How a Lawyer Wrote Her First Code with AI and Automated Compliance

A lawyer from the IT industry decided to write his own tool for automating compliance processes — and opened an IDE for the first time in his life. Without a technical education, without programming courses. Spoiler: the code worked.

When a Lawyer Opens an IDE

Most lawyers live in a world of Word, Excel, and twenty-page contracts in small print. Their tools are templates, precedents, internal knowledge bases. Programming, as a rule, does not figure into this list.

The author of the Habr article faced hours of routine: check each clause of a contract for compliance with internal policies, document deviations, generate a report. Every time, manually. He could have submitted the task to the IT department — but understood that it would end up in the backlog for several quarters.

So he decided to do it himself. First step: download VS Code. Second step: accept that he couldn't do it without help.

At this stage, an AI assistant entered the process: something between a mentor and a pair programmer.

AI as a Pair Programmer

The approach turned out to be straightforward: describe the task in plain language, get working code, figure out why it works (or why it doesn't), move forward. Without the feeling that you're "asking the wrong thing" — AI doesn't judge for basic questions. What helped at each stage:

• Describing the logic of the task in words → AI translates it into code
• An unclear error → explanation without technical jargon
• Need a regex for parsing legal text → a ready-made template with an explanation of what and why
• Unclear project structure → step-by-step advice on organizing files
• Wrote the first version yourself → AI does a review and explains what can be improved and why

Key takeaway: AI doesn't write code instead of you. It helps you understand what exactly needs to be written — and why exactly like that. The difference is fundamental.

What Came Out in the End

After a few weeks of work, a working tool appeared. A Python script reads a contract in text format, searches for clauses by keywords and patterns, checks them against the company's internal policies, marks deviations, and generates a report in Excel. This is not a production solution. There's no beautiful interface, not all possible file formats are handled, there are no autotests (the author learned what those are during the process). But the tool works — and saves several hours a week.

"I didn't become a developer.

But I became a person who can explain to a developer what exactly needs to be done — and in his language."

What This Means

The story of a lawyer-coder illustrates one of the major shifts that AI brings to the profession: lowering the entry barrier for creating tools. Previously, writing a working program without a technical education was practically impossible. Now it's long, not easy, requires patience — but realistic.

For companies, an important shift: automation can come not just from the IT department. Specialists from law, finance, operations begin to solve their own tasks independently — faster than waiting in the backlog queue.

For lawyers — a separate signal. The ability to describe a task so that AI can solve it, and to understand the result obtained, is already becoming a competitive advantage. Not because lawyers should replace developers. But because the line between "who makes the rules" and "who codes them" is becoming thinner.