This excerpt from AI for the Rest of Us shares practical advice for implementing legal AI in your practice, covering costs, security, quality control and more.
Nextpoint partnered with Tom O’Connor to publish AI for the Rest of Us, a comprehensive guide to AI for legal professionals navigating this rapidly evolving landscape. This excerpt from Chapter 3 shares seven key tips to keep in mind before adopting AI tools in your legal practice.
The book explains the fundamentals of artificial intelligence, defines key terminology, dives into the ins and outs of legal AI, and covers AI risks and ethics. Finally, it hones in on actionable guidance that will help legal professionals get real results when using AI. Click the button below to learn more about AI for the Rest of Us and buy your copy.
1. Start with Your Problems, Not the Technology
Begin by clearly defining the problems you want to solve. This might seem obvious, but in the face of exciting new technology, it’s easy to get carried away with the latest flashy features. Instead of searching for tools to solve your biggest challenges, you end up looking for problems that the tool can solve – working backward rather than forward.
Of course, there’s nothing wrong with seeing a new innovation and feeling inspired to explore the possible benefits to your practice. But it’s still important to set the foundation of your AI search with a careful analysis of your team’s workflows to obtain a clear idea of your biggest needs.
Ask yourself: What slows you down every day? Which tasks eat up the most time? What percentage of your work feels routine versus intellectually challenging? Focus on identifying workflow weaknesses that require less human judgment – these are perfect candidates for AI automation. By handling the menial tasks, AI frees you up to do what you do best: practice law.
2. Know Your AI Types and Models
Understand the different models and types of AI utilized in legal tech. Not all AI is created equal, and understanding the technology behind your tools helps you make better choices. Does the platform use generative AI or a more traditional form of artificial intelligence? Is it built on a large language model or a smaller, more focused one? Has it been fine-tuned for legal work or enhanced with retrieval-augmented generation (RAG)?
AI tools that use fine-tuning or RAG are typically better suited for specific legal tasks. Smaller language models often provide more targeted results, and model size affects cost, speed, and accuracy. The more you understand these distinctions, the better questions you’ll ask when evaluating AI tools. (Check out the glossary section on page 30 to dive deeper into these terms.)
3. Get Your Team on Board
Increase user adoption by making a training plan and involving impacted individuals in the technology selection process. Even the most cutting-edge AI tool won’t help if your team refuses to use it. At Nextpoint, we’ve worked with legal teams led by enthusiastic leaders who understood the efficiency gains and cost savings our ediscovery software would bring. However, we’ve also seen these tech-forward leaders struggle to get the rest of their team on board with using the software in their daily workflows.
When this happens, our expert team travels on-site to conduct personalized training sessions, empowering new users to engage with the software confidently. In the ever-evolving world of AI, it’s more important than ever to choose software providers with reliable support and knowledgeable, hands-on training expertise.
Through our client work, we’ve observed that technology adoption is often more successful when relevant users are involved in the software selection process. Make your team feel heard and work to address their specific pain points with new technology – they’ll adopt the tools more readily, making your investment worthwhile.
4. Understand the True Costs
Think through the costs associated with implementing AI and consider which pricing models make the most sense for your practice. Of course, cost is a primary consideration when a legal team adopts any new technology. AI pricing typically follows two models: usage-based or subscription-based, which we explain in more detail in the next section of this chapter. Subscription pricing offers greater predictability, while usage-based models can be more cost-effective depending on your needs. Don’t just look at the upfront costs – consider how AI can reduce your overall expenses.
Most are aware of the abundance of free AI tools in the market today, often used by the lay citizen to replace Google searches or generate derivative art for social media trends. But before using any free AI tech, ask if there is a trade-off. The free ChatGPT model is far more likely to produce a fake case citation than AI-powered legal research tools that have been specifically honed to support legal practitioners. Free tools also often come with data privacy concessions and fewer controls over how your information is used. For legal work, the security and accuracy of paid, specialized tools are usually worth the investment.
Calculate potential savings by measuring time spent on tasks that AI could handle. During your initial workflow analysis, track how much time goes to routine work – this helps you gauge the return on investment.
5. Prioritize Security and Ethics
Consider all the security, confidentiality, and ethical implications of using AI in your practice. We explained the risks and ethical considerations of AI in the previous chapter, but this point is always worth revisiting. Make sure the tools you are exploring can meet the rigorous standards for data protection in the legal industry. Check your local bar association’s guidelines about client notification and consent for AI use – even if it’s not required, transparency with clients is always smart.
6. Plan for Quality Control
Make a plan for quality control and verification of the AI’s outputs before implementing new tech. The possibility of hallucinations, inaccuracies, and other mistakes presents another key ethical consideration in addition to the potential data security risks associated with AI. Don’t invest in AI tools without a solid verification plan – you don’t want to spend more time checking AI outputs than you would have spent doing the work yourself.
Ask software providers how they evaluate their AI’s effectiveness and how those insights translate to you as the end user. For example, one of Nextpoint’s AI researchers built a comprehensive evaluation tool and interface for our AI-powered transcript summary feature. Our back-end team uses this tool to measure results and continuously improve the model. While this is not a part of the client-facing interface, our team is happy to walk users through the tool to demonstrate our quality assurance process and evaluate specific transcript summaries on request.
Your quality control process should combine methodologies offered by your software provider with your own internal verification workflows. As you develop these processes, consider what level of inaccuracy you find acceptable in different contexts. Are you comfortable handing over TAR-reviewed ediscovery files to opposing counsel, knowing some relevant documents might be missed? (When in doubt, establish clear standards with opposing counsel in your ESI protocol.) Would you submit an AI-drafted brief without verifying every case citation? (That’s a trick question – you should always verify citations, or risk facing serious consequences like the attorneys in the Mata v. Avianca case discussed in the last chapter.)
7. Leverage Your Existing Tech Stack
Explore integrations or existing tools within your current tech stack. Integration is key to successful AI adoption. If your current software providers are developing AI features, start there – you already trust them with your data and know their support quality. Many legal tech companies are adding AI capabilities to their existing platforms, which can be more seamless than implementing entirely new systems.
However, be wary of companies simply slapping an “AI” label on basic features without offering real transparency or support. Not every AI company can deliver on its promises or keep your data secure. Stick with proven providers who have track records in the legal industry.
Look for integration opportunities with your current tools before committing to new platforms. The modern AI landscape allows for more cross-functional technology usage and app integrations than ever before – explore these options to maximize your existing investments.
Want to Learn More? Get the Full Book
AI for the Rest of Us is available to purchase on Amazon. The full book expands on the advice mentioned in this excerpt, with detailed explanations on AI costs, security, ethics, and more.
