Gen AI Engineered Prompts Repository
Background
Role: Business Analyst, API Data Acquisition
Objective: Independently initiated the design and curation of a specialized repository of engineered prompts to optimize the use of ChatGPT for workforce efficiency as a side project. This project later became a cornerstone in an enterprise-wide generative AI initiative I lead, culminating in the development of an internal AI assistant.
Duration: Ongoing since inception.
Key Responsibilities:
Prompt Engineering: Developed and iteratively refined a diverse array of prompts to facilitate various business applications of off-the-shelf AI assistances like ChatGPT.
Repository Structure: Conceptualized and established a logical and user-friendly repository framework to store and categorize prompts.
Community Building: Encouraged and supported contributions from across the company to build a comprehensive prompt library.
Documentation and Training: Created documentation and training materials presented in enterprise-wide workshops to enable users to effectively utilize the repository.
Overview
Issue Identification: Identified the need for a structured and shared resource of AI prompts to support the practical application of generative AI across various business units.
Methodology: Utilized a blend of technical knowledge and project management skills to develop a repository using tools that support collaborative contribution and version control.
Challenges: Highlighted the need for a user-friendly interface that caters to both technical and non-technical users, fostering widespread adoption and ease of use.
Impact: The repository has become a crucial asset for employees, empowering them with ready-to-use prompts that enhance efficiency in tasks such as content creation, data analysis, and automated customer interactions.
Future: This repo has now transformed into a prompt library feature built into our enterprise AI assistant.
Skills and Tools
Skills: Prompt Engineering, Product Management, Generative AI, Instructional Design.
Tools: Azure DevOps, ChatGPT, PowerPoint.
Highlights
-
Comprehensive Prompt Catalogue
The repository boasts a comprehensive catalogue of prompts, tailored to a variety of use cases across different roles within the organization. From Product Owners to Lead Developers, each role can find specialized prompts that are designed to streamline their specific workflow. This extensive assortment ensures that regardless of the task at hand, there is a prompt ready to be deployed, enhancing productivity and consistency across projects.
-
Centralized Collaboration Hub
The repository serves as a centralized hub for collaborative prompt development, enabling teams to work together seamlessly. With a clear structure and naming convention for files, contributors can easily navigate and contribute to the growing library of prompts. This hub facilitates the sharing of knowledge and prompt engineering techniques, allowing teams to leverage collective expertise and reduce redundant efforts.
-
Templatized Prompt Structure
The repository includes sophisticated templatized prompts, designed to standardize the output while minimizing the effort required for each use. For example, the User Story Generator prompt, as depicted, exemplifies this approach. Users perform a one-time setup, entering their project details, preferences, and technical stack into clearly defined fields. Once saved, this customized template becomes a ready-to-use foundation for future tasks. For subsequent user stories, users need only modify the 'user_story_details' section to generate a new narrative, with all other fields pre-populated to ensure consistency and accuracy in output, streamlining usage.