:

  • Brand mentions per model — which AI engines mention your brand and how often
  • Citation sources — the exact URLs that AI models reference in their answers
  • Competitor visibility — your share of voice vs. competitors
  • Trend charts — visibility changes over time
  • Full AI responses — read the exact text each AI model generated

Prompt Library

All your monitored prompts are saved in the Prompt Library (sidebar). From there you can review past results, re-run prompts, or archive ones you no longer need.

Monitoring tools that run automatically

For each monitored prompt, AI Rankia also collects:

  • Shopping data — AI Shopping results (ChatGPT, Perplexity, AI Mode)
  • Citation Share of Voice — your % of total citations vs. competitors
  • Query Fan-Out — the sub-queries AI generates from your prompt
  • Map data — local/maps visibility in AI search