How I batch-write a month of blog posts with AI without it sounding like AI

Generic AI-written blogs are easy to spot — same rhythm, same hedge-everything tone, same “in today’s fast-paced world” openers. Yet AI genuinely helps me write faster. The difference is process, not the tool. Here’s the exact monthly batching routine I run for this site.

Step 1: the topic list comes from real gaps, not a keyword tool

Every batch starts by checking what I’ve already covered — I pull my existing post titles and slugs and ask which real, useful angles are still missing, favoring comparisons, explainers, and how-tos over generic listicles. Writing about something I haven’t actually done or tested gets cut here, before a word is drafted.

Step 2: I write the take first, AI writes the draft second

This is the part that actually prevents the “sounds like AI” problem: I decide the opinion, the verdict, the specific thing I tested, before prompting. The AI prompt includes my actual conclusion (“Perplexity wins for research, ChatGPT for conversation — here’s why”) rather than “write a comparison of X and Y.” An AI with no opinion to work from defaults to safe, hedged, generic prose. Give it your specific take, and it writes toward that instead.

Step 3: batch-draft, don’t batch-publish

I draft 5–7 posts in one sitting using a small set of reusable prompts, but I never publish a whole batch unedited. Drafting in batch is efficient because the voice and structure stay consistent across posts written back-to-back; publishing unedited is how a site ends up with 40 posts that all read identically.

Step 4: the honesty pass

Every draft gets one specific edit: find the sentence that sounds like marketing copy, and replace it with something true and specific — a real limitation, a real number, a real “this didn’t work for me.” This is the single edit that does the most to make a post not read like AI content, because generic AI text almost never volunteers a downside unprompted.

Step 5: internal links, added by hand

AI drafts don’t know my full back catalog, so every post gets 3–5 manual internal links to existing guides plus a link to the Library hub. This step is also what keeps a growing site from turning into 80 orphaned islands — a lesson I learned the hard way when most of this site’s early posts had zero real internal links and sat un-indexed for months.

The honest limitation

This process is slower than pure “prompt and publish” — a real batch still takes hours, not minutes. If your goal is volume over quality, an AI content mill will always out-produce this approach. Whether AI content ranks increasingly depends on exactly this gap: genuine specificity vs. generic filler, and Google is getting better at telling them apart.

Related reading

• Is AI Content Good Enough to Rank on Google in 2026?
• The 10 Claude Prompts I Copy-Paste Every Week as a Content Creator
• Browse every AI guide — the Library