Claude Prompt Templates for Blog Posts

Generic AI blog posts get generic results -- somewhere between page 8 and never. The reason isn't the AI; it's that most blog post prompts don't specify the one thing that separates ranking content from invisible content: a specific angle that makes this post genuinely more useful than what's already on page one. A well-built claude prompt template for blog posts forces that specificity before a single word of the draft is written.

Quick-Start (Copy This Right Now)

I'm writing a blog post for [target audience] about [topic].

My angle -- what makes this post different from what's already ranking: [describe your unique angle, experience, or data point]
Primary keyword: [keyword]
Publication: [your blog/site name and brief description of what it covers]

Step 1 (do this first): Propose 3 possible outlines for this post. For each, name: the structural logic, the first paragraph's opening hook, and the one section that would be genuinely hard for a competitor to replicate.

I'll choose an outline before you write the draft.

What this does: Proposing three outlines rather than one gives you meaningful creative input before any substantial content is written -- you're choosing a direction, not just reviewing a draft and accepting it or starting over. The "hard to replicate" question forces each outline to have something genuinely differentiating.

Understanding the Variables

The "angle" field is the most important and most skipped. "I'm writing about remote work productivity" is a topic. "I'm writing about why the standard advice about remote work productivity is counterproductive for people with ADHD, based on my six months testing it" is an angle. The first gets you a generic roundup article. The second gets you content that could actually win a reader who's already read the generic roundup.

If you don't have a clear angle, ask Claude for help before running the main prompt:

I want to write about [topic]. What angles on this topic would be genuinely hard for someone without [expertise/experience/data] to write? What would require something most content on this topic currently lacks?

⚠️ Common mistake: Pasting your keyword and expecting the outline to inherently include a reason for someone to read past the headline. Keywords describe what a piece is about. Angles describe why it's worth reading instead of the five other articles about the same thing. These are different problems requiring different inputs.

Step-by-Step: Full Draft from Approved Outline

Use this outline: [paste chosen outline]
Voice sample (write like this): [paste 200-word sample of your actual writing]
Primary keyword: [keyword] -- use it naturally in the first 100 words, at least one H2, and 3-4 times total

Draft the full post (1,200–1,800 words). For each section:
- Start with the key point of that section, not a transition phrase
- Use specific numbers and named examples rather than vague claims
- Flag [NEEDS YOUR EXAMPLE HERE] anywhere I should add personal experience or proprietary data rather than inventing it

SEO instructions: Every H2 should answer a question a reader would type into Google. No H1 in the body -- it's rendered from the title field.

What this does: The

[NEEDS YOUR EXAMPLE HERE]

⚡ Pro tip: After the draft is complete, run one more pass: "Read this draft and tell me: which paragraph sounds most like AI-generated content, and why? What would make it sound more like the voice sample I gave you?" This catches voice drift that accumulates across a long draft in ways that are hard to notice while writing.

Pro-Level Variations

Intro rewrite for stronger hook:

Here's my blog post intro: [paste]
Here's the hook style I want: [pick one -- surprising stat, counterintuitive claim, specific scenario, direct question the reader is already asking]

Rewrite the intro using that hook style. Don't change anything after the first three sentences -- just the opening. Give me two versions so I can compare.

Meta description and title tag:

Write a meta title (under 60 characters) and meta description (under 155 characters) for this post about [topic] targeting [keyword].

The meta description should make a specific promise about what the reader gets -- not "learn everything about X" but "here's exactly what this post delivers." Three variations, each with a different approach: one benefit-focused, one curiosity-based, one using a specific number or detail from the post.

Troubleshooting Common Issues

If drafts consistently need heavy editing, check the voice sample first. A 200-word sample that's actually representative of how you write (informal, technical, opinionated, whatever your real style is) produces dramatically better first drafts than a sample you picked because it seemed "professional" rather than authentic.

If posts feel long but shallow, the outline is usually the problem. Add a field to the outline step: "For each section, name the single most specific claim this section makes." Sections that can't answer that question with something specific tend to be filler rather than content -- find and remove them before drafting.

Three Real Scenarios

A content strategist managing a blog with 12 contributors used the three-outline structure to standardize the angle-selection step across writers who had wildly different instincts about what was interesting. By requiring every writer to propose three outlines before drafting, the review step moved from "I don't think this angle works" (subjective and hard to give) to "I'm choosing outline 2, which has the strongest differentiating section" (specific and actionable). The rejection rate of first drafts dropped significantly.

A solo consultant building a thought-leadership blog used the

[NEEDS YOUR EXAMPLE HERE]

A developer relations manager used the meta-description variation prompt to actually A/B test title tags -- generating three variations for each post and testing them in a small batch of social shares before finalizing the SEO metadata. The specificity-focused variation (using a specific number from the post) consistently outperformed the curiosity-based variation for their technical audience, which was a genuine finding they wouldn't have discovered using a single-pass prompt.

A Post-Draft Quality Check

I've written a blog post draft: [paste]

Run three checks:
1. Read the first 150 words -- does it earn the reader continuing? What's the specific reason to keep reading, and is it stated within the first three sentences?
2. Identify the single weakest paragraph -- the one with the vaguest claim or the least useful information. Quote the first sentence.
3. Does the post deliver on whatever the headline promises? If not, what's the gap?

Don't rewrite anything -- just identify the three issues so I can fix them myself.

Your Turn

A claude prompt template for blog posts that's calibrated to your specific voice, audience, and SEO strategy becomes more valuable the more you use it, because you refine it with each post that needed heavy editing. Storing the current best version somewhere accessible -- PromptABCD keeps prompt versions organized and shareable across a team -- means each post starts from the best version of your template, not a reconstructed approximation of it.

The angle pre-work -- deciding what makes this post genuinely different before writing a word -- is the investment most blog post processes skip because it feels like it should happen automatically. It doesn't. Spending ten minutes on the angle question before opening any prompt produces better outlines, better drafts, and less editing than spending those same ten minutes cleaning up a draft that started from a weak angle. The template is what makes that discipline consistent rather than occasional.

The post-draft quality check -- specifically the three-point version that flags the weak paragraph and checks whether the post delivers on the headline's promise -- is worth running on every post before publishing, not just ones that felt difficult to write. Posts that were easy to write are often the ones with the weakest individual paragraphs, because ease of writing sometimes signals you covered familiar ground rather than genuinely useful ground. That check catches the difference.