Build an AI Affiliate Marketing Agent That Finds, Publishes, and Tracks Revenue Opportunities

PROJECT CONTEXT:
Build an AI-powered affiliate marketing agent for a solo operator or small team. The app should manage affiliate offers, generate audience-specific content, maintain disclosure-compliant recommendations, track outbound clicks and conversions, and surface which campaigns actually make money.

GOAL:
Launch a real working affiliate marketing agent with operational workflows for offer intake, content generation, link tracking, performance reporting, and optimization based on measurable revenue signals.

REQUIREMENTS:
- Create a database schema for affiliate programs, offers, campaigns, content assets, clicks, conversions, and earnings
- Provide a way to add affiliate programs manually including payout terms, landing URLs, tracking links, and disclosure requirements
- Build an AI workflow that turns offer details into content briefs and multiple content variations for at least 3 channels
- Store generated content and link every asset back to a specific offer and campaign
- Implement real outbound link tracking using redirect URLs or tracked click events before forwarding users to the affiliate destination
- Support conversion and earnings logging either via manual entry, imported CSV, or webhook simulation
- Show a dashboard with clicks, conversions, EPC, conversion rate, and total earnings by campaign and offer
- Include compliance safeguards such as affiliate disclosure text and restricted claims validation
- Add campaign status states such as draft, ready, active, paused, and archived
- Provide filtering and search across offers, campaigns, and content assets

CONSTRAINTS:
- Do not rely on fake analytics; every displayed metric must come from stored events or entered earnings data
- Do not claim direct affiliate network integrations unless they are actually implemented
- Keep the initial version focused on one operator workflow rather than a full enterprise team system
- Generated content must include disclosure handling and avoid unsupported income or product claims

DEFINITION OF DONE:
- A user can create an affiliate offer with destination URL, payout model, commission details, and disclosure text
- The AI agent can generate campaign ideas and content assets tied to a real offer record in the database
- Tracked outbound links record click events before redirecting to the destination URL
- The system can record conversions or earnings and reflect them correctly in campaign reporting
- The dashboard calculates and displays real metrics including clicks, conversions, conversion rate, EPC, and earnings
- Campaigns move through real status states and can be filtered in the interface
- Compliance text appears in generated content or campaign output where required
- At least one complete end-to-end test proves the workflow from offer creation to tracked click to recorded revenue

COMMON FAILURES TO AVOID:
- AI builds a pretty dashboard with hardcoded sample metrics instead of computing from stored click and revenue data
- AI generates content but does not attach it to campaigns, offers, or persistent records
- AI adds affiliate links directly without implementing tracked redirect or click logging
- AI claims conversion tracking exists but only stores a placeholder field with no input path or import flow
- AI ignores disclosure requirements and creates risky promotional copy
- AI builds campaign cards and statuses visually but does not persist status changes in the database