February 25, 2026 · Legal Technology

Why I’m Building the Legal Case Management System I Wish I Had

By Asaf Fulks, Esq.

Originally published February 25, 2026. Updated May 2026 to reflect v1.0.0 production release.

I spent over twenty years running a recording studio before I became an attorney. Since passing the bar, I’ve practiced across a range that most lawyers never touch — lemon law, products liability, personal injury defense, insurance coverage, construction defect litigation, entertainment law, and intellectual property. I’ve handled Song-Beverly warranty cases against automakers, defended premises liability claims, and managed multi-party construction disputes with layered insurance towers. I’ve drafted wills and trusts, advised on wealth management, and put my two decades of music industry experience to work on IP and entertainment matters.

I’ve worked cases at every stage — pre-litigation demands, early litigation strategy, written discovery, depositions, motions, mediations, and trial. I’ve filed in small claims and state superior courts, and handled copyright, trademark, and patent matters from the transactional side. Every practice area had different demands. Every stage had different workflows. And every one of them used the same generic software that couldn’t keep up with any of it.

So I started building my own.

The Problem Nobody Talks About

Every legal case management platform on the market was built by software companies. That’s not a knock — it’s just a fact. And that fact shows up in the product every single day.

When I open a construction defect case, I need fields for Privette doctrine analysis, SB 800 notice compliance, indemnity tracking between general contractors and subcontractors, and multi-layer insurance coverage mapping. When I open a personal injury defense file, I need lien tracking, medical chronologies with treating physician hierarchies, and IME scheduling workflows. When I switch to a lemon law case, I need Song-Beverly warranty tracking, repair attempt logs, and mileage offset calculations.

What does the typical platform give me? A blank text field and a calendar.

That’s not case management. That’s a filing cabinet with a subscription fee.

The tools most attorneys use weren’t designed by people who have actually managed a case from intake to verdict. They were designed by people who interviewed attorneys, synthesized the feedback into a product requirements document, and built something generic enough to sell to everyone. The result is software that technically works for all practice areas but doesn’t truly serve any of them.

What I Wanted to Exist

I wanted a system that understood that a maritime injury case and a trade secret misappropriation case have almost nothing in common — and that the software should reflect that reality. I wanted practice-area-specific data capture across every type of case a firm might handle, not a one-size-fits-all form with custom fields bolted on as an afterthought.

I wanted deadline calculations that actually knew the rules. Not just “add 30 days to this date.” I’m talking about a system with jurisdiction-specific court rules for all fifty states and D.C. — a system that understands the difference between California’s CCP deadlines and New York’s CPLR timelines, that accounts for court holidays and local rules, and that flags hard deadlines versus soft ones without me having to think about it.

I wanted a legal reference library built into the platform — not a separate tab, not a Google search, not a prayer that I remembered the right statute number. I’m talking about thousands of verified legal references with working URLs, covering everything from general civil procedure through specialized areas like bankruptcy, immigration, healthcare compliance, and intellectual property.

And more than anything, I wanted AI integration that didn’t require me to send my clients’ privileged information to someone else’s server.

So I Built It

CaseDesk™ shipped to production in March 2026 — a personal project years in the making, now in daily use in my own practice.

The numbers reflect what comprehensive coverage actually looks like:

That last part matters more than most people realize.

The AI Question

Everyone in legal tech is racing to add AI. Most of them are doing it by piping your case data to a cloud API and hoping their terms of service are good enough to protect attorney-client privilege. I wasn’t comfortable with that approach.

CaseDesk’s AI architecture gives the attorney the choice — every time, on every workflow. The system supports local language models running on the attorney’s own GPU with no data leaving the machine, and it also supports cloud providers like Anthropic’s Claude (including AWS Bedrock VPC deployment for sensitive documents) when the attorney prefers that path. The choice stays with the attorney, not the platform.

When the AI answers a question about your caseload, it’s referencing your actual data — active cases, upcoming deadlines, recent communications, settlement activity — injected directly into the conversation. Not hallucinating case numbers from a generic training set.

And because legal AI hallucination isn’t just embarrassing — it’s potentially sanctionable — the system is engineered from the ground up to stay grounded in the data it can actually see. The 27 MCP integrations also mean external AI assistants like Claude Desktop can query the case database directly through a controlled interface, while CaseDesk itself is closed. That changes what an AI workflow looks like for an attorney who wants real automation, not just a chat sidebar.

Why This Matters for the Industry

I’m not writing this to sell software. CaseDesk is shipped and in use, but commercial distribution to other firms is still ahead. I’m writing this because I think the legal technology industry has a fundamental design problem, and I don’t see enough attorneys talking about it.

The problem is this: the people building legal software don’t practice law, and the people practicing law don’t build software. That gap produces tools that look impressive in a demo and feel frustrating in daily use. Features get prioritized based on what sells, not what actually reduces the cognitive load of managing a complex case docket.

When an attorney is monitoring excess coverage on a construction injury case with a $25 million exposure, they don’t need a prettier dashboard. They need a system that understands what monitoring counsel does — that tracks carrier reporting deadlines, flags reservation of rights issues, maps the coverage tower, and surfaces the information that matters without requiring twelve clicks to get there.

That kind of depth doesn’t come from user surveys. It comes from doing the work.

What’s Next

CaseDesk v1.0.0 is in production and proving itself in my own daily practice. The roadmap now points outward: commercial distribution to other attorneys, multi-device support, and continued refinement of the workflows that matter most. The product is real. The work continues.

If you’re an attorney who’s felt the same frustrations I have, or if you’re interested in the intersection of legal practice and technology, I’d love to hear from you. You can reach me through this site or connect with me on LinkedIn.

The tools we use should be as sophisticated as the work we do. It’s time they caught up.


Asaf Fulks is an associate attorney in California specializing in insurance defense, construction defect litigation, personal injury defense and intellectual property. He is also a software developer and the creator of CaseDesk. He can be reached at asaf@asaffulkslaw.com.