Pharos - effortlessly analyze, store and collaborate on film scripts
PROBLEM
Currently script reading, managing and storing is a tedious and manual task, with few tools targeting this problem area. Pharos aimed to build a web application to store and manage scripts, and incorporate an LLM product to analyze new scripts.
OUTCOME
I delivered competitor research, user journey, and a fully functional prototype. I collaborated with the founding Engineer to implement this product. This product has been shipped and is has 100+ active users.
ROLE
Product designer
TIMELINE
Feb ' 24- September ‘ 24
TEAM
1 Designer (me), 1 Engineer
Research
& Competitor Analysis
Before designing, I jumped into speaking with script writers and exploring key workflows used in:
01
Traditional script coverage services
Sample: WeScreenplay, Script Reader Pro, Avail
02
Google Drive + Notion
Key workspace for collaboration and organization
03
private Studios Internal Tools
Limited Resources
Key Challenges
What are the key challenges faced by those in the film industry?
01
Lack of Centralization
There is no industry specific centralized place to store and analyze scripts
02
Manual Process
Script analyzing, note taking and formatting is extremely time consuming and repetitive
03
Collaboration
Difficult to collaborate with team members asynchronously. Need a place for notes, rating and edits.
04
Status Visibility
Poor visibility into script statuses, genres, or writer contributions. Need for search and status tracking.
Solution
Pharos combines AI-powered script analysis with a collaborative workspace to:
01
Ingest & analyze PDFs using LLMs
02
Break down screenplays into editable sections
03
Enable team-based commenting, note-taking, and review
04
Sort, filter, and manage hundreds of scripts
05
Centralize writer data across uploaded content
Key Features & Design Decisions
We built Pharos around the idea that uploading a script should be the start of insight—not just storage. When a user uploads a script (PDF), a custom LLM model processes the file to extract story structure, themes, characters, dialogue density, and genre fit. This provides an instant breakdown of the screenplay without needing to read 100+ pages.
01
Script Upload & LLM Analysis
-> Users can upload a PDF script.
-> Pharos automatically uses an LLM to extract structure, categorize elements and present it in an editable format.
-> The system saves each script analysis as a “living document” for collaboration.
Design Decision: To keep the experience intuitive, I wanted the upload flow to be minimal: a single-click CTA followed by automatic AI parsing and categorization. The user never sees a “loading” screen—instead, they’re brought directly into a dynamic dashboard of results.
Key Feature: I introduced a “Status” dropdown (Submitted, Reading, Packaging, Passed) to track each script's progress across teams.
02
Interactive Script Breakdown View
-> Clicking on a title opens the full analysis.
-> Content is broken down into categories: Writers, Genres, Logline, Themes, Summary, Character Breakdown, etc
-> Editors can leave inline comments, tag teammates, and mark the script as Passed ✔️.
Design Choice: I prioritized a modular layout so teams could quickly scan insights, or dive deep into specific sectiolns. I had to ensure that we had enough section designs depending on the length and format of the LLM response.
03
Library Sorting and Filtering
-> Large teams deal with hundreds of scripts. To improve navigation, I implemented filters by Genre, Status and Submission Date.
-> Sorting by column helps track high priority or newly added content.
Design Choice: I drew inspiration from tools like Notion and Airtable to keep the table highly readable whille supporting quick actions like download and Status Update. I optimized for team collaboration rather than individual writing.
04
Writer's List
-> The Writers List gives users a bird’s-eye view of all contributing writers, showing known work, associated agencies and and genres they frequently write in.
-> Clicking a writer's name opens a dedicated profile, complete with personal information and a table with their uploaded scripts.
