Use AI
The Use AI column type runs prompts on every row in your table. Fast, low-cost, and perfect for transformations that don't require external research.
Need multiple fields from one prompt?
Use JSON Output Mode when one Use AI run should return multiple related fields such as subject and body.
How it works
We take data from your table and generate new content based on your prompt. Single step - no web searches, no article fetching, just your data in and output out.
| What it can do | What it can't do |
|---|---|
| Summarize text in your table | Search the web |
| Generate pitch angles from existing data | Read external articles |
| Reformat or transform content | Multi-step research |
| Score or classify journalists | Access tools beyond text generation |
Need research capabilities?
If your task requires searching the web or reading articles, use Medialyst Agent instead.
Adding a Use AI column
- Open your workflow table
- Click Add Column → Use AI
- Write your prompt using
{{Column Name}}variables - Run on a few rows to test
- Run on the full table
Variables
Reference any column in your table using double curly braces:
{{Name}}— Journalist name{{Publication}}— Publication name{{Email}}— Email address{{Recent Articles}}— Recent articles (if enriched){{Any Column Name}}— Any column you've added
Example: Pitch angles
The most common use - generating pitch angles that reference a journalist's actual coverage:
Based on this journalist's recent articles, suggest 2-3 pitch angles for {{Story}} that would resonate with their coverage patterns.
Inputs:
- Journalist: {{Name}}
- Publication: {{Publication}}
- Recent Articles: {{Recent Articles}}
- Story: {{Story}}
Output Format:
- Angle: One-line pitch angle
- Why it fits: Connection to their recent coverageSample output:
Angle 1: Why mid-market teams adopt this faster than enterprise
Why it fits: Their recent pieces focus on practical ROI over trend narratives.
Angle 2: The operational mistake this launch removes for PR teams
Why it fits: They consistently cover workflow bottlenecks and tooling shifts.Example: Subject line generator
Write 3 email subject lines for pitching {{Story}} to {{Name}} at {{Publication}}.
Keep them under 50 characters. Make them specific to their beat.
Recent coverage themes: {{Recent Articles}}Example: Coverage classifier
Classify this journalist's primary beat based on their recent articles.
Journalist: {{Name}}
Recent Articles: {{Recent Articles}}
Output one of: Tech, Business, Lifestyle, Finance, Health, Other
Include a confidence score (High/Medium/Low) and brief reasoning.Chaining columns
Chain multiple Use AI columns together. Each column references outputs from the previous one. You can also use formulas for more complex data transformations:
- Pitch Angles — generates angles from
{{Recent Articles}} - Best Angle — picks the strongest angle from
{{Pitch Angles}} - Subject Line — writes a subject line using
{{Best Angle}}
Pitch Angles → Best Angle → Subject LineEach downstream column references the output of the previous one.
If you want one prompt to produce multiple related values at once, use JSON Output Mode instead of chaining extra AI columns.
Tips for better prompts
- Be specific about output format - tell the model exactly what structure you want
- Add constraints - "no generic statements", "must reference recent coverage"
- Keep prompts focused - one task per column works better than mega-prompts
- Test on 5-10 rows first - refine before running on your full list
Save prompts to library
When you've written a prompt that works, save it for reuse:
- Open the prompt editor
- Click Save to Library
- Give it a descriptive name (e.g., "Pitch Angle Generator")
- The prompt is now in your library
Next time you create a Use AI column, insert saved prompts instead of writing from scratch. Your team can access shared library prompts too.