Stilta: The Swedish Legal AI Startup Using AI Agents to Transform Patent Litigation

What Stilta Does

Stilta builds agentic AI software for patent practitioners, focusing initially on patent litigation and intellectual‑property analysis. Instead of a single AI assistant, the system uses specialized agents designed to handle different parts of complex patent workflows.

The platform is intended to help law firms, in‑house legal teams, and IP professionals perform research and analysis that traditionally requires hours—or days—of manual document review.

According to the company and early coverage of the launch, Stilta’s tools can automate several key tasks involved in patent disputes and IP strategy.

Patent Litigation Tasks the AI Automates

Stilta’s AI agents are designed to assist with multiple stages of patent analysis, including:

• Patent infringement analysis: identifying whether a product or technology likely violates an existing patent.
• Patent invalidity analysis: searching for prior art and evidence that could invalidate a patent claim in litigation.
• Prior‑art searches: scanning large volumes of technical documents to locate earlier inventions related to a patent.
• Claim analysis: mapping evidence and technical details directly to patent claims.
• Document drafting: helping generate legal or analytical reports used in patent disputes.
• Portfolio analytics: evaluating large patent portfolios to identify strategic value or licensing opportunities.

The system produces source‑backed analysis, meaning evidence and citations are attached to the results so legal teams can verify findings before using them in cases.

Founders and Background

Stilta was founded by Oskar Block, Tobias Estreen, Petrus Werner, and Oscar Adamsson, engineers who previously worked at McKinsey’s AI division, QuantumBlack.

The company launched in December 2025 and joined the Y Combinator Winter 2026 cohort soon afterward, positioning itself in the rapidly growing legal‑AI sector.

Although some startup listings reference San Francisco, most reporting identifies Stockholm, Sweden, as the company’s headquarters and primary base.

Why Investors See a Big Opportunity

Patent litigation is widely considered one of the most data‑heavy and expensive legal processes. Lawyers must analyze technical documents, research prior art, and build detailed arguments that map evidence to specific patent claims.

Missing a key patent or piece of prior art can have enormous consequences, with disputes sometimes involving billions of dollars in damages or licensing value.

Much of the underlying work—reading documents, searching databases, and building analytical arguments—remains manual and time‑intensive, making it a natural target for automation.

Stilta’s pitch is that AI agents can act like a team of junior analysts, rapidly scanning large datasets and producing structured, auditable research.

Beyond litigation, investors also see potential in portfolio and licensing analysis. Many companies and research institutions hold patents that may have licensing value, but identifying who might be using those technologies—and gathering evidence—has historically been expensive and difficult to do at scale. AI systems that automate that discovery could unlock new revenue opportunities from dormant IP assets.

A Growing Wave of Specialized Legal AI

Stilta’s launch reflects a broader shift in legal technology. Early AI tools focused mostly on document drafting or general legal assistance, but newer startups are targeting specific, high‑value workflows where accuracy and domain expertise matter.

Patent litigation fits that pattern well: it is highly technical, document‑heavy, and costly, making it an attractive niche for specialized AI systems.

By focusing narrowly on patent analysis and litigation workflows, Stilta is attempting to build a category‑defining platform for IP professionals—one where AI handles the research groundwork while lawyers focus on strategy and argument.

If that model works, patent law could become one of the earliest legal specialties significantly reshaped by agentic AI tools.