Templates

AI SEO Content Prompt Pack (Free + Copy-Paste Ready for 2026)

8 min read · Feb 19, 2026· AO Network Editorial Team

AI changed SEO content production faster than most teams have adjusted to. The teams that win in 2026 are not using AI to write more articles. They are using AI to do the parts of the SEO workflow that used to take hours and rarely got done well.

The six prompts below cover those parts. Keyword research, SERP-derived briefs, content optimization scoring, gap analysis, headline testing, and refresh prioritization. Copy any prompt directly from the page.

If your SEO program is still figuring out cadence, the always-on SEO playbook is the prerequisite. For the broader prompt library across marketing workflows, see the AI prompt library.

What AI is good at in SEO

Three things, specifically. Pattern detection across SERP results. Restructuring existing content into better-organized drafts. Scoring content against measurable criteria.

What AI is bad at: producing original perspective, knowing what your customers actually care about, and judging whether a piece is worth writing in the first place. The prompts below stay in the lane where AI helps.

Keyword research

Prompt

Keyword cluster builder from a seed list

Recommended model: Claude Sonnet 4.7 or GPT-5

I will paste 50 to 200 keywords I pulled from [Ahrefs / Semrush / Search Console] below. The seed keyword is [SEED KEYWORD].

Your job is to cluster these into topical groups that can be covered by individual pages.

Return:

1. A list of 5 to 10 topical clusters. For each cluster:
   - The cluster's parent topic in one sentence
   - The 5 to 15 keywords that belong to it
   - The single best target keyword for a pillar piece on this cluster
   - 2 to 4 supporting keywords for follow-on articles

2. Keywords that did not fit any cluster (orphans). These are either off-topic or potential new clusters that need more research.

3. Two to three cluster suggestions that the data hints at but is incomplete on. Where would I research next to confirm?

Do not invent clusters that the keyword list does not support. If a cluster has fewer than 3 keywords, do not create it.

Keywords:
[PASTE KEYWORDS, ONE PER LINE]

Prompt

Search intent classifier

Recommended model: Claude Sonnet 4.7

I will paste 50 keywords below.

For each keyword, classify the search intent:

- I (Informational): the user wants to learn something
- N (Navigational): the user wants to find a specific site or brand
- C (Commercial investigation): the user is evaluating options before buying
- T (Transactional): the user wants to buy or sign up

For each keyword, return: keyword, classification, the SERP feature most likely to dominate (featured snippet, video carousel, product listings, etc.), and a one-line note on why.

Flag any keyword where the classification is ambiguous. Those are usually worth their own analysis.

Keywords:
[PASTE 50 KEYWORDS]

Content briefs from SERP analysis

Prompt

SERP-derived content brief

Recommended model: Claude Sonnet 4.7 or GPT-5

I am targeting the keyword [TARGET KEYWORD]. I will paste the content from the top 5 ranking pages below.

Build a content brief that beats them.

Return:

1. Topical coverage map: every subtopic the top 5 pages address. Mark each:
   - Frequency (how many of the 5 pages cover it)
   - Depth (shallow, medium, deep on average)
   - Differentiation opportunity (where we can go deeper or take a different angle)

2. Coverage gaps: subtopics the audience would expect that none of the top 5 pages address.

3. The angle we should take. One sentence. Should be different from at least 3 of the top 5 pages.

4. Recommended structure: 6 to 9 H2 sections in the order they should appear.

5. Word count target. Compute the median of the top 5 word counts and add 20%.

6. Internal linking opportunities: list of 5 to 10 specific topics on aonetwork.com that this piece should link to.

7. Title options: 5 candidates that signal the angle and contain the target keyword.

Do not just summarize the existing top pages. The brief should produce a piece that is better than each of them on at least one dimension.

Top 5 ranking pages:
Page 1 (URL: [URL]):
[PASTE CONTENT]
---
Page 2 (URL: [URL]):
[PASTE CONTENT]
---
Page 3 (URL: [URL]):
[PASTE CONTENT]
---
Page 4 (URL: [URL]):
[PASTE CONTENT]
---
Page 5 (URL: [URL]):
[PASTE CONTENT]

Content optimization

Prompt

On-page optimization scorer

Recommended model: Claude Sonnet 4.7 or GPT-5

I will paste the draft of an article below. The target keyword is [TARGET KEYWORD].

Score the draft on the following criteria, 1 to 10:

1. Keyword in title (presence, placement, naturalness)
2. Keyword in first 100 words (yes/no)
3. Keyword density (computed across the body)
4. H2 and H3 structure (sequential, descriptive, scannable)
5. Internal linking depth (number of internal links, contextual placement)
6. Topical coverage breadth (compare to the brief if I paste one)
7. Readability (sentence length variation, paragraph length, complexity)
8. Original perspective (does the piece have a sentence-form claim no other ranking page would make?)
9. FAQ section presence and quality
10. Conclusion that is not a generic summary

For each, give the score, a one-sentence reason, and the specific fix.

Then give a summary verdict: ship as is, ship with edits, or rebuild. The threshold for shipping is an average score of 7.5+.

Draft:
[PASTE DRAFT]

Optional brief from earlier prompt:
[PASTE BRIEF IF AVAILABLE]

Gap analysis and refresh

Prompt

Content gap analyzer

Recommended model: Claude Sonnet 4.7

I will paste a list of articles I have already published on a topic cluster below.

The cluster is [CLUSTER PARENT TOPIC, e.g., 'always-on email marketing'].

Return:

1. Coverage map: subtopics within this cluster that my existing articles address. For each, note which article covers it and how deeply.

2. Gap list: subtopics the cluster should logically cover that my existing articles do not. For each gap, propose a target keyword and a working title for an article that would fill it.

3. Overlap warning: articles that may be cannibalizing each other (covering the same subtopic). For each pair, recommend whether to consolidate, redirect one, or differentiate them.

4. Refresh candidates: articles older than 12 months that should be updated. Rank by potential ROI of refresh (think: stale data, missing recent context, top-of-funnel articles that have been overtaken by competitors).

Do not invent gaps that are not natural to the cluster. Do not suggest articles I would not realistically write for my audience.

My audience is [VP / Head of Marketing at B2B SaaS, 100 to 1,000 employees].

Existing articles:
[PASTE ARTICLE TITLES + SHORT DESCRIPTIONS, ONE PER LINE]

Headlines

Prompt

Headline rewriter with intent matching

Recommended model: Claude Sonnet 4.7 or GPT-5

I have a draft headline: [DRAFT HEADLINE].

The target keyword is [TARGET KEYWORD]. The search intent is [Informational / Commercial investigation / etc.].

Generate 10 alternative headlines following these rules:

- Include the target keyword in the first 60 characters
- Match the search intent (no clickbait for informational searches)
- Vary the patterns across the 10: 'How to', 'What is', 'X vs Y', listicle, question form, benefit-led, etc.
- No marketing tropes: 'ultimate guide', 'everything you need to know', 'definitive'
- Vary length: include 3 short (under 50 characters) and 3 long (60+ characters)

For each, give a one-line note on which audience or intent angle it serves.

Then rank the 10 by predicted CTR for the target intent. Be honest about uncertainty.

The brand voice is direct, opinionated, skeptical of marketing tropes.

What this pack does not include

AI prompts for backlink outreach. The category exists and works for some teams. The output quality has not been good enough to recommend yet.

AI prompts for technical SEO audits. The right tools (Screaming Frog, Sitebulb, the best SEO tools) already produce the audits. AI is the wrong layer for that work.

Full-article generation prompts. The AI content brief template covers what brief to give the AI. The actual writing happens separately and benefits from a different prompt than the SERP-derived brief above.

How to actually use these in a workflow

Pick one cluster you publish in. Run the keyword research prompts to expand your target list. Run the intent classifier to prioritize informational vs commercial pieces. Run the SERP-derived brief on the highest-priority keyword. Pass the brief to the AI content brief template for the actual draft. Run the optimization scorer after writing. Ship.

Compare this to the manual version of the same workflow. Three to four days of analyst time. The AI-assisted version is two to three hours of operator review on top of AI output.

Frequently asked questions

Can these prompts replace an SEO specialist?

No. They make an SEO specialist 3 to 5x more productive. A specialist who knows what good looks like can run these prompts confidently. Someone without that judgment will accept bad output.

Do these work for ecommerce SEO?

The keyword research and intent classifier prompts transfer. The SERP-derived brief and optimization scorer prompts are tuned for content pages and need adjustments for product pages.

How does this fit with the content marketing playbook?

The playbook decides what content to produce. These prompts compress the production. The prompts cannot substitute for strategy; they make execution faster.

Which of these prompts is closest to a task your team currently underinvests in because it takes too long? That is the one to start with.

The Always-On Brief

Weekly strategy, tool picks, and playbooks. 6,000+ marketers subscribed.

Templates

Retention Cohort Analysis Template

Jul 24, 2026

Templates

Incrementality Test Plan Template

Jul 9, 2026

Templates

Marketing Budget Template: An Allocation Worksheet

Jul 5, 2026