Neomatter logo

Legal and professional services · YC F26

Neomatter: The Legal Work Is Done Before You Open It

Meet Neomatter, the AI that picks up every request a company sends to its in-house legal team, does the work and hands counsel a decision to approve.

Neomatter’s public homepage introducing its AI workspace for in-house legal teams
Neomatter’s public homepage.Neomatter · Official website ↗

Every part of a company sends work to legal. Sales sends contracts, HR sends employment agreements, finance needs board approvals, and every team has questions. [2]

Much of it is repetitive: reviewing the same terms, finding old agreements, answering familiar questions and chasing information. When the team runs out of hours, the company hires another lawyer or pays an outside firm. [2]

Neomatter does that work first.

It plugs into the places requests already arrive, such as Gmail, Outlook, Slack, Teams and Salesforce, and picks up each one as it comes in. It gathers the context, applies the company’s playbooks and past decisions, and hands counsel completed work to approve. [1][2][3]

Instead of starting from a blank page, the lawyer starts from a decision. [1]

A matter that arrives ready for a decision

Neomatter routes each request to the right lawyer, with no tagging and no new tool to learn. [2]

Then it does the work. It reads everything, marks up the contract, checks it against the team’s playbooks and precedents, drafts the reply and waits for the lawyer’s call. The lawyer checks the sources, makes any edits and decides. [2]

Each matter shows what Neomatter did and what’s still open. Some need a decision, some have a reply ready to send, and some are waiting on another team or the other side’s counsel. [1]

It handles commercial, employment and corporate work and everyday questions, from a reseller agreement to a director’s written consent. [1][2]

Built by founders who talked to 200 legal teams

Luke Wagner cancelled plans for medical school to move to San Francisco and start Neomatter. [3]

Luke Mainwaring left a computer science degree and was previously CTO of DAHLIA Intelligence, which built an agentic personal assistant for students. [3]

The two talked with more than 200 in-house legal teams and came away convinced that AI could change how lawyers work. [2]

Neomatter is based in San Francisco and is part of Y Combinator’s Fall 2026 batch. [3]

How an in-house team would use it

Here’s how an in-house legal team would use it.

A customer pushes on the liability cap

A customer in an expansion deal wants a $10 million liability cap and uncapped indemnity for AI output. Neomatter works through the playbook and proposes a defensible package: keep the fees-paid cap and add one $2.4 million super-cap. [1]

Counsel chooses the next step: authorize the playbook fallback, ask the CFO to approve $5 million, or hold the current position. [1]

A director needs a written consent

A UK subsidiary is appointing a new director. Neomatter drafts the written consent from last year’s, checks the articles, confirms both directors must sign, and drafts a reply noting that the appointment must be filed within 14 days. [1]

Can we use their logo?

Marketing wants a customer’s logo for a launch blog post. Neomatter finds that the customer approved it by email and that the contract doesn’t require a signed form, so the email is enough. The reply is ready to send. [1]

Where Neomatter ranks

View the comparisons and methodology

Most legal AI waits for a lawyer to ask. Neomatter picks up the request before it reaches the lawyer. [3]

The bigger picture

In-house lawyers are hired for their judgment. Much of their day goes to assembling context instead. [2]

When they run out of hours, the work goes to new hires or outside counsel. [2]

Neomatter is betting that AI can do the assembly, inside the company’s own playbooks, so lawyers spend their time deciding.

The work is done. The decision is yours.

Written by SOTA2 from Neomatter’s public materials. Customer experiences and product plans are reported by Neomatter. Ranking analysis and comparison data are from SOTA2.