Y Combinator has accepted four legal technology companies into its summer 2026 cohort, the incubator confirmed this week. The group spans firm-specific AI training systems, plaintiff-side practice tools, business development agents, and workflow automation - a signal that YC sees durable demand for AI products built around lawyers' actual workflows.
Here's what each company does, and where the opportunities and open questions sit.
Perceptron ML: Private AI trained on firm data
Perceptron ML builds AI systems trained on each law firm's own matters and deployed privately in the firm's environment. The tools cover timekeeping, intake, matter monitoring, research, discovery, and drafting.
The company says its "grounding engine verifies every fact against a primary source before a model can cite it, which is why firms trust us with work that off-the-shelf legal AI keeps getting wrong." That positioning taps the growing interest among law firms in "AI sovereignty" - owning and controlling the models that handle their knowledge. For lawyers exploring how to apply these systems in their own practice, resources like this AI for Legal overview can help separate practical applications from vendor claims.
The pitch is straightforward: off-the-shelf legal AI tools make mistakes on firm-specific work, and Perceptron's approach aims to fix that by grounding every output in verified sources. Whether firms will pay for that control at scale is the open question.
Erinys: AI-native plaintiff-side network
Erinys describes itself as "building the first AI-native plaintiff-side litigation law firm network. We use sophisticated technology to deliver high-quality legal representation faster, more affordably, and to more people."
The company is pulling together several trends at once. Plaintiff-side firms in no-win, no-fee cases benefit directly from any cost reduction, because lower work costs improve the economics of every matter. Smaller firms and solo lawyers increasingly want AI tools but lack the time to build their own platforms. Erinys appears to offer the AI infrastructure rather than a true network of affiliated firms - the founders haven't developed the network concept in their public materials.
Osmaura: Agents for business development
Osmaura's agents scan the web to identify new client and cross-selling opportunities for law firm partners. The platform tells lawyers who to contact, why now, and what to say.
"Partners are trained to practice law - not to prospect, sell, and manage pipelines," the company said. "We're here to help them grow while they focus on what they do best."
This is a new approach to a well-developed category: law firm CRM and business development automation. The agentic angle lets the system dig deeper into firm data and surface opportunities that static CRM tools miss. The recurring challenge for any CRM remains unchanged - lawyers must actually log their key data for the system to work. The agentic layer doesn't solve that problem, but it makes the payoff for logging data more immediate.
Async: Automation across regulated industries
Async builds AI agents for law firms, healthcare clinics, and real estate companies. The company says its agents "automate complex work that's never been economical to staff with humans, letting us capture labour spend that didn't previously exist."
The three target industries share meaningful characteristics: all are regulated in the US, all handle confidential information, and all involve complex documentation. That commonality suggests Async is betting on a repeatable playbook across sectors rather than a bespoke solution for each one.
Why this matters for legal professionals
The cohort's diversity is the takeaway. Perceptron targets larger firms with control and verification; Erinys focuses on plaintiff-side economics; Osmaura addresses business development; Async goes after back-office workflow. For lawyers evaluating AI tools, the practical question is which model fits your firm's structure - not which vendor has the most impressive demo. Perceptron is likely to draw the most immediate interest from larger firms given the push toward AI sovereignty, but the others address real pain points that don't require enterprise scale. Legal professionals looking to build these skills can start with an AI Learning Path for Paralegals to understand how these tools apply to document review and compliance work.
Y Combinator's continued investment in legal tech suggests the category is moving beyond experimental pilots into repeatable business models. The startups that succeed will be the ones that prove they can reduce costs or generate revenue in ways firms can measure.
Your membership also unlocks:
