Skip to content
Uncapped with Jack AltmanUncapped with Jack Altman

Building an AI-Native Software Company With Legora CEO Max Junestrand | Ep. 44

At 23, with no legal background, Max Junestrand co-founded Legora to transform how lawyers work. Legora recently (March 2026) raised $550 million at a $5.55 billion valuation in a Series D funding round to accelerate its expansion across the United States. Over the past year, Legora has grown from 40 to 400 team members across the globe and the platform supports tens of thousands of lawyers each day across 800 customers in more than 50 markets. Max shares the story of building Legora, what it really means to build AI-native software from day one, why legal work is uniquely suited for AI, and how a small team from Stockholm convinced some of the world’s largest law firms to change how they work. Timestamps: (0:00) Intro (0:31) Legora's origin story (9:05) Building an AI-native company (18:16) No sacred cows, the models will be amazing (27:36) Winning pilots and global expansion (36:43) Starting in Europe (47:15) Stockholm culture and "blodsmak" Links: https://x.com/MaxJunestrand https://x.com/chetanp https://x.com/jaltma https://legora.com/ https://uncappedpod.com/ friends@uncappedpod.com

Max JunestrandguestJack Altmanhost
Mar 11, 202649mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Legora CEO on building an AI-native legal software juggernaut fast

  1. Legora bet early that foundation models would rapidly improve, so differentiation would come from application-layer workflows, guardrails, and enterprise-grade trust rather than training bespoke models.
  2. The company built an engineering- and research-led organization that expects features to become obsolete quickly and optimizes for speed, low ego, and continuous re-learning.
  3. Legora’s initial breakthrough came from narrowing focus to a few high-value legal workflows (e.g., tabular review and deep Word/Outlook integration), leading to rapid revenue doubling and U.S. expansion.
  4. Go-to-market relied on frictionless competitive pilots and forward-deployed “legal engineers” to deliver immediate, measurable value that created strong usage-based stickiness.
  5. Starting in Sweden forced multi-country and multi-language readiness from day one, enabling parallel global expansion while maintaining a distinctive Stockholm-centered culture (#blodsmak).

IDEAS WORTH REMEMBERING

5 ideas

In vertical AI, assume foundation models will outpace bespoke tuning.

Legora rejected the 2023–2024 “train your own model” playbook, arguing fine-tuning didn’t work well enough at their scale and that real value was in compliance, ingestion, parsing, citations, and workflow design around fast-improving models.

Differentiation shifts from model IQ to trustworthy execution environments.

Max notes a “flip” where models are no longer the bottleneck; the hard work is building software that lets models execute with the right tools and lets humans review outputs with confidence (auditability, UI, workflow controls).

Build for both human users and agent users.

Legora treats new features as dual-use: humans may initiate or review, while agents increasingly operate tools like tabular review and document editing directly—changing product requirements and interface priorities.

An AI-native company must be comfortable deleting major work.

They expect capabilities to become native to the next model release, so the organization optimizes for humility and adaptability—shipping fast, then ripping out scaffolding when the model makes it obsolete.

Roadmaps compress when models “flip”; execution cadence becomes daily.

Instead of long-term planning, Legora recalibrates continuously as new model behavior unlocks new product primitives; this requires strong evaluation infrastructure and fast decision loops.

WORDS WORTH SAVING

5 quotes

For two reasons we were like, 'Fuck that,' right?

— Max Junestrand

If AI can do something, it will do it.

— Max Junestrand

The models are now intelligent enough where they're no longer the bottleneck.

— Max Junestrand

Legora has two users. It's human users and agent users.

— Max Junestrand

At Legora, we wake up with a metallic taste of blood in our mouths.

— Max Junestrand

Why not to fine-tune legal-specific modelsApplication-layer differentiation: RAG, citations, context, trustAI-native org design: engineer-led, minimal product roadmapKilling features as models improve (low-ego culture)Due diligence at scale: tabular/matrix review paradigmCompetitive pilots, forward-deployed legal engineers, migrationsEurope-first global expansion and Stockholm culture

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome