This AI Workflow Replaces 10 Hours of Ad Research

What It Covers

Mike Futia demonstrates how to automate TikTok ad research using AI workflows that scrape trending videos, analyze content with Gemini, and generate creative briefs in minutes. The system uses Apify for scraping, processes video transcripts and comments, and outputs campaign strategies based on viral content patterns for ecommerce brands and agencies.

Key Questions Answered

• •Automated trend scraping workflow: Build a system using Apify API to scrape TikTok by keyword and date range, pulling 20-200 videos with view counts, likes, and comments. Gemini then analyzes each video to extract visual hooks, product demonstrations, messaging themes, and funnel stage classification. This replaces manual scrolling and note-taking that previously consumed hours of researcher time.
• •Comment analysis for product insights: Configure Gemini to analyze video comments alongside content to identify common customer questions, complaints, and preferences about competitor products. This reveals gaps in market offerings and validates gut-skin connection beliefs or similar product hypotheses. The AI synthesizes patterns across thousands of comments that would take days to review manually.
• •Brand bible context system: Create a comprehensive brand profile including tone of voice, target audience definitions, customer pain points, and past campaign performance data. Feed this context to AI alongside trending video analysis to generate campaign briefs that match brand standards. The system produces 80 percent complete briefs requiring only human refinement before execution.
• •Creative brief automation pipeline: Connect trend research directly to brief generation by combining scraped video data, AI analysis, and brand context in one workflow. Output includes campaign names, objectives, target audience profiles, pain points, and key messaging. This compresses the research-to-strategy timeline from days to under 15 minutes for Facebook and TikTok ad campaigns.
• •Vibe coding over workflow builders: Start learning AI automation through Claude Code or Cursor rather than node-based tools like n8n or Make. Use AI to help build AI solutions by running Claude Code in one monitor and regular Claude for debugging in another. This approach proves more accessible for non-technical marketers than mastering API integrations and webhooks.