Automated Ideation and Trend Spotting

The biggest bottleneck in content creation isn't writing; it's deciding what to write. Instead of staring at a blank screen, we automate the discovery phase.

I set up an automated trigger using Make.com. Every morning, the system scans specific RSS feeds (e.g., TechCrunch, Industry Blogs) and subreddit discussions relevant to my niche.

• Trigger: New popular post on Reddit (>100 upvotes) or news headline.
• Action: Make.com sends the headline to an AI agent via API.
• Evaluation: The AI analyzes the topic for "Content Fit" based on my pre-set criteria (Search Volume Potential, Audience Relevance).

If the score is high, it creates a new page in my Notion "Ideas" database with a proposed title and a status of "To Review." When I start my day, I have a menu of vetted ideas waiting for me.

The "Deep Dive" Research Brief

Once I approve an idea, the Research Agent kicks in. This is where Perplexity shines. We don't want generic AI fluff; we want data-backed insights.

The system queries Perplexity to analyze the top 5 ranking articles for the chosen keyword. It extracts key headers, data points, and—most importantly—Content Gaps (what competitors are missing).

This brief is saved to Notion. I spend 5 minutes reviewing it to ensure the angle matches my brand strategy before the writing begins.

The Recursive Writing Process (The "Claude" Step)

This is the critical differentiator. Most people fail because they ask for a 2,000-word article in a single prompt. LLMs degrade in quality as output length increases. To solve this, I built a Modular Writing Workflow.

I use Claude 3.5 Sonnet for its ability to adhere to complex style guides.

The "Chaining" Technique:

• Step 3.1 - The Hook: The AI writes 5 variations of the introduction. I pick the best one.
• Step 3.2 - Sectional Drafting: The AI writes Section 1 based on the approved brief. It maintains the context of the introduction.
• Step 3.3 - Critique Loop: Before moving to Section 2, a separate AI "Critic Persona" reviews Section 1 against my style guide (e.g., "Use active voice," "No buzzwords like 'game-changer'"). It auto-corrects the draft.
• Step 3.4 - Repeat: This loop continues for every section until the full draft is complete.

This recursive process ensures that the intro, body, and conclusion are coherent and high-quality, avoiding the repetitive "fluff" common in one-shot AI articles.

Visual Contextualization with Midjourney

A wall of text kills user experience. We need visuals. While the text is being drafted, the workflow identifies "Visual Opportunities."

The system scans the draft and suggests image concepts. For example, if the text discusses "scaling a mountain of data," it writes a detailed Midjourney prompt.

The "Human Polish" & SEO Optimization

At this stage, we have a 2,500-word draft and custom images sitting in Notion. Now, the human steps in. This is not editing for grammar; it's editing for impact.

I inject personal anecdotes that the AI couldn't possibly know. I verify the logical flow of complex arguments. I ensure the tone feels authentic, sometimes adding "rough edges" or controversial opinions that AI typically smoothes over.

The SEO Layer: Once polished, the text is run through an optimization tool (like Surfer SEO or NeuronWriter). It checks keyword density not by stuffing, but by semantic relevance (NLP), ensuring we cover the topic depth Google expects.