Key Concepts
- UGC (User-Generated Content): Authentic-style video content that mimics organic social media posts rather than traditional polished commercials.
- Winning Hunter: A market research tool used to identify trending products and high-performing ads across platforms like TikTok Shop and Meta.
- Claude (AI): An advanced LLM used here as an orchestration layer to research ad strategies and generate prompts.
- Hicksfield MCP (Model Context Protocol): A connector that integrates Claude with the Hicksfield Marketing Studio to automate video ad production.
- Marketing Studio: A specialized feature within Hicksfield that provides presets (e.g., unboxing, tutorials, product reviews) for AI video generation.
- Hook: The critical first two seconds of a video designed to stop the user from scrolling.
1. Market Research Methodology
The process begins by identifying products with proven market demand to minimize the risk of creating ads for products that won't sell.
- Platform Selection: The creator uses Winning Hunter to filter for products on TikTok Shop, focusing on metrics like "Revenue Growth" and "Items Sold."
- Data Analysis: By analyzing top-performing videos, the user identifies which creative angles and hooks are currently driving sales. For example, the video highlights a "PDRN Pink Collagen" product generating $3.74 million in 30 days as a prime candidate for ad creation.
2. The AI Automation Framework
The workflow leverages Claude’s reasoning capabilities combined with the Hicksfield MCP to bridge the gap between research and production.
Step-by-Step Process:
- Connector Setup: Install the Hicksfield MCP in Claude by adding a custom connector via the "Manage Connectors" menu using the provided remote MCP server URL.
- Prompt Engineering: Use a structured three-step prompt:
- Research: Instruct Claude to search for current high-performing ads and identify the top 3–5 recurring hooks.
- Conceptualization: Ask Claude to draft a one-paragraph UGC-style concept that mirrors the successful elements of existing ads.
- Generation: Command the Hicksfield MCP to fetch the product URL and generate the video using the appropriate Marketing Studio preset (e.g., "Product Review" or "UGC").
- Execution: Claude executes the research, selects the best preset based on the product category, and triggers the video generation within the Hicksfield environment.
3. Ad Creative Strategy
The creator emphasizes that the structure of the ad is non-negotiable for high conversion:
- The Hook (0–2 seconds): The most vital component. It must grab attention immediately to stop the scroll.
- The Demo (Middle): A clear, concise demonstration of the product’s value proposition.
- The CTA (End): A direct Call to Action (e.g., "Link in description") to drive traffic to the purchase page.
4. Customization and Iteration
The system allows for significant post-generation refinement:
- Regeneration: If the initial output is unsatisfactory, users can click "Recreate" and modify the prompt to change specific elements like the actor, the script, or the claims being made.
- Voiceover Editing: The script generated by the AI can be manually adjusted within the Hicksfield interface to better align with the brand’s voice or specific marketing goals.
5. Notable Quotes
- "Instead of guessing what kind of ad to make, we'll just start with what's already working and use AI to create something similar with our own angle."
- "The hook is the first two seconds. That's what you want people when they come on your video, they stop scrolling."
6. Synthesis and Conclusion
The core takeaway is that successful AI advertising is not about creating content from scratch, but about data-driven replication. By using tools like Winning Hunter to find proven winners and Claude/Hicksfield to automate the production of UGC-style ads, creators can bypass the manual labor of filming and editing. This framework allows for rapid testing and scaling of ad creatives, enabling users to monetize AI by providing high-quality, conversion-focused video services to businesses.
AI summaries can miss context or contain errors. Check important details against the original video.





