Back to projects
AI
AdIntel AI
AI-powered campaign intelligence platform for media buyers — Meta, Google, TikTok, Taboola
TypeScriptNext.jsPostgreSQLOpenAI APIMulti-Platform Ad APIsAI Analytics
Overview
A campaign intelligence platform that consolidates ad performance data from Meta, Google, TikTok, and Taboola into one view, using AI to surface root causes of performance shifts, generate executive-ready briefs, and recommend budget reallocation for winning campaigns.
Problem
Media buyers manage campaigns across multiple ad platforms with no unified view, making it slow to spot underperformance and explain 'why' to stakeholders.
Solution
A single dashboard that pulls data across ad platforms, applies AI-driven root-cause analysis to performance changes, and auto-generates executive briefs and budget-optimization suggestions.
Architecture
- 1Ingestion layer normalizing data from Meta, Google, TikTok, and Taboola APIs into one schema
- 2AI analysis layer detecting performance anomalies and attributing likely causes
- 3OpenAI-powered briefing layer summarizing findings into executive-ready reports
- 4Budget reallocation recommendation engine for winning campaigns
Technical Challenges
- Reconciling inconsistent metric definitions across four different ad platform APIs into one comparable schema.
- Generating root-cause explanations that are genuinely useful rather than generic AI summaries.
Results
- Consolidates four ad platforms' data into a single AI-analyzed view for faster campaign decisions.