I built a customized contract clause editor for less than one cent.
And it works.
My thesis is that lawyers should be learning to develop DIY AI tools for their own practices. And while this skill set complements legal work, it differs in one deeply satisfying way: unlike many of the heavily qualified, noncommital answers we often have to give clients as lawyers, these development of these tools is binary: either they work or they don’t. That’s not to say there isn’t room for testing, evaluation, and iteration.1 But when you line up all the parts and finally click “run” for the first time, the tool either returns an answer or throws an error—and there’s something refreshingly concrete about that.
So what did I build? A tool that, during a contract negotiation, analyzes a counterparty’s proposed edits to the indemnification provision in my base form contract, identifies the type of change made, and—drawing on my internal playbook—returns both marching orders and a revised version of the clause consistent with that guidance. It’s not fancy (yet), but it works.
I built the tool for less than one cent (plus the sunk cost of my monthly OpenAI ChatGPT subscription, currently about $20/month—which I use daily across countless other tasks). I wrote some rudimentary code with AI assistance, and the fraction of a penny was the variable cost OpenAI charged to process my data. And importantly, when I send data through the OpenAI API (a secure, professional gateway for sending data to OpenAI—much safer than pasting text into ChatGPT), it isn’t used to train their models or exposed to other users—an essential safeguard for anyone working with sensitive or proprietary information.
Of course, that “less than a cent” doesn’t capture the real value embedded in the tool: my playbook for responding to edits on a particular clause—in this case, indemnification—in a way that minimizes the impact of concessions from my client’s perspective while still moving the draft toward a deal I know, from experience, is achievable.
The important takeaway is this: the overwhelming majority of the value lies in your knowledge and expertise as a lawyer—built slowly over years of practicing and honing your craft. The AI is just the commoditized bolt-on. Your legal judgment is the asset.
I’ll cover the trials, tribulations, and plans for future improvements in upcoming posts. But for now, here’s a preview:
Chris Archer is a lawyer and business executive whose work spans real estate, construction, and technology, including practical applications of AI in legal practice.
The opinions and perspectives presented in this article are entirely Chris’s own and do not represent or reflect the positions of his employer. This isn’t legal advice—just ideas and observations—and it does not create an attorney-client relationship. For actual legal guidance, speak with a qualified lawyer.
I hope to cover these topics in the future.