General AI Models Aren't Enough for Legal Work, Says Legora CEO
Max Junestrand, CEO of legal AI platform Legora, says building tailored solutions on top of foundational models is the only way to serve law firms effectively. General-purpose AI models and fine-tuning approaches don't work at scale for legal applications.
Legora raised $550 million at a $5.55 billion valuation in its Series D round, backing Junestrand's thesis that the legal sector demands purpose-built tools. The company now serves 800 customers across more than 50 markets.
Why General Models Fall Short
Legal workflows are too complex for off-the-shelf AI. "There was so much application that you had to build on top of the models to make them useful in your environment," Junestrand said.
Fine-tuning general models proved ineffective at Legora's scale. The company instead invested in understanding what foundational models could and couldn't do, then built specific applications on top of that foundation.
This approach differs fundamentally from how traditional software companies operate. AI software firms must continuously track model capabilities and adapt products accordingly, since features can become irrelevant as models improve.
Legal Firms Are Hungry for AI Differentiation
Law firms historically had few software options. That underinvestment created an opening for AI-driven solutions to address problems that were expensive or impossible to solve before large language models existed.
The legal market is also uniquely positioned for AI adoption. Law firms operate in a low-differentiation sector where most competitors offer similar services at similar prices. One firm adopting AI to deliver better work at a lower cost creates pressure on rivals to do the same.
Lawyers, particularly tech-savvy ones, won't accept AI products that merely match foundational models. Any legal AI tool has to outperform what lawyers can already do with ChatGPT or Claude, or firms won't pay for it.
Product Quality Over Speed to Market
Junestrand prioritized engineering and product development over rapid sales growth. Legora spent six months not selling because the company wasn't ready to onboard a thousand lawyers per day reliably.
That decision reflects a core insight: AI software companies live or die on reliability and product readiness. Building a culture around quality matters more than early revenue when the underlying technology is moving this fast.
For product teams working in AI, this means understanding what your models can actually do, building differentiated applications on top of them, and accepting that your roadmap will shift as capabilities improve. Speed matters less than getting the fundamentals right.
Learn more about AI for Product Development and how to structure AI software teams differently than traditional companies.
Your membership also unlocks:
