The Access Gap Isn’t a Lawyer Shortage. It’s a Triage Problem.

Most conversations about the access-to-justice gap are arguments about supply: not enough lawyers, not enough funding, not enough free hours. After decades of that argument, the gap is roughly the same size it has always been. There is a reason. The diagnosis is wrong.

A Tuesday in the 92%

Picture someone opening a debt-collection letter on a Tuesday. The letter says they have thirty days to respond. They don’t know what “respond” means in this context. They don’t know if the debt is real, already paid, or so old the law no longer cares. They don’t know whether answering it the wrong way will make things worse than not answering at all.

So they Google. They find document templates and don’t know if they need them. They find lists of lawyers and don’t know how to pick one — or what kind of lawyer even handles their problem. They find an AI chatbot that confidently answers questions they didn’t know to ask. They read three forum threads from 2017, none of which match. They close the laptop and tell themselves they’ll deal with it tomorrow.

They become part of the 92%.

The Legal Services Corporation’s 2022 Justice Gap Report found that 92% of the substantial civil legal problems faced by low-income Americans get no or inadequate legal help. The reflexive read of that statistic — “92% can’t afford a lawyer” — isn’t exactly wrong. But it misses what actually happened to the person above. They didn’t fail to afford a lawyer. They never got far enough to know they needed one.

That is the access gap as it actually exists. And it is not a lawyer shortage.

The decades-old misdiagnosis

The mainstream framing of access to justice has been remarkably stable for a long time. The diagnosis: there aren’t enough affordable lawyers. The prescription: more legal aid funding, more pro bono hours, more allowances for paralegals to do work that’s currently lawyer-only, more law school clinics, more court-run self-help centers.

These are good things. They have not closed the gap.

Federal funding for the Legal Services Corporation — the largest single funder of civil legal aid in the country — was $560 million in fiscal year 2024 and again in fiscal 2025, against tens of millions of income-eligible Americans. LSC’s grantee organizations report turning away roughly one out of every two eligible people who come to their doors, simply because they don’t have the resources to help. Those numbers haven’t shifted meaningfully in years.

It’s worth asking why an effort this sustained, with this much goodwill behind it, has produced this little movement. The honest answer is uncomfortable: the diagnosis is a true description of a different problem.

There aren’t enough lawyers to serve every individual legal need at the prices individuals can pay. That’s a built-in limit, not a temporary one — the math of small-dollar legal matters doesn’t work for any business model that puts a billable lawyer on every step. You can’t fund your way past that math.

But most people in the 92% never get to the point of needing a lawyer’s hour at all. They fail at an earlier step. They fail to figure out what their problem actually is in legal terms, whether it’s worth pursuing, what kind of help they’d even ask for, and whether the time-pressured letter in their inbox is a real signal or a bluff.

That is the bottleneck. It is not legal services. It is the layer before legal services.

What the bottleneck actually is

Think about what an average person actually needs to know in the moment they realize they may have a legal problem. The questions are not the questions a lawyer answers. They’re the questions the person at the front desk does — the one who decides whether you should even be sitting in the lobby:

Do I have a real claim, or am I just upset?

Is this worth pursuing? What’s the realistic outcome, and what will it cost me — in time, money, and stress — to find out?

What kind of lawyer handles this? Most people don’t know that “employment law,” “labor law,” and “civil rights” are different specialties with different bars. They don’t know which kind of lawyer fits a story they can’t yet tell in legal terms.

Is my situation normal, or unusual? Is the demand being made of me a standard one, or an aggressive one? Is the deadline real, or invented to scare me into compliance?

What’s my window? Statute of limitations. Response deadline. The narrow stretch of time during which the evidence still exists and the law still cares.

None of these are legal advice in the sense the regulators mean. They’re triage questions. The legal field already has a name for the layer they live in — “advice and counsel,” or just intake. It’s the most overstretched part of every legal aid clinic, the desk where the line forms and stays long.

And it’s the layer almost no consumer-facing product has ever attempted to build, because it doesn’t look like a product. It looks like a conversation.

Why every existing product misses it

Walk down the list of consumer legal products that exist today, and notice what each one assumes the user has already done.

Online document-template services — the fill-in-the-blank category — assume the diagnosis is done. You already know you need an LLC operating agreement, a residential lease, or a power of attorney. They optimize the form-filling step. Useful, if you’ve already triaged.

Attorney directories and lawyer-matching marketplaces assume the diagnosis is done. You already know what kind of lawyer you need. They optimize search and matching. Useful, if you’ve already triaged.

General-purpose AI chatbots will answer whatever question you ask. They don’t walk you through structured discovery. They don’t know what to ask next, or what you should have asked but didn’t. A person who can’t describe their situation in legal terms can’t get a useful answer from a tool that requires them to ask the right question.

Single-purpose legal automation tools — the ones built around one specific problem like a debt-collection lawsuit, a single-form immigration application, or a particular landlord-tenant scenario — solve that problem well. They’re no help to anyone whose problem isn’t exactly that one.

Legal aid intake — the one layer that does the diagnostic work properly — does it well, and is fundamentally constrained by people. It can’t scale to the size of the gap because it’s humans doing patient, individualized work in offices that close at five.

The pattern is consistent. Every existing product either assumes triage is done, skips it to optimize a downstream step, or tries to do it at the speed humans can manage. Nobody is doing the diagnostic layer at the scale the gap requires.

Why this technology, for this bottleneck, now

The case for AI in legal is usually argued in the abstract: AI is powerful, AI will help, AI is the future. That’s not a case. Every category that was “the future” twenty years ago and never arrived can tell you that.

The case for AI doing the triage layer specifically is concrete. It rests on three things modern AI happens to be good at, which line up with three things the existing system is bad at.

First: patience and depth. An AI assistant can ask thirty follow-up questions and never get tired or annoyed. A human intake worker on a Wednesday afternoon can’t. The kind of structured back-and-forth a good legal intake takes — twenty to forty questions, each one shaped by the previous answer — is exactly what AI is best at, and exactly what humans can’t sustain at scale.

Second: situational awareness. A modern AI model can hold someone’s specific facts in mind, tell similar-sounding situations apart, and know when the next question matters. The difference between “I was fired,” “I was fired after I reported safety violations,” and “I was fired after I reported safety violations and there was a verbal threat last week” points to three different paths. A general chatbot blurs that. An intake-shaped tool doesn’t.

Third: plain-English translation. The gap between “what happened to me” and “what part of that the law cares about” is exactly what these tools are best at — taking messy human stories and reframing them in the right shape. That happens to be the single biggest thing the existing system can’t do at scale.

This isn’t AI getting good at law. It’s AI getting good at the kind of patient, structured conversation that triage actually is. That’s new.

Why fixing triage rewires the rest of the system

The reason the triage layer is interesting isn’t just that it would help individual users — although it would. It’s that everything else in consumer legal sits downstream of it. Fix triage, and the shape of everything below it changes.

Better triage produces better attorney matching. Someone who walks into a consultation with an organized story, a tentative legal framing, and the right kind of lawyer in front of them is a different client from someone who walks in confused. The first hour of that meeting — the one currently spent doing intake the client could have done with software the night before — is recovered. Lawyers can take cases at a lower cost per case because the inputs are cleaner.

Better triage produces a different funnel for lawyers, too. A pre-organized, specialty-matched case is more economically viable to take than a cold one. Cases that previously sat below the threshold for any private attorney to bother with become viable when the intake cost has been removed. The economics of consumer legal services start to change.

And triage is the one place in the system where genuine 100x scale is possible. You can’t 100x lawyers. You can’t 100x pro bono. You can’t 100x legal aid funding without a political miracle that hasn’t arrived in decades of trying. You can 100x software. The bottleneck that has held the gap constant for years is the one place real scale is finally possible.

The reframe

Access to justice has been a moral cause for a long time. It deserves to be one. But the reason the gap has stayed open isn’t that the cause has been insufficiently moral. It’s that the cause has been pointed at the wrong layer.

The gap is not a shortage of lawyers. It is a shortage of the diagnostic layer that decides whether someone needs a lawyer in the first place. That layer was not previously buildable at scale. It is now.

The 92% statistic, read correctly, is not a story about people who couldn’t afford counsel. It’s a story about people who couldn’t enter the system at all. The fix isn’t at the lawyer end of the funnel. It’s at the front door — the part of the system that, until recently, didn’t really exist.

That’s the part worth building. That’s what we’re building at CC My Attorney — a privacy-first AI legal research assistant whose entire job is the triage layer the system has been missing. Not a robo-lawyer. Not a substitute for counsel. The front door.

Try it at ccmyattorney.ai.