Prompts

Prompts are the questions Cleotic sends to AI models on your behalf. They're the foundation of your monitoring -- the quality and variety of your prompts directly determines how comprehensive your visibility data is.

Why prompts matter

AI visibility starts with asking the right questions. Your prompts should reflect the kinds of questions real people ask AI assistants about your industry, products, or category. The more relevant and varied your prompts, the more accurate your visibility picture becomes.

For example, if you're monitoring a CRM brand, good prompts might include:

• "What's the best CRM for small businesses?"
• "Compare Salesforce and HubSpot features"
• "Which CRM has the best email integration?"
• "What do experts recommend for enterprise CRM?"

Three ways to create prompts

1. Write your own

Type a prompt directly in the prompt form. You can also add tags (comma-separated) to categorise your prompts -- for example, "product-comparison", "feature-specific", or "recommendation". Tags help you organise and filter prompts later.

2. Generate with AI

If you have a general idea but want help expanding it, use AI-powered generation:

1. Type a seed prompt (e.g., "Best CRM for startups")
2. Click Generate Related
3. Cleotic suggests 3-5 related variations, each with auto-generated tags
4. Select the ones you like (or select all)
5. Click Add to create them

This is the fastest way to build a comprehensive prompt set. Start with one idea and let AI expand it into several related angles.

3. Batch add

When viewing generated suggestions, you can use Select All to pick every suggestion, then add them all at once. This is especially useful during initial setup when you want to quickly populate a monitor with diverse prompts.

Running prompts

Prompts run in two ways:

• Automatically -- Based on your monitor's refresh frequency (daily or weekly), all prompts in the monitor are sent to the configured AI models.
• On demand -- Click the Run button on any individual prompt to execute it immediately. This is useful for testing new prompts or getting fresh data right away.

When a prompt runs, Cleotic sends it to each AI model in the monitor and captures the full response. The response is then analysed for brand mentions, sentiment, and citations.

If some models fail during a run (for example, due to a temporary API issue), you'll see a partial success notification showing which models succeeded and which had errors.

Tags

Tags are optional labels you can attach to prompts for organisation. They're especially useful when you have many prompts and want to group them by topic or intent.

Some examples:

• product-comparison -- Prompts that ask AI to compare products
• recommendation -- Prompts asking for suggestions
• feature-specific -- Prompts about particular features
• industry-trend -- Prompts about market trends

Tags are set when creating a prompt (either manually or auto-generated by AI) and appear as badges on prompt cards.

Viewing prompts

From the monitor page, the Prompts tab shows all prompts with:

• The prompt text
• Tags
• Last run date
• Actions to run or delete

You can switch between card, vertical, and table view modes depending on your preference.

Tips

• Think like your customer. Write prompts that reflect how real people ask AI assistants about your category. Use natural language, not keywords.
• Cover different angles. Don't just ask "best X" -- also ask about specific features, comparisons, alternatives, and use cases.
• Use AI generation liberally. It's great at expanding a single idea into multiple angles you might not have thought of.
• Review and refine over time. Check your responses to see which prompts generate the most insightful data. Remove prompts that consistently produce irrelevant results and add new ones targeting gaps you discover.

Plan limits

See Plans and usage for full details.