Back to case studies
Case Study · 02

Legal Intake, Triage & Assignment

A production intake pipeline that screens, enriches, triages, estimates, recommends, reviews, and assigns legal enquiries end-to-end.

Architecture

Enquiry to assignment, in one pipeline.

Every enquiry flows through screening, ML classification, forecasting, retrieval, ranking, and a human-in-the-loop review before being assigned to a fee-earner.

legal.intake · triage pipeline
days → minutes
  1. stage 01auto
    Inbound enquiry
    email · form · portal
  2. stage 02auto
    Relevance filter
    ML classifier
  3. stage 03auto
    Case classification
    department + case type
  4. stage 04auto
    Cost & time forecast
    regression models
  5. stage 05auto
    Legal retrieval
    UK laws · case law
  6. stage 06auto
    Fee-earner recommendation
    ranking model
  7. stage 07HITL
    Department head review
    human-in-the-loop
  8. stage 08auto
    Assignment + client update
    structured record
Custom ML inference
Email filter · case classifier · cost/time regression · fee-earner ranker
Retrieval layer
UK laws + court cases indexed for contextual case support at assignment.
Automation surface
Follow-ups · structured records · department alerts · client updates.
Impact

Measured outcomes.

+50%
Onboarding capacity
£300k+
Yr-1 cost reduction
Days → Min
Triage time
How it works

From inbox to fee-earner.

Anonymised summary of the system's stages and design choices.

01

Problem

Inbound enquiries queue up across email and portals. Manual triage takes days, blocks fee-earner assignment, and leaks revenue on missed matters.

02

Screening & classification

Custom ML filters relevant emails and classifies case type and department, reducing reliance on open-ended LLM reasoning and hallucination risk.

03

Forecasting

Regression models estimate cost and time-to-resolution per matter, feeding downstream ranking and department planning.

04

Retrieval

Relevant UK laws and court cases are retrieved at assignment time to give the fee-earner immediate context on the matter.

05

Recommendation & HITL

A ranking model recommends fee-earners; department heads review and approve assignments in a lightweight human-in-the-loop step.

06

Assignment & records

Structured case records, department alerts, and client updates are generated automatically - minutes instead of days.