Enterprise engineering, translated for law firms

Before building for law firms, I built workflow, AI and data systems for a Fortune 100 aerospace and defense company. The problems rhyme: multi-stage approvals, audit trails, documents that need reading, and numbers that have to hold up when someone checks them.

$2.5Msaved per aircraft by an AI-driven data pipeline

< 10 sto price a bid that used to take days

±22 minmodel accuracy on 4,687 historical records

30–40fixes and features in under 24 hours of engineering

01 · workflow portal

A multi-stage approval portal with automatic sign-in

An enterprise web portal that coordinates checklists through a seven-stage process with peer reviews and approvals. Tasks can be split into numbered subtasks with parent-child rules, and every answer has a question-by-question audit trail.

Windows sign-in through the company's existing accounts, so nobody types a password twice
A "My Work" queue showing each person only what's waiting on them
16+ related database tables with schema changes applied automatically on startup
30 to 40 user-requested fixes and features delivered in under 24 hours of focused engineering, with no regressions

For a law firm

This is matter workflow: stages, checklists, review and approval, a "My Work" queue for each attorney and paralegal, and an audit trail of who changed what. Firms on Microsoft 365 can get the same automatic sign-in.

02 · ai pipeline

AI-driven measurement that saved $2.5M per aircraft

An automated measurement and data processing pipeline, built with agentic AI tools, that improved wire-cutting accuracy and manufacturing analytics. The result was $2.5 million in savings per aircraft.

For a law firm

The same approach, AI plus careful data handling, is what reads financial documents, pulls figures out of PDFs and flags what's missing, so staff check results instead of starting from scratch.

03 · cost model

Bids priced in under 10 seconds instead of days

A statistical labor cost model trained on 4,687 historical records replaced days of manual engineering estimates. On drawings it had never seen, it was accurate to within about 22 minutes on average, and every factor in the model was statistically significant, so each number could be defended in an audit.

For a law firm

Your own history holds the answers: how long each kind of matter really takes, what it really costs, which matters are profitable. I build the reporting and models that turn that data into fee and staffing decisions.

04 · document extraction

Reading measurements straight out of PDF drawings

A computer vision pipeline that finds the right sheet in a multi-page PDF drawing package, works out its scale from the background grid and measures each cable automatically, with no manual clicks.

For a law firm

Getting structured data out of messy PDFs is the hard part of most migrations and most document review. It's work I've already done at scale.

05 · learning platform & devops

A sector-wide learning platform and automated deployments

Designed and built an enterprise learning management system now deployed sector-wide, which earned a nomination for the company's President's Award. Built automated build-and-release pipelines that test and deploy every change to development, staging and production servers.

For a law firm

Staff training that lives inside your system, and updates that ship safely without anyone staying late to copy files onto a server.

Figures come from internal project reviews. Employer and project details are described generally to respect confidentiality.

Enterprise engineering, at law firm scale

Book an intro call About Jonathan Simon