JamSocial: AI Prompts Generator

Goal  The goal of this project is to develop an AI-powered prompts generator for Jam Bingo, designed to enhance networking and engagement at events. The tool will create dynamic, customized prompts that help attendees connect by encouraging interaction through fun and meaningful activities. These prompts will be based on key event details (name, start/end time, location, description) and any additional guidelines provided by event organizers.  Deliverables  1. AI-Powered Prompt Generator :  A system that generates prompts such as:  Find someone (e.g., "Find someone who has attended this event before.")  Ask someone (e.g., "Ask someone their favorite moment from last year’s event.")  Challenge someone (e.g., "Challenge someone to share a unique fact about themselves.")  Combinations like "Find someone and ask them about their goals for the event," or "Challenge someone and ask them to share a fun fact."  Prompts dynamically generated based on input fields:  Event name  Event Demographics (Optional) Start/end time  Location (city, province, country)  Event description  Organizer’s guidelines (e.g., specific themes or types of interactions to encourage)  2. Prompt Filtering and Categorization :  Categories to ensure diversity in prompts (e.g., icebreakers, professional, personal, fun challenges). Filtering options to exclude certain types of prompts based on organizer preferences.  3. Program Integration :  Ensure compatibility with a Node.js application. If the generator is not written in TypeScript, create a façade to facilitate seamless interaction between the Node.js application and the AI-powered generator.  4. Testing and Feedback Mechanism :  A pilot program to test the generator’s effectiveness at a live event. Feedback collection from attendees and organizers to refine the tool.  Requirements  The AI prompts generator must:  Accept event data as inputs (e.g., name, time, location, description, guidelines).  Generate prompts that are contextually relevant, diverse, and engaging. Operate as a backend program with clearly defined APIs or endpoints.  Be compatible with integration into a Node.js application, either natively or through a TypeScript-compatible façade.  Allow for future scalability and updates based on user feedback. Timeline  Week 1:  Define technical requirements and design the AI system.  Conduct research on effective networking prompts.  At request, Melvin from JamSocial will share a list of prompts used at past client events.  Week 2-5:  Build the AI model using natural language processing (NLP) techniques to generate 50 prompts as quickly as possible per query.  Week 6-8:  Test the system with mock event data and collect feedback.  Refine prompt generation algorithms based on feedback.  Discuss and review unwanted and wanted biases in prompts generated. Week 9-12:  Integrate AI prompt generator with fascade which will be utilized by our current systems.  Evaluate its performance and gather post-event feedback for future improvements.  Success Metrics  Relevance to Event Contexts: Prompts are directly aligned with the specific details of each event (e.g., themes, location, and goals).  Depth and Thoughtfulness: Prompts encourage meaningful conversations and deeper connections among attendees.  Organizer Satisfaction: Positive feedback from organizers regarding how well the prompts fit their guidelines and event objectives.  Keep in mind that these suggestions will be integrated into our website, and organizers can generate prompts, and we're measuring their satisfaction based on the rate at which suggested prompts are added after viewing.  Prompt Quality: High diversity and uniqueness in prompts with minimal repetition in a batch.