New Way Now: KeyBank saves 400+ hours on deployments with Gemini CLI and Code Assist

Google Cloud · Beginner ·💻 AI-Assisted Coding ·6h ago
*Featured in this video:* Azim Shaik, Lead Enterprise Architect, KeyBank *Executive summary:* KeyBank, a leading financial institution, needed to modernize legacy codebases and eliminate manual bottlenecks that slowed down development cycles. By adopting Gemini CLI and Gemini Code Assist, KeyBank transformed its software delivery process, reducing the time to move from ideation to proof of concept (POC) from months to just days. The bank is now saving 300-400 hours on internal deployment processes, allowing engineers to shift from repetitive manual tasks to more meaningful work. These tools have also mitigated risk by generating real-time documentation for legacy code and reducing release cycles through AI-augmented peer reviews. *Challenge:* As a regulated financial institution, KeyBank manages critical legacy code where outdated documentation presents a significant operational risk. If a system breaks without clear documentation, the time to recovery increases exponentially. Furthermore, developers were burdened by context switching, forced to navigate multiple external websites to find patterns for specific problems. Internal deployment processes were also hampered by a high number of manual steps that consumed hundreds of development hours, while the traditional ideation and prototyping phase for new features typically took a couple of months to complete. *Solution:* KeyBank chose to adopt Gemini CLI and Gemini Code Assist to automate these manual hurdles. The team implemented a framework around KeyBank’s internal manual processes, using Gemini CLI to handle tasks that previously required extensive human intervention. Developers now use Code Assist within their IDE to solve programming tasks and optimize code without switching between multiple browser tabs. The bank also uses Gemini CLI to analyze legacy code and generate detailed, up-to-date documentation, ensuring that the logic behind 30-year-old systems remains transparent and manageable as the code evolv
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Related AI Lessons

The era of the Indian H-1B programmer is over. US graduates will take their place.
The rise of AI automation is replacing H-1B visa holders in programming jobs, making way for US graduates to fill the gap, highlighting the need for American education to focus on skills that are complementary to AI
Medium · Programming
Claude Strangelove or: How I Learned to Stop Worrying and Love Coding with AI
Learn to leverage AI in coding with Google Certifications and overcome initial worries
Medium · Programming
Built a Portfolio Site With My Own AI Chatbot Using Claude Code
Learn to build a portfolio site with an AI chatbot using Claude Code in 3 steps
Medium · AI
Built a Portfolio Site With My Own AI Chatbot Using Claude Code
Learn to build a portfolio site with an AI chatbot using Claude Code in 3 steps
Medium · LLM
Up next
Clean Code vs Technical Debt!
Real Python
Watch →