The Inspiration

Small service businesses can lose potential customers when inquiries and quotes go unanswered.
Following up manually can be time-consuming, and business owners may struggle to decide which leads need attention first.
Lead Rescue AI was created to address this problem with a focused, practical AI-powered tool.

What It Does

Displays leads and their current statuses in a simple dashboard.
Flags quotes unanswered for 72 hours or more as High Priority.
Identifies inquiries unanswered for 48 hours or more and quotes between 48 and 72 hours as needing follow-up.
Explains why a lead needs attention.
Generates personalized messages in three tones: Friendly Check-in, Gentle Nudge, and Value-Focused.
Allows users to edit and copy messages for use in their preferred communication channel.
Uses Google Gemini for AI-generated messages, with a template fallback when AI access is unavailable.
Supports manually adding leads and resetting the sample data.

The prototype prepares messages; it does not automatically send them through WhatsApp or other channels.

Followed the Devpost Learn Skill Pack's six-stage workflow: Start, Scope, PRD, Spec, Build, and Ship.
Created planning documents before implementation.
Built the backend using Python and FastAPI, with HTML, CSS, and JavaScript for the dashboard.
Integrated Google Gemini while keeping AI credentials on the backend.
Developed the application incrementally in five build slices.
Reported 17 passing automated tests before the final hands-on review.

Implementing consistent, time-based lead prioritization.
Testing the boundaries between recent, follow-up-needed, and high-priority leads.
Integrating AI generation without making the application dependent on API availability.
Providing useful fallback messages when AI generation fails.
Keeping the project focused enough to deliver a working prototype within the hackathon deadline.

Building a working end-to-end prototype within the hackathon timeframe.
Implementing deterministic lead prioritization with clear reasons for why a lead needs attention.
Integrating Google Gemini to generate personalized follow-up messages in three different tones.
Implementing a template fallback so message generation can continue when AI access fails.
Supporting lead creation, status updates, message editing, and copying through a simple dashboard.
Achieving 17 passing automated tests before the final hands-on review.
Following the Devpost Learn Skill Pack and using five incremental build slices instead of attempting the entire application at once.

Why planning and defining scope before coding makes implementation more manageable.
How to break a project into small, testable increments.
How to guide a coding agent, review its implementation, and use automated tests to verify behavior.
Why reliable fallback behavior matters when building applications that depend on external AI services.
How to turn an identified business problem into a working, testable software prototype.

Real-world validation: Test the prototype with plumbing and other home-service businesses to learn whether the prioritization rules fit their workflows.
CRM integration: Connect to existing systems so businesses can import real leads instead of entering them manually.
Communication integrations: Explore WhatsApp, email, and SMS integrations, with user approval before sending messages.
Smarter prioritization: Allow businesses to configure follow-up thresholds and eventually incorporate lead value and engagement history.
Performance measurement: Track follow-up completion and responses to determine whether the product actually helps businesses recover opportunities.
Deployment and security: Prepare a secure hosted version suitable for real business data and ongoing use.