← BACK TO WORK

AI Social & Media
Intelligence Monitor

Turning fragmented media and social conversations into structured, actionable intelligence.

BUILDn8n automationAIGeminiSOURCESGoogle News · Bluesky · YouTube
Person observing projected media
LISTENING / SIGNALS / INTERPRETATION

Social listening creates a lot of noise before it creates insight.

Monitoring conversations often means searching across multiple platforms, reviewing large volumes of mentions and trying to identify which narratives actually matter.

I wanted to explore whether that process could be automated without reducing every conversation to a simple positive-or-negative sentiment score.

One configurable system.
Two ways to listen.

I designed an n8n workflow that collects conversations from multiple media and social sources, filters and deduplicates the results, and uses Gemini AI to transform relevant mentions into structured intelligence.

OPEN LISTENINGDiscover what is emerging.

Monitor an entity or topic without predefined secondary themes so narratives can surface naturally.

FOCUSED LISTENINGInvestigate a specific issue.

Combine the monitored entity with selected themes or keywords to examine a particular conversation.

LISTEN→COLLECT→FILTER→ANALYZE→STRUCTURE
Wall of digital media imagery
MULTI-SOURCE MONITORING
INPUT SOURCES

Google News
Bluesky
YouTube

Incoming mentions are normalized into a common structure, filtered by the chosen lookback period, deduplicated and volume-controlled before AI analysis.

Built to move from search to structured intelligence automatically.

The workflow dynamically builds search combinations, gathers mentions across three sources, standardizes incoming data and processes relevant mentions through an AI analysis loop.

20-SECOND WORKFLOW / N8N
Person with projected media fragments and text
FROM MEDIA MENTION TO INTERPRETATION

Sentiment is not the same as stance.

A negative-sounding headline can still support the entity being discussed. The analysis therefore separates emotional tone from the position taken toward the monitored entity.

RAW MENTION

“Nike shareholders reject climate proposal backed by Norway wealth fund”

TOPICCorporate GovernanceSENTIMENTNeutralSTANCENeutralREPUTATION IMPACTNegativeCLASSIFICATIONRisk

Beyond positive, neutral and negative.

01Relevance
02Topic
03Narrative
04Sentiment
05Stance
06Emotion
07Audience perspective
08Related entities
09Reputation impact
10Risk / opportunity

From listening question to working automation.

Concept development, workflow architecture, API integration, data normalization, prompt design, AI classification framework, testing and QA.

N8NGEMINI AIGOOGLE NEWS RSSBLUESKY APIYOUTUBE DATA APIGOOGLE SHEETS

Less time collecting mentions.
More time understanding what they mean.

The project combines marketing intelligence, automation and AI analysis in a reusable system that can monitor a brand, person, organization, product, campaign or broader topic.

← BACK TO SELECTED WORK