Here is a detailed summary of the video transcription in markdown format, broken down into sections for better readability:
Introduction
- Kik from Accenture, leading global asset engineering
- Over 20 years of experience, specializing in AI and Automation in software development and platform management
- Filed over 115 patents in this space
- Collaborating with Luke Higgins, the Chief Arch Tech responsible for Accenture's Global Asset and Automation Deployment
Approach to Generative AI
- 2 years ago, the team started exploring the potential of generative AI, especially with the launch of ChatGPT
- Worked with Stanford to create a scientific model on applying generative AI across the technology delivery lifecycle
- Analyzed over 50,000 projects to understand where generative AI can complement traditional automation
- Collaborated with AWS to understand their capabilities around foundation models (e.g., CodeWhisperer)
- Conducted 70+ controlled pilots to learn and create playbooks
- Currently scaling the approach across 1,200+ projects
Key Capabilities Enabled by Generative AI
-
Digitizing Knowledge:
- Extracting explicit and implicit knowledge from systems, documents, and people
- Reverse engineering code to document functionality (e.g., 50% accuracy on Assembler code)
-
Managing Applications and Platforms:
- Automating ticket processing and root cause analysis
- Generating insights from voice calls and other unstructured data
-
Software Development Lifecycle:
- Generating new features and user stories faster
- Creating architecture as code and generating UI code
- Reverse engineering platform customizations and configurations
-
Rationalization, Modernization, and Data Engineering:
- Comparing and finding commonalities across different applications
- Rewriting software from one technology stack to another
- Automating data engineering tasks
Platform Architecture
- Built on top of AWS services, utilizing foundation models like CodeWhisperer
- Consists of a data fabric, AI layer, and various use case-specific capabilities
- Provides user experiences through integration with tools like Jira and ServiceNow, as well as custom interfaces
Demonstration
-
Knowledge Digitization:
- Reverse engineering a C-based application using a multi-step process involving different foundation models
- Generating functional descriptions, user stories, and test cases from the reverse-engineered knowledge
-
Application Documentation:
- Digitizing knowledge from recorded sessions and extracting standard operating procedures
- Automating the creation of workflows for ticket resolution
-
Software Development:
- Generating new features and user stories based on existing functionality
- Creating functional and technical architecture artifacts as code
- Generating code using tools like AWS CodeWhisperer
-
Application and Platform Management:
- Enriching the CMDB with reverse-engineered and process-mapped information
- Generating recommendations for application rationalization and modernization
Live Demonstration
- Showcasing the platform's capabilities in incident management, event ops, and workflow automation
- Demonstrating the reverse engineering of a large-scale Java application with millions of lines of code