Introduction
SEO optimization is no longer just about keyword stuffing or guessing what Google wants. In 2026, search engines rely on sophisticated algorithms that evaluate content quality, semantic relevance, and technical signals. The challenge for developers and content creators is to produce optimized material consistently — without spending hours on manual research or rewriting.
This is where prompts come in. By using the right prompts with AI tools like ChatGPT, Claude, or specialized SEO assistants, you can automate routine SEO tasks: generating meta tags, structuring content for featured snippets, and auditing page structure for crawlability. Below is a battle-tested collection of 15 prompts I use in my own workflow, each with a real usage example and context.
Prompts for Meta Tags (Titles and Descriptions)
1. Title Tag Generator
Prompt:
Generate 5 title tag options for a blog post about [topic]. Each title must be under 60 characters, include the primary keyword '[keyword]' at the beginning, and follow a 'How to' or 'Number + Adjective + Noun' pattern. Output as a bullet list.
Example: For a post about 'cloud cost optimization', with keyword 'AWS savings'. Output includes titles like 'How to Reduce AWS Costs in 2026' and '7 Proven AWS Savings Strategies'.
2. Meta Description from Content
Prompt:
Write a meta description (150–160 characters) for the following article. Include the primary keyword '[keyword]' and a clear value proposition. Use active voice and a call to action. Content: [paste first 300 words].
Example: For a SaaS onboarding guide, the description becomes 'Learn how to reduce churn with a 5-step onboarding flow. Discover proven tactics for SaaS user retention in 2026.'
3. Open Graph and Twitter Card Copy
Prompt:
Create an Open Graph title (max 60 chars) and description (max 200 chars) for this page. Also provide a Twitter Card title and description (separate). Primary focus: [topic]. Target audience: [audience].
Example: For a product launch page, you get separate optimized text for LinkedIn, Facebook, and Twitter sharing.
Prompts for Content Optimization
4. Keyword-Optimized Outline
Prompt:
Act as an SEO content strategist. Create a detailed H2/H3 outline for a 2000-word article targeting the keyword '[keyword]'. Include semantic LSI terms, common questions from 'People Also Ask', and a section for '[keyword] vs [related term]'. Format as markdown headings.
Example: Targeting 'serverless monitoring tools' yields sections on 'Top 5 Serverless Monitoring Tools in 2026', 'How to Choose a Tool', and 'Serverless vs Traditional Monitoring'.
5. Content Gap Analysis
Prompt:
I have a page about [topic]. Compare it with the top 3 Google results for '[keyword]'. List missing topics, questions not answered, and content angles I should add. Output as a table with columns: Gap, Priority (High/Medium/Low), Suggested Action.
Example: For a 'React state management' page, the analysis reveals missing coverage of Zustand and Jotai, which competitors include.
6. FAQ Schema Content
Prompt:
Generate 5 FAQ items with questions and answers (each answer 40–60 words) for the topic '[keyword]'. The questions should match common 'People Also Ask' queries. Output in JSON-LD format with @context and @type.
Example: For 'Docker networking', you get structured data ready to paste into the page.
Prompts for Technical Structure
7. Internal Linking Suggestions
Prompt:
Analyze this page content: [paste content]. Suggest 5 internal links to other pages on my site (list existing URLs). For each link, propose anchor text and explain relevance. Prioritize links that improve topical authority.
Example: For a 'Python decorators' guide, links to 'Python functions' and 'Advanced Python patterns' are suggested.
8. URL Slug Optimization
Prompt:
The current URL slug is '[current slug]'. Suggest 3 alternative slugs that are shorter, include the primary keyword '[keyword]', and remove stop words. Explain why each is better for SEO.
Example: Changing '/blog/how-to-use-aws-lambda-for-backend' to '/aws-lambda-backend-guide'.
9. Heading Structure Audit
Prompt:
Audit the heading structure of this HTML: [paste HTML]. Check for: missing H1, multiple H1s, skipped heading levels (e.g., H2 to H4), and non-descriptive headings. Output a table with issues and fixes.
Example: Finds an H1 missing on a product page and suggests adding one containing the primary keyword.
Prompts for Structured Data & Schema
10. Article Schema Generator
Prompt:
Generate JSON-LD Article schema for this page. Include headline, description, datePublished, dateModified, author, publisher, and image. Use this data: [provide title, date, author name]. Output only valid JSON.
Example: You get a ready-to-copy snippet that adds rich results in Google.
11. BreadcrumbList Schema
Prompt:
Create BreadcrumbList schema for a site with this structure: Home > Category > Subcategory > Page. Use these URLs: [list]. Output JSON-LD.
Example: For an e-commerce site with 'Home > Electronics > Laptops > Gaming Laptops', you get structured breadcrumbs.
12. HowTo Schema from Instructions
Prompt:
Convert the following step-by-step instructions into HowTo schema JSON-LD. Steps: [list steps]. Include a total time and tool list. Output valid JSON.
Example: For a 'How to install Node.js' guide, you get schema that can appear as a rich snippet.
Prompts for Performance & Crawlability
13. Robots.txt Suggestions
Prompt:
My site structure is [describe]. Suggest robots.txt rules to block duplicate content (e.g., filter pages, pagination) and allow access to key sections. Output as plain text with comments.
Example: Blocking /tag/ and /category/ while allowing /blog/ and /products/.
14. Canonical Tag Check
Prompt:
I have these two URLs: [URL A] and [URL B]. They have similar content. Which should be the canonical? Explain based on content depth, internal links, and traffic data.
Example: Recommends URL A as canonical because it has more backlinks and higher organic traffic.
15. Core Web Vitals Improvement Prompt
Prompt:
Analyze this page's Lighthouse report: [paste summary]. Suggest 3 specific, actionable improvements for LCP, FID, and CLS. Prioritize changes that can be implemented in under 2 hours.
Example: For LCP issues, suggests lazy-loading hero images and preloading the primary font.
Case Study: Real Implementation
Problem: A mid-size e-commerce site (1000+ product pages) had poor click-through rates (CTR) from search results — average 2.1% for product pages. Meta descriptions were generic ('Buy [product] online').
Solution: Using prompts #1–3, the team generated unique meta titles and descriptions for the top 200 product pages. Each meta description included a specific benefit (e.g., 'Free shipping over $50', '30-day returns'). They also added HowTo schema (prompt #12) for product usage guides.
Result: After 6 weeks, average CTR increased to 3.8% (an 81% relative improvement). Organic traffic to those pages grew by 34%. The structured data also generated 12 product-rich snippets in Google.
Tools used: ChatGPT for prompt execution, Screaming Frog for initial audit, and Google Search Console for tracking. ASI Biont supports connection to Google Search Console via API — more details at asibiont.com/courses.
Conclusion
These 15 prompts cover the core areas of SEO: metadata, content structure, technical audit, and schema markup. The key is to adapt them to your specific tools and workflows — whether you use ChatGPT, Claude, or a dedicated SEO AI. Start with the meta tag prompts (1–3) for quick wins, then move to content optimization (4–6) and technical structure (7–9). Structured data (10–12) and performance prompts (13–15) are for more advanced optimization.
Remember: prompts are only as good as the context you provide. Always include the target keyword, audience, and any existing content. Test each prompt with a few pages, measure CTR and rankings, and refine. In 2026, the difference between average and great SEO is often just a well-crafted prompt.
Comments