AI Content Generation

Define 5 content topics aligned with your ICP's problems; research top-performing articles on each topic using Ahrefs or BuzzSumo to understand what resonates.

Create a prompt template for Claude or GPT-4 that includes: target audience, topic, key points to cover, desired tone, word count (800-1,200 words); test the prompt with 3 variations to find the best structure.

Generate 5 blog posts using your refined prompt; for each post, review and edit for accuracy, brand voice, and SEO optimization (include target keyword in title, H1, and body).

Publish the 5 posts on your blog or Ghost; add relevant images, internal links, and clear CTAs; ensure each post is optimized for Core Web Vitals.

Set up PostHog to track page views, time on page, scroll depth, and conversions (signup, download) for each AI-generated post.

Set pass threshold: ≥300 total page views and ≥3 conversions from the 5 posts within 3 weeks.

Share each post on LinkedIn and Twitter; track social referral traffic in PostHog to measure distribution effectiveness.

After 3 weeks, analyze performance in PostHog: identify top-performing post by traffic and conversions; note what made it successful (topic, format, CTA placement).

Compare AI editing time vs. writing from scratch; document time savings and quality trade-offs.

If you hit the threshold, document the prompt template and editing process and proceed to Baseline; if not, refine prompts or topic selection and retest.