The Challenge
By the time a cultural moment makes it into a brand deck, the window to act has often already closed. The problem isn't awareness: it's knowing which trends actually matter for your brand, in your category, before competitors act on them.
Standard social listening platforms surface volume without context. Quarterly trend reports arrive months after the moment has passed. Manual research is reactive and time-consuming. Most tools address discovery or relevance, but not both: they find what's trending without knowing if it matters to you, or monitor your category and miss what's emerging outside it. I wanted to build something that solved both: a system that starts with the brand, not the trend.
Implementation
I built the system with Claude Code using the WAT (Workflows, Agent, Tools) framework, which creates a clear separation between instructions, intelligence, and data collection. Workflow files define each stage of the pipeline in plain language. The intelligence layer runs through OpenRouter, which routes each task to the right model:
- Claude Sonnet 4.6 handles trend extraction, the most analytically demanding step
- Gemini 2.5 Flash handles brand analysis and activation
- Gemini 2.5 Flash Lite handles the simpler merging step
Apify handles the data collection, pulling posts from social platforms via its scraping actors.
The pipeline starts with a Brand Identity Scan, which I consider the system's core differentiator. It begins by scraping the brand's own TikTok, Instagram, and Facebook presence and using Claude to build a brand profile: category, tone, audience, and content themes. Everything downstream is filtered through this profile. It's finding trends for this brand.
From that profile, the system generates search terms and uses them to scrape trending content across TikTok and Instagram, either globally or focused on a specific country. Claude then identifies distinct trends from those posts and scores each one across four dimensions:
- Volume — how many posts feature it
- Magnitude — how much engagement it's generated
- Velocity — whether that engagement is growing or fading
- Breadth — how widely it's spread across platforms
The four scores combine into an Opportunity Score out of 100.
Brand relevance is evaluated separately. Claude assesses each trend against the brand profile across four criteria: category fit, competitive edge, audience match, and timing. Trends that fall below a relevance threshold are removed. For those that make the cut, Claude generates a set of activation ideas: concrete suggestions to help marketers act quickly and spark creative thinking.
Business Impact
The output is a ranked Trend Brief for any brand, generated in minutes. Each trend includes:
- Opportunity and Brand Relevance scores with full reasoning
- A trending timeline showing presence across three timeframes
- Example content with direct links
- Competitors already activating the trend
- Three to five specific activation ideas
- A Copy Brief export, ready to paste into a strategy document or creative brief
The result:
- Campaign planners can find creative hooks before competitors identify them
- NPD teams can surface product opportunities while demand is still emerging
- Brand strategists can run a scan before any major brief, without commissioning research or waiting on a trend report
For brands that want ongoing coverage, the system can also be configured to run on a schedule and deliver briefs automatically, or packaged as an agent skill, making it available on demand from within the tools your team already works in.
See it in action below: