Uploaded June 2026 | Updated September 2026, 3 weeks ago
*Featured in this video:* Fred Miskawi, Vice President of AI Innovation, and Steve Zemanick, Director, National AI Strategy & Technical Lead, CGI
*Executive summary:* CGI, a global IT and business consulting firm, uses Gemini Code Assist and Gemini CLI to manage the complexity of codebases developed over 30 years. By integrating these AI tools across the entire software development lifecycle, CGI has seen a significant increase in developer velocity, producing 30-40% more output after just three months of ramp-up. The solution allows developers to navigate legacy estates, automate mundane tasks like test coverage, and perform sophisticated agentic development — all while maintaining high code quality and avoiding task switching.
*Challenge:* CGI manages a massive estate of legacy applications for thousands of global clients, many of which contain layers of complexity built up over three decades. These clients face the constant need for upgrade cycles, but navigating and understanding such vast, complex codebases often slows down development velocity. Furthermore, because over 20% of CGI’s revenue is derived from its own Intellectual Property (IP), the company must constantly innovate and maintain its own product suites while ensuring high levels of test coverage and code quality. Developers previously struggled with task switching and the repetitive aspects of coding — such as writing boilerplate and documenting APIs — which detracted from strategic problem-solving.
*Solution:* CGI implemented Gemini Code Assist and Gemini CLI to augment the software development lifecycle. Developers use Code Assist within their IDE to get real-time answers about APIs, look up reference documentation, and optimize algorithms without switching tasks. To handle the most complex legacy codebases, the team uses Gemini CLI for deeper code understanding and more sophisticated agentic use cases. Beyond development, this AI-powered ecosystem is also used for test automation and validating the value delivered to clients, ensuring that AI-driven insights are applied the right way across industries.
*Result:* The adoption of Gemini has transformed CGI's delivery model, with teams producing 30-40% more output after a three-month ramp-up. The tools have significantly increased test coverage and code readability across both client projects and CGI’s own IP, while Gemini’s ability to handle the context of large codebases allows developers to navigate 30 years of industry development effectively. By automating repetitive tasks like troubleshooting errors and creating boilerplate code, developers can also focus on higher value tasks.
*Highlights and key takeaways from our interview with Fred Miskawi and Steve Zemanick:*
→ “There's an organic growth of [Gemini] Code Assist that's evolving right now, and that's thanks to the quality and the value of what Google is producing.”
→ “With Gemini Code Assist and Gemini CLI within one place, [developers are] able to understand code and work on specific problems all without task switching.”
→ “Gemini seems to be handling the context of our larger, more complex code bases in a way that some of the other tools are not.”
*Google Cloud products used:* Gemini Code Assist, Gemini CLI
*Learn more:*
https://codeassist.google/
geminicli.com
*Featured in this video:* Fred Miskawi, Vice President of AI Innovation, and Steve Zemanick, Director, National AI Strategy & Technical Lead, CGI
*Executive summary:* CGI, a global IT and business consulting firm, uses Gemini Code Assist and Gemini CLI to manage the complexity of codebases developed over 30 years. By integrating these AI tools across the entire software development lifecycle, CGI has seen a significant increase in developer velocity, producing 30-40% more output after just three months of ramp-up. The solution allows developers to navigate legacy estates, automate mundane tasks like test coverage, and perform sophisticated agentic development — all while maintaining high code quality and avoiding task switching.
*Challenge:* CGI manages a massive estate of legacy applications for thousands of global clients, many of which contain layers of complexity built up over three decades. These clients face the constant need for upgrade cycles, but navigating and understanding such vast, complex codebases often slows down development velocity. Furthermore, because over 20% of CGI’s revenue is derived from its own Intellectual Property (IP), the company must constantly innovate and maintain its own product suites while ensuring high levels of test coverage and code quality. Developers previously struggled with task switching and the repetitive aspects of coding — such as writing boilerplate and documenting APIs — which detracted from strategic problem-solving.
*Solution:* CGI implemented Gemini Code Assist and Gemini CLI to augment the software development lifecycle. Developers use Code Assist within their IDE to get real-time answers about APIs, look up reference documentation, and optimize algorithms without switching tasks. To handle the most complex legacy codebases, the team uses Gemini CLI for deeper code understanding and more sophisticated agentic use cases. Beyond development, this AI-powered ecosystem is also used for test automation and validating the value delivered to clients, ensuring that AI-driven insights are applied the right way across industries.
*Result:* The adoption of Gemini has transformed CGI's delivery model, with teams producing 30-40% more output after a three-month ramp-up. The tools have significantly increased test coverage and code readability across both client projects and CGI’s own IP, while Gemini’s ability to handle the context of large codebases allows developers to navigate 30 years of industry development effectively. By automating repetitive tasks like troubleshooting errors and creating boilerplate code, developers can also focus on higher value tasks.
*Highlights and key takeaways from our interview with Fred Miskawi and Steve Zemanick:*
→ “There's an organic growth of [Gemini] Code Assist that's evolving right now, and that's thanks to the quality and the value of what Google is producing.”
→ “With Gemini Code Assist and Gemini CLI within one place, [developers are] able to understand code and work on specific problems all without task switching.”
→ “Gemini seems to be handling the context of our larger, more complex code bases in a way that some of the other tools are not.”
*Google Cloud products used:* Gemini Code Assist, Gemini CLI
*Learn more:*
https://codeassist.google/
geminicli.com
