30 SEO Production Prompts: AI Workflow Templates 2026

# 30 SEO Production Prompts: AI Workflow Templates 2026

TL;DR: 30 production-ready prompts covering all 6 SEO workflow phases. Each tested on Claude Sonnet 4.6 and GPT-4o, with expected input and output type. Copy-paste into your workflow or custom GPT. For the methodological skeleton: SEO Workflow Guide 2026.

Research Prompts (1–5)

Prompt 1 — Keyword expansion

"Target keyword: [KEYWORD]. Generate 25 probable long-tail variants (informational + commercial intent), grouped by intent. Format: 3-column markdown table (keyword, estimated volume, intent)."

Prompt 2 — Intent classifier

"For these 20 keywords: [LIST], classify intent as informational / commercial / transactional / navigational. Justify in one sentence each."

Prompt 3 — PAA extractor

"Give me 12 probable People Also Ask questions for the query [KEYWORD], ranked by estimated frequency on Google US/UK 2026."

Prompt 4 — Competitor gap

"These are the 5 titles from my top SERP competitors on [TOPIC]: [LIST]. What 8 angles are they NOT covering that would have business value for a site that [POSITIONING]?"

Prompt 5 — Keyword cluster

"Here are 40 keywords: [LIST]. Group them into semantic clusters. For each cluster: cluster name, head keyword, 4–8 supportings, dominant intent."

Brief Prompts (6–10)

Prompt 6 — SERP-driven outline

"Here are the 5 outlines from the top SERP for [KEYWORD]: [PASTE]. Generate a unified outline covering 90% of patterns + 2 differentiating sections. H1/H2/H3 markdown format."

Prompt 7 — Short brief (cluster article)

"Generate a 250-word brief for a cluster article on [KEYWORD]. Include: H1, 4–6 H2s, secondary keywords, 1200-word target, 3 angles to develop, 2 citable E-E-A-T sources."

Prompt 8 — Pillar brief

"Generate a full brief (800 words) for a pillar article on [HEAD KEYWORD]. Sections: intent, 8–12 H2 structure, sub-clusters to link, schema candidates, 2500-word target, sources, CTA."

Prompt 9 — TL;DR generator

"Given this brief: [PASTE]. Generate a TL;DR of 80–110 words summarizing the article's promise, mentioning the primary keyword once, ending with a clear angle statement."

Prompt 10 — FAQ candidates

"For an article targeting [KEYWORD] with angle [ANGLE], generate 6 natural user FAQ questions + answers of 60–90 words each. No bullet points in answers."

Production Prompts (11–18)

Prompt 11 — Section drafter

"Given this brief: [PASTE]. Write ONLY the [H2 NAME] section in 250–350 words. Style: direct, no filler phrases ('in today's digital landscape'). Include 1 sourceable figure and 1 concrete example. No conclusion to this section."

Prompt 12 — Style enforcer

"Rewrite this paragraph in direct professional style: no filler, short sentences, active verbs, remove vague adjectives. Same substance."

Prompt 13 — Expansion

"This section is 120 words and lacks depth: [PASTE]. Expand to 280–320 words adding: 1 figure, 1 example, 1 nuanced counterpoint. No fluff, no repetition."

Prompt 14 — Anti-hallucination

"Given this paragraph: [PASTE]. Identify each factual claim (figure, statistic, citation) and indicate: (a) probability it's accurate, (b) source to verify, (c) reformulation if unsourceable."

Prompt 15 — Tone of voice

"Here are 3 reference paragraphs of the house style: [PASTE]. Rewrite this text [PASTE2] in the same tone: sentence length, formality level, active/passive ratio, vocabulary."

Prompt 16 — Anecdote injector

"This section [PASTE] is technically correct but cold. Inject a realistic mini-case (3–5 sentences) illustrating the point with a character and a figure. Don't invent known brand names."

Prompt 17 — CTA generator

"For an article on [TOPIC] with audience [AUDIENCE], propose 3 CTA variants: (1) soft inform, (2) lead magnet, (3) free consultation. No prices mentioned."

Prompt 18 — First-person experience

"Rewrite this paragraph [PASTE] adding a first-person angle (author voice) with a personal observation or experience marker that demonstrates E-E-A-T. Keep < 50 words added."

Optimization Prompts (19–23)

Prompt 19 — Title optimizer

"Give 5 title tag variants for [TOPIC, ANGLE]. Constraints: ≤60 characters, primary keyword in first 30 characters, modifier (guide, 2026, complete) included. Table format."

Prompt 20 — Meta description

"For article [TITLE], generate 3 meta descriptions ≤155 characters, each with a different angle (curiosity, benefit, urgency) + implicit CTA."

Prompt 21 — Schema picker

"Given this article: [PASTE]. Recommend applicable schema.org types (Article, FAQPage, HowTo, ItemList, Review). Justify each."

Prompt 22 — Internal link suggester

"Given this article [PASTE] and my 80 other articles with slug + meta description: [LIST]. Suggest 5 relevant internal links with descriptive anchor text."

Prompt 23 — Alt text generator

"For an image illustrating [CONCEPT] in an article on [TOPIC], propose 3 alt texts ≤125 characters, descriptive and SEO-friendly."

Publication Prompts (24–26)

Prompt 24 — LinkedIn post

"For this article [TITLE + TLDR], generate a LinkedIn post 180–220 words: strong opening hook, 4–5 bullet points, article link CTA. Max 1 emoji in hook."

Prompt 25 — Newsletter intro

"Generate a newsletter intro (90–110 words) for this article: [TITLE + TLDR]. Conversational tone, 1 hook, 1 promise, 1 CTA 'read the full article'."

Prompt 26 — Tweet thread

"Convert this article [PASTE] into an 8-tweet max thread. Each tweet ≤270 chars. First tweet: strong hook. Last tweet: article link CTA."

Measurement Prompts (27–30)

Prompt 27 — GSC interpreter

"Here is a 90-day GSC export for [URL]: [DATA]. Identify: 3 opportunity queries, 2 risks (declining), 1 priority action."

Prompt 28 — Refresh diagnostic

"Here is a 2024 article [PASTE] and the current SERP for the same keyword [SERP DATA]. Identify: 5 sections to update, 3 sections to add, 2 claims to re-verify."

Prompt 29 — Cannibalization detector

"These 4 articles target close keywords [DATA]. Is there cannibalization? If yes: which to keep as primary target, which to 301 redirect, which to reframe as sub-cluster?"

Prompt 30 — Monthly SEO report

"Here are the month's KPIs [DATA: traffic, top pages, top queries, conversions]. Generate an executive report of 250 words: highlights, lowlights, 3 next-month actions."

---

How to Use This Library

1. Create a Claude Project or Custom GPT with these 30 prompts as system context 2. Version your prompts (Git or Notion) — note which adjustments work 3. Measure real time savings per prompt over 10 uses 4. Eliminate prompts with < 5 min saved per use

For use in an industrialized pipeline, see SEO Content Automation Pipeline 2026 (Layer 3 — Generation).

Articles liés sur ce thème