From: redpointai
The legal industry is undergoing a significant transformation due to the rapid advancement and application of artificial intelligence. Companies like Lorra are at the forefront of this shift, demonstrating how AI can streamline complex legal processes and redefine the roles of legal professionals.
Lorra: A Pioneer in Legal AI
Lorra, led by CEO and co-founder Max Unistron, is a portfolio company that applies AI to the law industry, working with many of the top law firms globally and having raised over $100 million [00:00:23]. It is recognized as one of the fastest-growing AI applications available [00:00:35].
Max Unistron, a former engineer, observed a lack of software development in the legal space prior to founding Lorra [00:04:27]. This observation highlighted a unique opportunity for AI application development in the sector.
Evolution of AI in Law
When Lorra started, AI models were nascent, with early BERT models from Google being “decent in English, horrendously bad in Swedish” in 2020 [00:01:40]. The arrival of GPT 3.5 marked a “paradigm starter,” moving the industry from full experimentation to implementing practical, end-to-end solutions [00:01:52].
Current State of AI in Law
Today, the focus has shifted from mere experimentation to implementing tangible, end-to-end work deliverables. For example, due diligence can now be performed by uploading documents to a platform like Lorra, which identifies findings and generates reports, eliminating the need for physical data rooms or traditional software like F [00:02:02].
AI Capabilities and the Future of Legal Work
The advancements in AI models, coupled with surrounding frameworks like function calling and tool calling, are significantly changing the legal landscape [00:02:53]. The legal software space, traditionally “incredibly fragmented” with separate tools for translation, comparison, searching, and reviewing, is now seeing all these functions consolidated [00:03:05].
AI can automate tasks across a spectrum of complexity in legal work:
- Bottom Quartile (Simple tasks): Fully automated, including data extraction and simple finding [00:03:39].
- Top Quartile (Complex tasks): Gradually being automated, such as drafting share purchase agreements [00:03:34].
Legal professionals must determine where their specific expertise, context, and instruction for AI models add the most value, and where “off-the-shelf LLMs” are sufficient [00:03:50].
Why Law is Suited for AI
The legal field is uniquely well-suited for AI due to several factors:
- Lack of Prior Software Development: Historically, there hasn’t been significant software development in this space, often due to industry-specific incentives that didn’t align with efficiency [00:04:34].
- Repetitive Tasks: In-house counsels frequently handle repetitive tasks like NDA reviews and MSAs [00:05:02].
- Categorizable Workflows: Legal work can be broadly categorized into reviewing, reading, drafting, writing, or researching [00:05:20].
- AI’s Cross-Functional Capabilities: AI can perform tasks across this entire stack, enabling platforms like Lorra to emerge as comprehensive solutions rather than point solutions [00:05:36].
Client Pressure and Market Shift
“Law firms want to lead in this new paradigm… clients are putting pressure on law firms to make this shift because they’re starting to use tools like this internally.” [00:05:54]
The legal business is seeing a significant shift where software is a $20 billion market, while legal services is a trillion-dollar market [00:06:43]. This creates immense opportunities for AI to blur the lines between software and service.
Addressing Challenges and Fostering Adoption
The traditional “hourly billing problem,” where increased efficiency could lead to less billable hours, was once thought to hinder software adoption [00:06:50]. However, client pressure is now driving the shift. Clients, especially large private equity firms, are less willing to pay for manual contract reviews or due diligence, making it unprofitable for firms to staff expensive associates on such tasks [00:07:31]. This creates a “prisoner’s dilemma,” forcing firms to adopt AI to remain competitive [00:07:59].
Adoption Rates
In traditional software rollouts, 5-10% adoption was considered good for legal firms. With Lorra, adoption rates are “increasingly hitting numbers like 70-80%,” driven by lawyers actively seeking access to these tools [00:12:57]. This indicates a strong demand for effective AI solutions.
Lorra’s Strategic Advantages
Lorra’s journey began in the Nordics, a smaller market that allowed the company to “eat our way to become a bit of a bigger fish” before expanding globally, including to the US [00:09:00]. This approach allowed them to become “enterprise ready” before entering new markets [00:11:52].
Being a “fast second mover” enabled Lorra to observe what was working and not working for others [00:09:43]. While many initial AI companies focused on training their own LLMs, Lorra prioritized building an “application layer” that users would be excited about, especially given their initial limited funding [00:09:50]. Their non-legal background also fostered humility and attentiveness to client needs [00:10:19].
Product Development and Architecture
A key challenge in AI application development is balancing immediate customer value with the rapidly evolving capabilities of models [00:13:31]. Lorra’s philosophy is: “if it’s something that the AI labs are going to build and at some point make available to builders like us, then we should not build it” [00:14:24]. This approach aims to build a flexible system where functionality improves as underlying models advance [00:14:38].
Model Pricing and Usage
While early expectations were that LLM prices would continuously decrease, newer, more capable models (like 03 for legal work) are often more expensive [00:19:53]. Lorra employs “classification algorithms and model pickers” to choose the best model for a given task, balancing cost and performance [00:20:29]. The ideal pricing model may evolve from seat-based to a platform fee with a usage element [00:19:38].
The “hardest part” of building these products is prioritizing among “a hundred different things” that offer high value and ensuring they form a “cohesive platform” rather than a “Frankenstein monster” [00:17:31].
Infrastructure and Modality
Lorra leverages tool-calling or “MCP” to enable its AI to access external tools, such as redlining documents or integrating with client-specific CRMs, knowledge databases, or template sets [00:20:57]. This significantly expands the scope of what the AI can accomplish [00:21:34].
Regarding AI application “modes,” Lorra aims to be a system of record and a central platform for collaboration and direct client interaction, similar to Figma in design [00:21:50]. This shifts the traditional legal workflow which has been heavily reliant on Microsoft Word, Outlook, and document management systems [00:22:40].
Looking ahead, multimodal use cases like voice and audio transcripts are exciting. For instance, uploading deposition audio files, transcribing them, and interrogating them as documents could automate tasks previously requiring manual notation [00:41:00].
The Future Lawyer
The impact of AI on education and legal careers is profound. Law firms are rethinking how they upskill and train new associates [00:41:48].
Essential Skills for Future Lawyers
Instead of hiring “underconfident overachievers that are… good at following instructions,” firms will increasingly need “entrepreneurial, creative people who maybe challenge the existing ways things have been done” [00:42:21]. The future lawyer will effectively be a “manager from day one of a bunch of AI agents,” requiring a different skill set than traditional diligence [00:43:06]. Every individual will need to “augment themselves” with AI [00:43:21].
Lorra collaborates with law firms and legal professionals to build this future [00:43:59].