Artificial intelligence is increasingly becoming part of workplace productivity across industries, including legal services. While software development has traditionally been associated with coding expertise, new AI-powered tools are allowing professionals without technical backgrounds to build practical solutions for repetitive tasks. One example comes from GitHub’s legal team, where lawyers and business professionals used GitHub Copilot CLI to create tools that improved efficiency without relying on traditional software development skills.
The experience demonstrates how AI can support knowledge-based work while keeping human expertise at the centre of decision-making.
Turning Repetitive Legal Tasks Into Practical AI Workflows
Many legal professionals spend a significant portion of their time reviewing contracts, answering recurring legal questions and reusing previous guidance. These repetitive responsibilities often create opportunities for automation, yet many non-technical employees have traditionally viewed software development as beyond their reach.
According to the team, GitHub Copilot CLI changed that perception by allowing users to describe what they wanted in plain language. Instead of writing code from scratch, they built customized tools around their existing workflows, helping improve consistency and reduce manual work.
The result was a growing culture of experimentation, with lawyers, program managers and business professionals developing solutions tailored to their own responsibilities.
Building a Contract Drafting Tool Around Legal Expertise
Creating a Consistent Drafting Style
Principal Product Counsel Ngandu Kasuku explained that increasing volumes of partnership agreements involving data, infrastructure and product integrations prompted him to rethink how he approached legal drafting.
Rather than relying on separate AI prompts for individual tasks, he built a contract drafting tool called terms-ai using GitHub Copilot CLI.
The project began by organizing drafting instructions, guidance documents and workflow resources into a structured repository. This central location allowed version control while giving the AI more consistent reference material, reducing the need for repetitive copying and pasting.
A key feature of the tool is an internal drafting style guide built around plain-language legal writing. Instead of relying on traditional legal expressions that can make contracts difficult to read, the guide encourages clearer language while maintaining legal accuracy.
Using Previous Agreements Securely
Kasuku also developed a library containing previously completed agreements. When similar contracts or amendments are received, the system can reference earlier approved work to support drafting.
Sensitive agreements remain within an approved, access-controlled internal environment and are not included in the project’s open-source repository.
According to Kasuku, the tool has reduced his drafting and review time by approximately half while improving consistency across contracts.
He said the biggest takeaway was not simply drafting documents more quickly, but using AI to create a tool that reflects his own legal judgment, experience and preferred working style.
Developing Legal Workflows Without Traditional Programming
Starting With a Specific Legal Challenge
Online Safety Counsel Jesse Geraci focused on a different problem: analysing source code efficiently when evaluating Digital Millennium Copyright Act (DMCA) notices.
The initial project consisted of structured GitHub Copilot instructions covering recurring legal tasks such as DMCA assessments, code comparisons, licence reviews and circumvention analysis. The objective was to replace inconsistent one-off prompts with standardized workflows capable of gathering relevant facts and producing more consistent legal analysis.
Geraci said he was surprised by how much progress could be made without engineering support.
Instead of writing software code, he created plain-language workflow instructions, policy references and reporting templates that embedded legal reasoning directly into the process.
Expanding Into a Broader Legal Platform
As the project evolved, the team introduced separate analysis modes for legal professionals and clients, offering faster recommendations for clients while providing more detailed legal analysis for lawyers. External data sources were also incorporated.
The workflow eventually developed into a desktop application that runs predefined legal processes through a streamlined interface. While building the application required software development, the workflow instructions themselves remain editable in plain language, allowing legal professionals to customize the system without programming expertise.
The platform now supports a range of in-house legal tasks, including contract reviews, non-disclosure agreement (NDA) assessments, risk evaluations, compliance reviews and response drafting.
Behind the scenes, reusable skills and AI agents assist with functions such as intake, playbook alignment, risk scoring, evidence verification, escalation routing and report generation. However, the legal team continues to manage the workflows using readable Markdown files rather than complex programming languages.
Geraci emphasized that the AI system is designed to support—not replace—legal judgment. Human review remains central, while the technology helps make legal analysis more consistent, transparent and scalable.
AI as a Productivity Tool for Non-Technical Professionals
The experiences shared by GitHub’s legal team illustrate that AI development is no longer limited to software engineers. By describing workflows in plain language and organizing existing knowledge effectively, professionals in legal, business and administrative roles can create tools that address repetitive work while preserving expert oversight.
For Canadian organizations exploring practical AI adoption, the examples highlight how productivity gains can come from improving everyday workflows rather than replacing professional expertise. As AI tools continue to mature, many knowledge-based workplaces may find similar opportunities to automate routine processes while keeping critical decisions firmly in human hands.
Nolan Fraser is a contributor at Angperyodiko.ca, covering a wide range of topics including news, politics, business, technology, sports, entertainment, and lifestyle. He focuses on delivering clear, accurate, and accessible reporting that helps readers stay informed about current events and developments that matter. With an emphasis on useful information and balanced storytelling, Nolan aims to provide timely coverage and relevant insights for audiences across Canada and beyond.