December 16th, 2025
New

Storytell now generates Magic Suggestions, AI-powered prompt ideas based on the Assets, Labels, and Concepts you’re working with, so you don’t have to guess what to ask next.
Instead of starting from a blank prompt, Magic Suggestions surface relevant questions that help you explore, analyze, and understand your content faster.
As soon as an asset finishes processing, Storytell generates Magic Suggestions based on what’s inside the file.
These suggestions reflect:
The key themes and phrases in the content
The type of question someone might reasonably ask (analysis, explanation, comparison, etc.)
A relevant persona or perspective to guide the question
Clicking Magic Suggestion drops it straight into the prompt bar, everything else about how prompting works stays the same.

Learn more at: Magic suggestions
Magic Suggestions are always scoped to what’s currently in focus.
When an asset is in scope, you’ll only see Magic Suggestions for that asset
When working with a Label or Concept, Magic Suggestions are generated from the most relevant connected content
Suggestions update automatically as assets, Labels, and Concepts change
This keeps ideas relevant and avoids pulling questions from unrelated parts of your knowledge base.
Magic Suggestions rotate and refresh as new ones become available, giving you multiple angles to explore your content.
You’ll also see light attribution (like category or persona) to help you understand why a Magic Suggestion was generated.
ⓘ Tip: If you’re not sure what to ask yet, try cycling through Magic Suggestions to uncover analysis paths you might not have considered.
Knowing what to ask is often the hardest part of working with complex content. Magic Suggestions reduce that friction by turning your existing knowledge into clear starting points, so you can move from ingestion to insight faster, without needing to fully understand every asset upfront.
This helps teams spend less time experimenting with prompts and more time actually learning from their data.
Magic Suggestions are especially useful for:
New users getting oriented in a project or unfamiliar dataset
Researchers and analysts exploring large or dense documents
Cross-functional teams who didn’t create the content but need to work with it
Anyone stuck on “what should I ask next?”