Case Study 04
Clio Legal
Intake Automation
Industry
Personal Injury Law (NY)
Integration
Clio Manage + Make.com
Scope
Full pipeline — PDF to populated matter
Review time (was 25-40 min)
Automated Clio operations
Claude Vision extraction
Data enters Clio without human review
The Problem
25 minutes of manual data entry for every new case — and mistakes that could misdirect an entire file.
A personal injury law firm in New York was spending 25-40 minutes per intake: reading police report PDFs, manually entering client and defendant information into Clio, calculating statute of limitations dates, generating retainer agreements, creating calendar entries, and emailing clients.
With 10-15 new reports per week, that's 5-10 hours of paralegal time on pure data entry. Worse, the manual process created inconsistency — some matters had incomplete custom fields, some clients received the wrong Calendly link, and occasionally client and defendant were swappedin the system. In legal operations, that's not a minor bug. It's a workflow-breaking mistake that can misdirect an entire case file.
The Architecture
Three layers. PDF in, populated matter out.
Three-Layer Pipeline
Extraction
Verification
Automation
Design Principle
Nothing enters Clio without human review. AI accelerates, humans verify.
Hard Problems Solved
The details that separate working from demo-ready.
Client vs. Defendant Identification
Police reports don't label parties as "client" or "defendant." Solved with dual-signal approach: narrative fault indicators ("struck by," "failed to yield") combined with filename-based identification using the firm's naming convention (SMITH_JOHN_v_DOE_JANE).
Handwritten & Scanned PDFs
NYC MV-104 reports vary wildly — handwritten fields, poor scan quality, multi-page layouts. Claude Vision handles all formats without format-specific parsing rules. No OCR pipeline needed.
Retainer PDF Timing
Clio's document automation requires a delay after triggering before the generated PDF is available. Added a 15-second buffer in Make.com to ensure the retainer is ready before the email module attempts to fetch and attach it.
Derived Field Propagation
Statute of limitations (accident date + 8 years), claim type classification, and seasonal Calendly links all recalculate live as the paralegal edits any field. Corrections propagate immediately.
Matter Deduplication
A Matter ID field allows the automation to update existing matters rather than creating duplicates — reflecting real workflows where matters may be opened before the police report arrives.
Tech Stack
Flask
Web app + API layer
Claude Vision API
Sonnet — PDF extraction
Make.com
Clio API orchestration
Clio Manage
Practice management system
Gmail API
Client email with attachments
Calendly
Seasonal booking links
Render
Production deployment
JavaScript
Live form recalculation
Design Decision
Why Make.com instead of direct API integration?
Law firms change their workflows frequently. Make.com lets a non-developer adjust the automation sequence, add conditional logic, or swap out modules without touching code. The blueprint is included for full transparency and reproducibility.
This was a deliberate choice: maintainability by the legal team over engineering elegance. The goal was not to build the most technically impressive system. It was to build the system that a real firm would actually use.
The Impact
What took a paralegal 25-40 minutes now takes under 2 minutes of review time. For a firm processing 10-15 new reports per week, that's 5-10 hours of paralegal time freed up every week.
Every matter gets the same complete set of custom fields populated. Every client gets the same professional email with the retainer attached on the same day. No more forgotten statute of limitations calculations. No more wrong Calendly links.
Time reduction per intake
Cases processed per week
Fields populated consistently
Next project