What ABA TECHSHOW 2026’s Startup Alley Tells Us About Where Legal Tech Is Going

From chatbots to agentic AI in legal technology 

Chatbots are easy to adopt. Lawyers have told us that. But there’s a long way between typing a question into a chat window and getting something you can actually file, send, or hand to a client. You still have to know what to ask, how to sequence the prompts, and how to stitch the outputs together yourself.

What we saw this year was different. Several legal tech AI startups have built beyond the chat model towards what’s being called agentic AI. If you don’t really know what agentic AI is defined as, you can think of it as stringing together tasks dynamically, as first explained that way by one of the analysts at Andreessen Horowitz, of Marc Andreessen fame. We’ve had agents for hundreds of years. A thermostat is one. TurboTax is one. They both string tasks dynamically. Large language models are the latest version, and the startups building on them are getting good fast.

• Lawdify, the third-place winner, was the clearest example. A document dump goes in. Litigation-ready output comes out: draft letters, interrogatories, timelines, profiles of every individual in the documents. Lawyers don’t need to do intermediate prompting or copying and pasting between windows.
• TwinCounsel took a different approach. It creates what amounts to a digital twin of the lawyer, and it operates through email. You delegate to it the same way you’d delegate to a junior associate by describing what you need, and it handles the work in the background. For a lawyer who already spends most of the day in their inbox, there’s nothing new to learn.
• Litmas, CaseCreate, and VoiceScript rounded out the category, each focused on different pieces of the litigation lifecycle. 

The common thread is that, with each of them, lawyers don’t have to orchestrate every step. The platforms chain the tasks together and hand them an output for review. For a deeper look at how AI tools are changing legal practice, explore our AI for lawyers hub.

Why startups are going vertical

Legal tech startups tend to start narrow. Founders with deep expertise in one area of law build the tool they wish they’d had when they were practicing.

LegalBridge pitched a full immigration workflow. EstateScribe and EstateMin both went after estate planning from different angles. And two dispute resolution tools, Immediator and StreamSettle, built shared platforms where negotiating parties can see how much alignment exists before they even sit down.

These startups solve a specific problem for a specific practice area or workflow, with the deep expertise founders have often gained in that area. This brings a lot of benefits, and especially when startups with deep domain knowledge can plug into the systems lawyers already use, everybody wins.

This is part of why Clio built an open API, an application programming interface that lets software connect together securely. It means a startup with deep immigration or estate planning expertise doesn’t have to build an entire practice management platform from scratch. They build the specialized tool, connect it to Clio, and the lawyer gets both in one workflow. With over 300 integration partners now listed in the Clio app directory, that strategy is working. The startup gets distribution. The lawyer gets a tool built specifically for how they practice. And the client gets better outcomes because their lawyer isn’t fighting the technology.

Meeting lawyers where they already work

Adoption is the oldest problem in legal tech. Lawyers don’t have afternoons to spend learning a new interface. The smarter startups have figured this out. Instead of asking lawyers to come to them, they go to the lawyer.

Candle AI is an email assistant that lives inside Outlook, gathering context from your inbox and acting on it. Sonar Legal does one-click house style formatting inside Microsoft Word, which sounds modest until you’ve spent three hours reformatting a brief because Word decided to override your styles. That pain point has existed since WordPerfect lost the word processing wars in the early nineties. There’s a reason WordPerfect files still show up in practice management systems today.

The email-as-interface approach is also telling for a different reason. Your inbox is where your case lives. Client communications, opposing counsel exchanges, court notices, attachments. When AI is only as useful as what it knows about the matter in front of you, tools that can read your inbox have an advantage over tools that ask you to re-explain everything from scratch.

We’re already used to typing messages to people we can’t see and getting responses. Adding an AI on the other end is a natural extension of how the work gets done.

What this year’s winner gets right

The overall winner of Startup Alley was Collbox, a tool that automates accounts receivable for law firms. You’ve billed the client. You’re waiting for payment. Collbox steps in and acts as a collection agent on your behalf.

It won for a reason. Lawyers are far more willing to bill than to collect. It’s an open secret in firm management. A lawyer would rather spend an afternoon generating $2,000 in new billings than pick up the phone and call the client who owes them $10,000. Collbox takes that discomfort and turns it into a system that runs in the background without the awkward phone call. We’ve solved this problem at Clio as well, so it’s good to see Collbox recognizing the same pain point and building for it.

CounselPro may have had the strongest demo of the evening. It’s a financial forensics tool. Drag in a year’s worth of bank statements and it shows you where the money went. In a divorce, that means categorizing spending by spouse, flagging the anomalies (hotels booked in someone’s own hometown, spending patterns that don’t match what’s been disclosed), and producing the financial picture a court needs to determine support and divide assets. It’s immediately useful for any practice area where the work involves following money, which is a lot of them.

CounselPro, Immediator, Collbox. They all go where the money is. A firm will spend a dollar to recover ten. That’s not a complicated pitch, and it’s a smart place for a startup to build.

Where legal tech startups go from here

What struck us most about this year’s Startup Alley wasn’t any single product. It was how specific the tools have gotten. A year or two ago, legal AI startups were mostly showing chatbots with legal training data. This year, they’re showing tools that do particular jobs in particular practice areas and produce particular outputs. The pitches have gotten specific, and so have the outputs. That’s a sign of a maturing industry.

If you want to understand the future of AI in law, this year’s Startup Alley was a preview. The tools that automate your intake, format your briefs, follow your money, and manage your litigation, those tools are here now. And the legal tech startups building them are only getting more focused.

For the startups themselves, the opportunity has never been bigger. If you’re building something worth showing, ClioCon this autumn in Boston is the best place to put it in front of lawyers who are already excited about legal technology and ready to adopt. We provide the booth, the branding, and the technology so startups can show up and focus on what matters: talking to customers. Whether you’re pitching to a solo practitioner who can switch today or a mid-size firm evaluating tools for next quarter, there’s no better room to be in.

And for the firms thinking about AI and the future of law, the clearest lesson from this year’s Startup Alley is that the best legal technology isn’t the tool that does the most but the tool that does the specific thing you need, inside the workflow you already have, without asking you to change how you practice.