Building Software for Lawyers: UX Lessons from a Non-Lawyer

I am not a lawyer. I've never been to law school. I've no legal training beyond what I've absorbed through obsessive research over the past six months. And yet, I am building software for the legal industry.

This is either incredibly naive or a genuine advantage. After months of user research and prototyping, I am increasingly convinced it's the latter.

The Curse of Expertise

Most legal tech software is designed by people who deeply understand the law. Former lawyers, legal ops professionals, people who have spent their careers inside the legal system. They build tools that make perfect sense to other legal professionals.

The problem is that LegalAgento isn't for legal professionals. It's for regular people navigating the legal system for the first time. And those people don't think like lawyers.

When a lawyer thinks about a housing dispute, they think in terms of causes of action, statutes, burden of proof, and procedural requirements. When a renter whose landlord is keeping their security deposit thinks about the same dispute, they think: "My landlord took my money. How do I get it back?"

The gap between those two framings is enormous, and most legal tech software operates entirely in the lawyer framing. It uses legal terminology without explanation, presents options without context, and assumes familiarity with processes that are completely foreign to most people.

Being a non-lawyer building for non-lawyers means I don't have to unlearn the expert framing. I naturally think about these problems the way the user does, because I am the user.

Lesson 1: Kill the Legal Jargon

The single most impactful UX decision in LegalAgento has been the relentless elimination of legal jargon from the user-facing interface.

This sounds obvious, but it's remarkably difficult. Legal concepts often don't have clean everyday equivalents. What do you call a "motion for summary judgment" in plain language? "Asking the judge to decide without a trial" is close but loses important nuance.

My approach has been layered communication:

• Primary layer: Plain language that anyone can understand. "Ask the judge to decide your case based on the documents, without a trial."
• Contextual layer: A brief explanation of why this matters. "This can save you time and money, but only works if the facts are clear."
• Technical layer: Available on request. "This is formally called a Motion for Summary Judgment under Rule 56."

Most legal software starts at the technical layer and maybe links to a glossary. LegalAgento starts at the plain language layer and lets users drill down if they want.

Lesson 2: Progressive Disclosure Over Information Dumps

Legal processes are complex. A small claims court case in Massachusetts involves understanding jurisdiction limits, proper service of process, filing requirements, hearing procedures, and potential appeals. Dumping all of this information on a user at once is overwhelming and useless.

LegalAgento uses aggressive progressive disclosure. At any given moment, the user sees only what they need for their current step. When they're filling out the complaint form, they don't see information about the hearing. When they're preparing for the hearing, they don't see information about appeals.

This required restructuring legal processes into clear, sequential stages with well-defined transitions. It's essentially the same design pattern I used in my aviation products: complex information, delivered incrementally, with clear progress indicators.

The interesting discovery was that lawyers often struggle to decompose legal processes this way. They see the entire landscape at once and have trouble isolating individual steps. As an outsider, I naturally chunked the process because I had to learn it step by step myself.

Lesson 3: Confidence Indicators Are Essential

One of the most valuable UX patterns I've developed is what I call confidence indicators. When LegalAgento provides information, it explicitly communicates how confident it's and why.

"Based on Massachusetts General Laws Chapter 93A, your landlord must return your security deposit within 30 days. This is a well-established statute that applies to your situation."

versus

"Whether this clause in your lease is enforceable may depend on several factors specific to your county's court interpretations. I recommend having an attorney review this specific clause before proceeding."

Users need to know when the AI is on solid ground and when it's entering uncertain territory. Legal professionals instinctively add qualifiers and caveats to everything, which paradoxically makes users less confident in the information. LegalAgento is explicit: this part is clear-cut, this part needs a human lawyer.

Lesson 4: Emotional Context Matters

Here is something that no legal tech product I've evaluated accounts for: the emotional state of the user.

Someone dealing with an eviction is scared. Someone going through a custody dispute is angry and heartbroken. Someone fighting wage theft feels powerless. These emotional contexts fundamentally affect how people process information, make decisions, and interact with software.

LegalAgento's conversational layer acknowledges this. Not in a therapy way. It's not a mental health tool. But it recognizes that saying "You have strong legal options here" to someone who feels powerless is as important as explaining what those options are. It validates the user's experience before diving into procedures.

This was something I learned from building consumer products in other domains. In travel tech, I learned that people booking a trip are excited and want the experience to feel aspirational, not transactional. In legal tech, people seeking help are stressed and want the experience to feel supportive, not clinical.

Lesson 5: The Handoff to Humans Must Be Smooth

LegalAgento isn't trying to replace lawyers. It's trying to make legal help accessible to people who currently have none. That means there are moments when the AI needs to hand off to a human attorney, and that handoff needs to be smooth.

The worst possible experience is: "I can't help with this. Please consult a lawyer." That leaves the user exactly where they started, knowing they need help but not knowing how to get it.

Instead, LegalAgento's handoff pattern is: "This specific question requires an attorney's judgment. Here is exactly what you need to ask them, here are attorneys in your area who offer unbundled services for this type of question, and here is the typical cost range. The rest of your case can continue with me."

That pattern (specific, actionable, and continuity-preserving) came directly from my experience building other products. In AvioSharing, when a user's request went beyond what the platform could automate, we didn't just say "contact us." We told them exactly what would happen next and who would help them.

The Outsider Advantage

Six months in, I believe the outsider perspective is genuinely valuable in legal tech. Not because lawyers are bad at building software (they're not), but because the people who need legal software most are outsiders to the legal system. Building from their perspective, with their mental models, in their language, produces fundamentally different products.

The risk of the outsider approach is getting the law wrong. I mitigate this through expert review, extensive testing, and clear boundaries around what the AI can and can't do. But the advantage, building products that actually work for the people who need them most, is worth the extra diligence.

Legal software doesn't need to be designed by lawyers. It needs to be designed for the people lawyers have not been able to reach.