Skip to content

Legal Tech for Transactional Lawyers Newsletter: #1

Picture of Matthew Basile

Brainstorming GenAI Use Cases, Document Automation, Prompt Engineering Tips and More!

Happy Friday!  I am excited to announce the launch of the very first edition of the Legal Tech for Transactional Lawyers Newsletter. As a legal tech professional and a practicing lawyer, my goal is to provide you with practical information about of how legal tech can transform corporate and real estate transactions.

In this edition I focus on thinking about what Generative AI use cases might work well in transactional matters and how it impacts traditional document automation strategies. I also included links to a few helpful articles from the last couple of weeks (especially relating to prompt engineering).

Hopefully you find it useful and interesting.  Please feel free to share with anyone that may be interested and have a great weekend!

Matt Basile

CEO, Naya Software, Inc.

If you find this content valuable, please connect with me on LinkedIn.

Leveraging Tech in Transactional Matters

📚 Brainstorming Generative AI Use Cases 

As we all start to think about how GenAI can be useful in transactional matters, I think it is helpful to do a deep dive into what GenAI is good at and what it is not good at in the context of transactional matters.  This article by Jack Shepard (and the various linked prior articles within) is all excellent content that will get you up to speed on how GenAI (and related LLM functionality) can be applied to drafting contracts.  Peter Duffy also breaks this article down nicely here in his must read newsletter that comes out every two weeks. Below are a few use cases in the transactional space that jump out at me for GenAI:

  • Improving document automation workflows (quick start to completing the questionnaire and checking the work product)
  • Generating bespoke language where you don't have precedent
  • Reviewing and summarizing due diligence documents
  • Extracting information out of documents for use in summaries (including items like S&P or other loan summaries, lease reviews, etc.)
  • Searching for precedent documents and clauses with semantic search powered by embeddings
  • Checking work (including looking for blanks and brackets, missing defined terms and other proofreading concepts)

💣Has GenAI Eliminated the Need for Document Automation Tools?

Definitely not in complex real estate and corporate finance transactions!  I had a great time discussing this topic with Jennifer Mendez and Catherine Bamford (Doc Auto expert from BamLegal).  I plan to expand on my thoughts in a separate post but as it relates to real estate and corporate transactions:

  • Traditional document automation is the way to go for automating complex sets of loan documents where formatting and specific language needs to be precise
  • Be careful using GenAI for generating any documents or text that needs to be consistent every time and based on vetted precedent documents as outputs are often not consistent
  • Don't wait to get started even if you do not have well defined templates, automate the core documents and build ancillary templates as you do transactions and as new language is used in real deals

More to come in the future on this compelling topic!

Stop the Robots! Has GenAI Eliminated the Need for Document Automation Tools?

Articles and News

📚 Summarizing Long Documents with GenAI

This blog post by Matt Payne of Width.Ai does a great job explaining some practical techniques for using LLMs to summarize documents that are bigger than the token limit. 

💱Change Management in the Legal Industry

Quotes from Darwin, Gretsky and Ferris Bueller are sprinkled through this Forbes article by Mark A. Cohen that discussed how growth, change and adaption are going to be key to staying ahead in the quickly changing legal market.  In the transactional space I think leveraging technology to get to more fixed fee arrangements will be key to staying ahead of the game.

⛓️Improve Summary Quality with Chain of Density Prompts 

Love this approach shared by Julian Horsey to refining summaries produced by GPT and excited to see how well a variation of this works for summaries of due diligence documents. Often the problem with simple summarization prompts for complex documents is that you don't get consistent topics and/or results.

👉Let's Think Step By Step 

If you are interested in digging in to prompt improvement, this is a great resource from OpenAI that outlines some great practical tips for improving performance of prompts.  Even small additions such as "Let's think step by step" or "Sit down and take a deep breath and think about your answer" can have a big impact on results.  Just like humans, given more time and space to think often leads to better results from the LLM!

My Work

❄️The Future of Transactional Deal Teams

AI won't replace lawyers anytime soon, but it should change how transactional deal teams are put together and push billing closer to fixed fee models.  Below are links to the first two parts of blog post that breaks down a couple common transactional deal team structures that exist now and also frame out how an ideal team will look in the future (which is actually being employed now by some of the more innovative firms).

Each post will look at specific tasks in a typical real estate loan transaction (drafting loan documents, due diligence, negotiating documents and closing process). In the final post, I will also discuss how each staffing model impacts efficiency, lawyer well-being and the long term business model as more technology becomes available.

Part 1 gives and overview of typical deal team structures and dives into the process of drafting loan documents.  Part 2 focuses on due diligence tasks that are performed by lawyers in a typical real estate transaction.

Related posts