From: redpointai
The legal industry is undergoing a significant transformation due to the rapid advancements in artificial intelligence (AI). Lora, led by CEO and co-founder Max Unistron, is a company at the forefront of applying AI to the law industry [00:00:26]. They work with many top law firms globally, have raised over $100 million, and are considered one of the fastest-growing AI applications [00:00:31].
Current State of AI in Law
When Lora started in 2020, early AI models like Google’s BERT were “horrendously bad” in languages like Swedish, though decent in English [00:01:46]. The arrival of GPT 3.5 marked a significant shift, moving the industry from full-on experimentation to actual implementation of AI solutions [00:01:52].
Today, AI is enabling end-to-end work deliverables. For example, due diligence no longer requires physically going into data rooms or using outdated software. Instead, all documents can be put into Lora, which finds the required information and generates reports based on those findings [00:02:08]. The focus has moved from simple queries to defining processes for the LLM to follow, utilizing agents with access to various tools for planning and execution [00:02:29].
The Future of AI in Law
While models continuously improve, the most significant leverage is coming from surrounding frameworks like function calling and tool calling [00:02:53]. The historically fragmented legal software space, with separate tools for translations, document comparisons, searching, and reviewing, is now consolidating as AI bakes these functionalities together [00:03:06].
Legal work can be viewed on a spectrum of complexity:
- Bottom Quartile: Simple tasks like data extraction and finding information [00:03:28]. Much of this is already fully automated [00:03:39].
- Top Quartile: Complex tasks such as drafting a share purchase agreement [00:03:34]. The industry is slowly moving up this scale [00:03:45].
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 Uniquely Suited for AI
The legal space has historically lacked significant software development [00:04:34]. Industry-specific incentives have not always aligned with efficiency and software adoption, with templating systems being considered cutting-edge [00:04:45].
Legal work broadly falls into categories: reviewing, reading, drafting, writing, or researching [00:05:20]. Historically, software focused on only one of these areas. However, AI’s capability to perform across this entire stack has led to the emergence of platforms like Lora, which provide wall-to-wall solutions rather than point solutions [00:05:36].
Challenges and Pressures Driving AI Adoption
The traditional hourly billing model was thought to hinder software adoption [00:06:50]. If AI makes a lawyer 50% more efficient, they bill 50% less [00:07:06]. However, similar to how lawyers adopted databases over manual library research, changes are inevitable [00:07:14].
Pressure points for law firms include:
- Write-offs and Price Pressure: Tasks like due diligence, once expensive, are now often not paid for by large private equity clients who prefer to pay for high-level advisory [00:07:26].
- Outsourcing: Large American firms are outsourcing tasks like contract review because it’s unprofitable to staff associates billing $800 an hour on such work [00:07:48].
- Client Adoption: Clients are increasingly using AI tools internally and putting pressure on law firms to adopt similar technologies [00:06:09]. CEOs are declaring “AI-first” strategies, requiring proof of efficiency before approving new headcounts [00:06:14].
- Competitive Pressure: If competitors adopt AI, other firms must adjust to keep pace, as there is low differentiation on specific tasks [00:08:00].
Law firms want to lead in this new paradigm. Firms not leaning in risk not upskilling their teams in new ways of working [00:05:54].
The Evolving Role of Legal Professionals
The traditional “underconfident overachievers” who are good at following instructions and working step-by-step are no longer enough [00:42:18]. The future lawyer will need to be:
- Entrepreneurial and Creative: Challenging existing methods of work [00:42:34].
- Fluent in AI: Augmented by AI elements in their processes [00:42:52].
- Managers of AI Agents: This represents a very different skill set than being a diligent associate [00:43:05].
Individuals in the legal field must take responsibility for augmenting their own work with AI and demonstrating how they are doing so [00:43:19]. This means a shift from relying solely on IT or innovation departments for AI integration to personal adoption and mastery.
AI Applications in Legal Workflows
Lora’s product strategy involves building an application layer that people are excited to use, rather than focusing on training proprietary LLMs [00:10:10]. This approach allows them to be agile with model improvements and provide value today [00:15:30].
Key aspects of AI in legal workflows:
- Agents and Tools: LLMs, given proper planning and execution with tool use, are capable of on-the-fly creating their own workflows [00:15:58]. The exciting prospect is clients providing their own tools (e.g., CRM, knowledge database, templates) for Lora to access [00:21:13].
- Playbooks: Lora has a “playbooks” feature where users can give it a set of rules and fallbacks for negotiation (e.g., for NDAs or MSAs) [00:14:46]. While this adds value today, the future might see models advanced enough to simply take a playbook from a document and cross-reference it with another, making current features unnecessary [00:15:08].
- Model Picking: Lora uses classification algorithms and model pickers to choose the best model for a specific task, balancing capability and cost [00:20:29]. More advanced models like 03 are “incredibly good” for legal work but also “very very expensive” [00:20:11].
- Reliability and Expectations: Initial challenges included reliability and infrastructure to deliver on user expectations [00:29:45]. If an attorney’s query doesn’t work, they might not return, making reactivation difficult [00:29:34]. Users have varied expectations; some are tech-savvy with templates and prompt libraries, while others expect the world from a simple query [00:38:25].
Adoption and Business Model
Lora’s adoption rates are significantly higher than traditional software rollouts. While 5-10% adoption was considered good for traditional legal software, Lora is seeing 70-80% adoption rates, with lawyers actively seeking access to the tools [00:13:04].
On pricing, Lora currently uses a seat-based model, which is easy to predict for clients [00:19:23]. However, a single user can incur $10,000 in LLM costs, suggesting a future shift towards a platform fee with a usage element [00:19:32]. While early expectations were for LLM prices to continually decrease, models are becoming better but also more expensive [00:19:51].
The ability to integrate with customers’ own data, as well as external databases for case law and legislation, forms a core “moat” or competitive advantage for AI applications in law [00:22:12]. This moves the system of record from traditional tools like Microsoft Word, Outlook, and IM Manage to AI-powered platforms [00:22:40].
Future Opportunities Beyond Text
Beyond text, multimodal AI offers exciting possibilities. Voice and audio transcripts are particularly promising. For example, uploading audio files of depositions, transcribing them, and then interrogating them as documents can streamline processes that previously required manual note-taking [00:41:00].
Lora’s Unique Growth Strategy
Lora’s growth strategy included starting in the Nordics, winning that market, and then expanding [00:09:00]. Being a “fast second mover” allowed them to observe what worked and didn’t, leading them to focus on the application layer rather than expensive LLM training [00:09:43]. This approach, combined with a non-legal background, fostered humility and attentiveness to client needs [00:10:19].
Unlike many legal tech companies that solve niche problems, Lora aimed to service “every lawyer” with a broad range of capabilities, aspiring to be for lawyers what Figma is for designers [00:11:14]. Starting with enterprise clients like Manheim Schwartling (the largest law firm in the Nordics) from day one meant they were immediately “enterprise ready” [00:36:17]. They invested half of their initial angel funding in certifications like SOC and ISO to meet enterprise standards [00:36:46].
The company prioritizes speed, intense focus on commercial aspects, and a culture of high urgency. This includes a “no remote work” policy and attracting employees who “love winning and hate losing” [00:34:07]. This intense pace has enabled Lora to grow from 10 to 100 people in a year [00:34:53]. This rapid growth, driven by AI, sets a new expectation for how quickly software companies can scale, creating entirely new categories rather than just replacing existing software [00:35:44].