AI workflow

For 2 months, I’d open Claude, type a topic, and ask for a draft. No outline. No research pass. No structure beyond whatever Claude improvised on the spot. Some of those articles were fine. Most weren’t — not because the writing was weak, but because the thinking behind it was thin. I’d cover the obvious … Read more

For a while, my content gap process was just me, six competitor tabs open, and a notebook. I’d read an article, feel like something was missing, and try to guess what. That’s not a process. That’s improvising with extra steps. Claude didn’t fix that by handing me a “here are 20 topics you’re missing” list. … Read more

For the first stretch of running my site, I published whatever I thought would rank. A prompt guide here, a comparison post there — useful on their own, disconnected from everything else. Google treated me accordingly: individual pages ranking inconsistently, nothing building on anything else. [PLACEHOLDER: swap in the specific moment that made you rethink … Read more

I almost published a paragraph that sounded completely correct. The tone was confident. The structure was clean. Every sentence read like something I’d already checked myself. Then I went looking for the source it was supposedly built on. I couldn’t find anything that actually said what the paragraph claimed. Nothing about that draft looked wrong. … Read more

The fastest way to ruin a good blog post is to ask Claude to “refresh it.” I’ve done exactly that. I’d paste one of my published articles into Claude, ask it to update everything for the latest changes, and get back a cleaner draft in minutes. The facts were current. The grammar was better. The … Read more

I spent three months rewriting the same prompt. Longer instructions. More examples. A paragraph explaining exactly how I wanted my sentences to sound. Every version came back more polished — and somehow less like me. I finally figured out the problem wasn’t the prompt. It was everything I wasn’t telling Claude: sentence rhythm, the words … Read more

Most articles comparing Claude and ChatGPT spend their word count on pricing tiers, context windows, and benchmark scores. That information matters to developers and product teams. It matters a lot less to writers who need to produce 2,000 words of clean, coherent content that sounds like it was written by a human. What writers actually … Read more

I spent six months trying to make Claude sound like me. I uploaded articles. I pasted in style notes. I wrote long prompt instructions explaining my tone, my sentence length, the way I open sections. The output improved. But it never felt right. It was close — the way a cover band is close. All … Read more