Week 34: Using AI to overhaul a marketing page

(through careful iteration and validation)

The Experiment

I had this important marketing page that wasn’t converting cold traffic. You know the type—visitors would land, squint at it for a few seconds, then bounce. The page was supposed to explain a complex solution, but instead it was creating more questions than answers.

I wanted to transform this underperforming page into something that actually converted, without needing a full redesign or months of back-and-forth with designers. Could ChatGPT and v0 help me create a data-driven overhaul that the team would actually want to implement?

The Process

Here’s how I tackled it:

Data Collection:

• Customer research from our support hub
• Intercom queries (the goldmine of “what are people actually confused about?”)
• Scroll analysis and ecommerce funnel drop-offs
• Team feedback and upcoming test briefs
• Screenshots of the current page that was causing all the problems

AI Analysis Framework:

• Cross-reference insights from different sources
• Rate each recommendation on a proof strength scale (★★★★★)
• Justify every change with actual evidence
• Write everything in “Insight → Action → Proof” format
• Check its own homework
• Loop back and validate recommendations

Content Strategy:

• Refine the messaging using actual customer language
• Build a coherent narrative based on winning testing of value proposition statements

The Outcome

After a day of back-and-forth iteration, I had:

• A prioritised insight list with supporting sources and ratings
• Section-by-section copy for both overview and features pages
• Updated above-the-fold brief with strategic CTA changes
• Implementation-ready wireframes
• A clear narrative that tied everything together

The best part? The team’s reaction was “this is great preparation.” This AI experiment delivered exactly what was needed; a comprehensive, evidence-backed plan that people actually wanted to implement.

Key Takeaway

AI tools like ChatGPT excel at synthesis and iteration when you make them check their work. The secret isn’t accepting the first output—it’s creating a feedback loop where the AI validates its own recommendations against your data. Combined with v0 for visualization, you can create compelling, actionable proposals that teams take seriously.

This wasn’t about replacing human judgment—it was about using AI to create better, more thorough preparation for human decision-making.

Pro Tips for AI-Powered Marketing Overhauls:

1. Start with ChatGPT for data synthesis and strategic recommendations
2. Use v0 to visualize your content strategy
3. Make the AI justify every recommendation with evidence
4. Focus on iteration and validation rather than accepting first outputs

What’s Next?

I’m planning to apply this same methodology to other underperforming pages, using the same data-driven approach to create evidence-backed recommendations that teams can actually implement.

The real value isn’t in the AI doing the work—it’s in creating a systematic approach to validation that ensures every recommendation is backed by real customer insights.