What counts as a prompt?
A prompt is one saved question your buyers might ask an AI assistant, such as “What is the best email tool for small shops?”. Your plan limits how many prompts can be active at the same time; disabled prompts do not count.
How often is each prompt checked?
Every active prompt is checked on each included AI platform at the frequency you choose: daily, every 3 days or every 7 days. Each check stores the answer, the brands mentioned, sentiment and cited sources.
Can I change my monitoring frequency?
Yes, from Monitoring settings at any time. On Brand plans the frequency is part of the price, so you see the new price before you confirm and the difference is prorated. History is kept and the next run is rescheduled from the last one; existing prompts are not all re-run at once.
Can I change my plan?
Yes. Upgrades are charged the prorated difference immediately; downgrades are credited on your next invoice. If you have more prompts or projects than the new plan allows, disable some first.
What happens if I run out of credits?
Agency plans pause new monitoring runs when the month’s credits are used, notify the account owner and resume automatically at the start of the next billing period. Failed or cancelled checks never use credits. Unused credits expire at the end of each period.
How does the Free plan work?
You get one project with up to 50 prompts. Narrashift analyzes real-user AI prompt responses from ChatGPT and Gemini to build your initial dashboard. Your plan card lists the requirements for recurring runs.
Which AI platforms are supported?
Paid plans include 3 of these 5 AI platforms: ChatGPT, Gemini, Claude, Perplexity and Google AI Mode. A daily plan with 150 prompts is up to 13,500 monitoring executions a month (150 prompts × 3 platforms × 30 days). The Free plan analyzes initial real-user AI prompt responses from ChatGPT and Gemini.
Are monitoring results identical to consumer ChatGPT, Gemini, or Claude apps?
No. API answers come from the providers’ developer models and search tools and can differ from what the consumer apps show, which also depend on account settings, memory, location and model version. Narrashift reports what the configured models answered, not a guarantee of what every user sees.