I've been working on one thing this past year: using AI to deliver real work, not just run experiments. Last week, I put that into practice by rebuilding my entire personal website using Claude's Cowork mode — moving from WordPress to clean, static HTML.
What I started with
My resume, my career background, my LinkedIn About section, and a photo. No design brief. No specs. Just a clear vision: clean, minimal, professional.
Here's what actually made the difference
I didn't just prompt. I directed.
- When the first draft felt safe, I pushed for a bolder redesign.
- When it offered layout options, I made intentional choices and explained why.
- When the copy drifted into generic territory, I forced it back to my real metrics — $100M+ portfolios, work across 3 continents — all sourced directly from my resume.
Then I gave it a harder challenge: turn 40+ LinkedIn recommendations from 17 years into a polished, dedicated page. It built a clean card layout, organized everything, and when the grouping created ambiguity, I caught it and had it rework the structure.
The finished site is live at www.kolaganti.com.
The real lesson
The value isn't that AI can build a website. The value is knowing how to work with AI — what to specify, when to push back, when to redirect, and when to accept. The better you can articulate what you want, the better the result. Simple as that.
For senior leaders, these are skills you already have: scope definition, alignment, quality review, iteration. The tools look different now, but the underlying discipline is the same.
If you're in leadership or program management and haven't explored how AI fits into your workflow yet, it's worth the time. It won't do your thinking for you, but it will do a lot of the heavy lifting.
What are you experimenting with?