AI agent for scriptwriters
Coverage Reader Feedback Pattern Agent
A ranked list of feedback themes with evidence from the script, ready for the writer to decide what to change
What it does
Five readers send notes on a script. One loves the opening, two find it slow, and four mention the ending in different words. This agent reads each reader's notes, groups them by theme such as character, pace, structure, dialogue and ending, and finds which points repeat across readers and which stand alone. It checks each point against the script text, for example finding the scene a reader called slow and measuring page count, and marks whether the script supports the comment. For contested points, it asks the writer to confirm what they intended and regroups after the answer. It then ranks themes by how many readers raised them and how closely the script supports them. The writer decides what to apply. Edge case: a note from a reader outside the target audience is shown but weighted lower if the writer says so.
How it works
Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.
Read the steps as a list
- Reader notes arrive
- Read each reader's notes and the script
- Split the notes into individual points and group by theme
- Count how many readers raise each theme
- Check each point against the script text and page references
- Does the script text support each point the readers made?If not: Mark the point as unsupported and look at other scenes the reader may have meant. Back to step 4.
- List contested themes where readers disagree
- Writer clarifies intent on the contested themesThe agent waits here for your OK.
- Regroup and rerank the themes with the writer's answers
- Is every theme ranked with evidence and a clear status?If not: Fill the missing evidence and rank again. Back to step 9.
- Feedback pattern report
How it decides
It counts a theme as strong when 3 or more readers raise it and the script supports the comment, and as contested when readers disagree or the script does not support it.
- Strong theme: 3 or more readers and supported by the script
- Contested theme: readers disagree or the script does not support the point
- Weight readers by the writer's target audience if specified
- Quote page references for every theme
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Themes list
- Reader weighting
- Strong theme threshold (default 3 readers)
- Report format
- How script evidence is shown
What keeps you in control
It always asks you first
- Writer decides what to apply and answers on contested themes
Hard limits
- Never decide what the writer should change
- Keep reader names within the writer's own files
It stops when
- Done: Themes are ranked with evidence and the writer decides what to apply
- Stop: Fewer than 2 readers' notes, so the agent waits for more feedback
Set it up
We guide you through the set-up, step by step
Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.
- One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
- The agent then walks you through connecting your own data, one source at a time
- A downloadable copy with the flow chart, the rules and the full guide