D'Alessio Law Group: Speed-to-Lead Architecture

Published April 2026 | Legal Tech Automation

How I replaced a manual lead triage workflow with a fully autonomous Speed-to-Lead pipeline—cutting lead-to-CRM sync from 48-72 hours to 10 seconds and delivering a 45% increase in qualified conversions for a mid-sized law firm managing 12 simultaneous intake channels.

01 Problem

Legacy Manual Intake Was Bleeding High-Value Leads

D'Alessio Law Group was running high-cost, targeted digital marketing campaigns generating 200+ leads per week. But the intake process relied on email forwarding and manual data entry by paralegals. A prospective client would submit a form at 9:00 AM. An intake specialist would read the email at 11:30 AM. They'd open Clio Grow, create a Contact, create a Matter, assign it to an attorney, and send a follow-up SMS. By the time the lead entered the system, 6-48 hours had passed. Conversion rates for leads contacted after 5 minutes were 10x higher than leads contacted after 5 hours. The firm was hemorrhaging cases to faster competitors.

02 Stack

The architecture uses Make.com as the primary orchestration layer for complex routing logic, with Zapier handling simple webhook passthroughs. Lawmatics serves as the lead intake and marketing automation engine, while MyCase handles back-end case management. Claude 3.5 Sonnet provides semantic analysis on intake forms to determine case complexity and attorney fit. Google Cloud Functions act as custom webhook endpoints for high-volume spike handling.

03 Bottleneck

Human-Keystroke Latency & Duplicate Entry Chaos

Three critical bottlenecks emerged:

1. Parallel Processing Limit: One paralegal could only process ~15 leads/hour manually. During campaign surges, 200+ leads would queue up creating a backlog lasting days.
2. Error Cascade: Manual form transposition caused wrong attorney assignments (criminal cases to family law attorneys) requiring rework and client confusion.
3. No Conflict Checking: The firm had no automated system to detect if a potential client was already represented by the firm or was an adverse party in an existing matter, creating ethical risk exposure.

04 Architecture

Modular Decision-Tree Pipeline with Semantic Routing

The system operates in four stages:

Conflict Resolution Layer: Before creating a new Matter, the system queries MyCase API to check if the email/phone exists in any opposing party role across active matters. If match found, route to managing partner for ethical review instead of auto-assignment.

05 Technical Artifact

Intake Routing Logic Engine - Core Webhook Handler

06 Before / After

07 Metrics

Performance tracked over 90 days post-implementation (6,000+ leads):

• Speed-to-Lead: Reduced from average 14.2 hours to 10 seconds (5,112x faster)
• Qualified Conversion Rate: Increased from 12% to 55% (3.6x improvement)
• Routing Accuracy: 99.2% correct attorney assignment (vs 87% manual)
• Paralegal Time Savings: 22 hours/week freed for high-value client work
• Conflict Detection: 47 potential ethical violations prevented in first 6 months
• Cost per Lead Processing: Reduced from $18.50 to $0.42

08 What Broke (Lessons Learned)

Three critical failures during implementation:

1. Webhook Flood on Campaign Launch: Our first major campaign triggered 847 leads in 4 hours. Make.com rate limits hit 429 errors. Fix: Implemented Google Cloud Functions as buffer layer with exponential backoff retry logic and queue-based rate limiting.
2. Claude Token Overflow: Rich text intake forms with pasted legal documents exceeded Claude's context window, causing silent truncation. Fix: Added pre-processing step to truncate and summarize long-form context before API call, prioritizing key fields.
3. MyCase API Field Mismatch: Custom Lawmatics fields didn't map 1:1 to MyCase, causing data loss. Fix: Built field translation dictionary and implemented fallback logic to store unmapped data in Lawmatics custom fields.

09 What I Built

The complete Modular Intake Orchestrator consists of:

• Make.com Scenario: Main routing engine with 47 decision nodes handling form validation, spam scoring, attorney availability checks, and CRM synchronization
• Claude Prompt Template: Enforced JSON schema for deterministic output: {complexity, damages_estimate, primary_issue[], jurisdiction, conflict_risk, seniority_required}
• Google Cloud Functions: Three serverless functions - webhook validator, queue manager, and conflict checker using MyCase REST API
• Conflict Database Schema: Supabase instance tracking parties, attorneys, and adverse relationships with real-time lookup API
• Attorney Availability Matrix: Real-time calendar integration checking OOO status, current caseload, and specialty matching
• SLA Timer System: Automated escalation to managing partner if lead not contacted within 15 minutes during business hours

10 Related Systems