The Small Automation That Changed How I Think About AI in Legal
The five-minute automation that saves me hours... and what it taught me.
My first foray into building a legal AI solution wasn’t flashy. But it changed the way I think about AI’s role in legal and set me on a path of building more.
It started with a simple pain point: every time a contract was signed, I had to manually download it from my email, rename the file, and store it in my company’s document management system. Sometimes I’d do it right away. Sometimes I’d forget. Sometimes it never got done at all.
So I built an automation, using only tools my company already had. Now, the second a signed contract hits my inbox, agentic AI grabs it, names it, and files it instantly. Every time. No exceptions.
It’s a small win. But it taught me lessons that now shape how I approach every legal AI project.
Want the template? Just drop a comment.
Lesson 1: Solutions Work Best When Built by Someone Who Has Walked in Your Shoes
AI, automation, or agentic projects only succeed if you approach them with both a project management lens and a deep understanding of your existing processes.
If you’re too far removed from the day-to-day reality of the work, you’ll struggle to deliver something useful. My automation works because it’s built for the exact pain points I face as a GC at a high-growth tech company. My constraints — budget, headcount, industry, customer expectations — are specific. They’re not the same as those of a public company GC or a GC in, say, the financial industry.
That’s why I’m skeptical of one-size-fits-all solutions.
The tools I actually trust? They’re built by people who have done the job themselves.
Two examples (not an ad - just a fan):
GC AI - founded by a former in-house tech GC who understands the nuance of redlining, reviewing, and negotiating tech contracts. Their new Word plug-in is the first I’ve seen that gets what in-house tech lawyers really want and need in an AI-assisted contract tool.
RelyanceAI - co-founded by an engineer and an in-house lawyer who both experienced the grind of maintaining GDPR records of processing activities and manual data maps. Their solution works because it’s built from lived pain.
When the builder understands the problem as deeply as the end user, the result both functional and genuinely usable.
Lesson 2: You Need to Understand the Tech Before You Use It
Building even a simple automation often requires elevated access to tools your company already has. Things like ChatGPT Enterprise, Zapier, Make, Relevance, n8n, Replit, etc.
IT and security teams don’t just hand out that access for fun. Nor should they.
If you want to work with your company’s data (or other sensitive data), you need to understand security requirements, role-based access controls, and where things can go wrong. That means learning the tech.
Teach yourself enough to understand what’s happening when you grab an API key, connect systems, and hook them into production. You need to know the downstream consequences of every connection you make.
If tech isn’t your thing, that’s fine! Fill your gaps. Partner with someone who has the skill set or hire an external resource.
And: be mindful of your IT and ops teams’ bandwidth. Their priority is usually building for customers, not the in-house legal department. If your internal teams don’t have the bandwidth to build for you, DIY or bring in the right external help.
Lesson 3: Start Small, Then Level Up
My contract automation wasn’t flashy, and that’s the point.
Starting with a small, low-risk project gave me space to experiment, learn the tools, and identify pitfalls before tackling something mission critical.
It also built credibility.
When cross-functional leadership sees you deliver something practical that works the first time, they’re more inclined to support the next, bigger project.
Think of it as building a track record in miniature… one small win at a time.
Lesson 4: Document, Measure, and Keep Humans in the Loop
Once you’ve built something, don’t just set it and forget it.
Document it - what it does, how it’s built, and what tools and connections it depends on. This isn’t just for you; it’s for whoever inherits it later or has to troubleshoot it.
Measure the impact - even rough metrics like “five minutes per contract, 500 contracts a year” tell a story about value delivered. Data is your best friend when you want to justify expanding AI use.
Keep human judgment in the loop - AI can execute, but only a human can set the rules, catch nuance, and course-correct when something goes sideways. The goal isn’t to remove the lawyer from the process; it’s to let the lawyer focus on the high-value work.
Bonus: I enjoyed this podcast episode, where the guests talk about delivering on legal ops so that your legal team can operate “at the top of their license.”
Where I Went Next
That first contract automation gave me confidence, credibility, and a blueprint for building AI into legal workflows without buying a single new tool. Since then, I’ve applied the same thinking to experimenting with things like:
AI-assisted SOW reviews — Helping deals close faster with automated knowledge bases, playbooks, customGPTs, redlines, and commentary.
Procurement automations — Automating feedback, insights, and status updates for business owners, with JIRA tickets progressing automatically based on triggered conditions.
RFI automations — Eliminating repetitive, manual FAQs and chasing people in Slack.
DIY CLM: Automating data entry points in contract management dashboards, the hub for cross-functional reporting and due diligence planning.
And more!
Each project started with the same principles: build for real pain points, leverage existing tools, and work within the realities of headcount, budget, and security requirements.
The Takeaway
You don’t have to have a huge budget, a massive team, or a moonshot project to start applying AI in legal. You just need a real problem and the willingness to learn…not only the tech, but also its implications.
And remember: small actions can deliver real results.
📣 Workshop This Friday: Building Your First Legal AI Automation
I’m joining the The L Suite (TechGC) this Friday for a live, hands-on session where we’ll walk through how to scope, build, and launch a legal AI or automation project.
You’ll learn:
How to pick a high-impact, low-risk starting point
Ways to work within budget and headcount constraints
Real examples (and pitfalls)
This is for in-house counsel, legal ops pros, and anyone curious about using AI to make legal workflows faster, smarter, and more reliable.
Seats are limited so you can get real Q&A time.
📚 Thank you for reading!
I hope you enjoyed reading it as much as I enjoyed writing it.
-Rachel
Sounds awesome. Would love to try it out.