A client came to us with a clear problem: their competitor had 200+ pages of SEO content capturing long-tail traffic, while they had 30 blog posts. They were losing thousands of potential visitors every month to content they simply had not created yet. They needed to close the content gap fast. We produced 50 SEO-optimized pages in one week — and 38 of them were ranking within 60 days. Here is the full case study.
The Strategy: Integration Pages at Scale
The client was a project management SaaS tool with 200+ integrations. Their competitors had dedicated pages for each integration — "[Competitor] + Slack integration," "[Competitor] + HubSpot integration" — all capturing search traffic from users looking for specific integration capabilities. Our client had a single "Integrations" page listing all partners. This was the opportunity.
The target: Create dedicated pages for the 50 highest-search-volume integration partners, each optimized for the "[Client Product] + [Partner] integration" keyword pattern.
Day 1: Data Collection and Template Design
Data Collection
We built an N8N workflow to collect data for each integration:
- Partner product description and key features (from the partner's website via web scraping)
- Integration capabilities (from the client's integration documentation)
- Common use cases for combining the two products (from the client's support team and community forums)
- Search volume and keyword variations (from Ahrefs API)
- Competitor page analysis — what do competitor integration pages cover? (Claude AI summarized competitor pages)
Template Design
We designed a page template with Claude AI that balances SEO requirements with user value:
- H1: [Client Product] + [Partner] Integration — keyword-rich, clear
- Introduction (150 words): What both products do and why integrating them matters
- Key integration features (300 words): Specific data flows and capabilities
- Top use cases (400 words): 3-4 practical use cases with examples
- How to set up (200 words): Quick setup guide with links to documentation
- FAQ section (200 words): 3-4 common questions (targets featured snippets)
- Related integrations (100 words): Internal links to related integration pages
Target word count per page: 1,200-1,500 words. Enough for SEO value, not so much that the content is padded.
Need help building your GTM systems? I build outbound and pipeline systems for B2B companies - and get results in 30 - 60 days.
Days 2-3: Content Generation
We ran the N8N content generation pipeline in batches of 10 pages:
- N8N loads the data profile for 10 integrations
- For each integration, Claude AI receives the template structure, the collected data, SEO guidelines (keyword density, heading structure, internal linking), and three examples of our best manually-written integration pages as style references
- Claude generates the full page content as structured HTML
- A validation node checks: word count above 1,200, all template sections present, target keyword appears in H1 and first paragraph, at least 3 internal links included
- Pages that pass validation are stored; pages that fail are flagged for prompt adjustment and re-generation
We processed all 50 pages across 5 batches over two days. The validation pass rate was 82% on the first run, with 9 pages requiring prompt adjustments and re-generation.
Day 4: Quality Review and SEO Optimization
Two team members each reviewed 25 pages, spending approximately 10 minutes per page:
- Verified factual accuracy of integration capabilities (this is critical — AI can hallucinate features)
- Improved introductions for pages where Claude's opening felt generic
- Added specific screenshots or diagrams for complex integrations
- Verified internal links pointed to correct pages
- Optimized title tags and meta descriptions for click-through rate
Separately, an N8N workflow handled technical SEO optimization:
- Generated unique schema markup (SoftwareApplication schema) for each page
- Created Open Graph tags for social sharing
- Generated alt text for any images
- Built the XML sitemap update to include all new pages
Day 5: Publication and Indexing
We published all 50 pages via the CMS API in two batches of 25, spaced 12 hours apart. Post-publication steps:
- Submitted updated sitemap to Google Search Console
- Used the URL Inspection API to request indexing for all 50 pages
- Added internal links from existing high-authority pages (homepage, main integrations page, relevant blog posts) to the new integration pages
- Shared 10 of the most interesting integration pages on social media to generate initial traffic signals
Results After 60 Days
- Indexation: 47 of 50 pages indexed within 14 days (94% indexation rate)
- Rankings: 38 pages ranking on page 1-2 for their target keyword within 60 days
- Traffic: 4,200 organic visits per month from integration pages alone
- Conversions: 127 sign-ups attributed to integration pages in the first 60 days
- Time investment: ~40 person-hours total (compared to an estimated 250+ hours for manual creation)
- Cost: $2,800 (Claude AI API + team time) versus an estimated $25,000+ for agency content creation
Lessons Learned
- Data quality determines content quality: The pages where we had rich integration data (detailed feature descriptions, real use cases) performed significantly better than pages where the data was thin. Invest in data collection before content generation.
- Human review is non-negotiable: Three of the 50 pages had factual errors about integration capabilities. If we had published without review, those errors would have damaged trust with users who landed on those pages.
- Template iteration matters: Our first 10 pages were noticeably weaker than the last 10 because we refined the Claude AI prompt based on what worked and what did not in earlier batches.
- Internal linking drives indexation: The pages we linked from high-authority existing content got indexed 2x faster than pages with no incoming internal links.
This approach works for any repeatable content pattern — integration pages, use case pages, comparison pages, location pages. The system scales to hundreds of pages once the template and data pipeline are dialed in. We are not advocating for replacing all human writing with AI. We are advocating for using AI to handle the content that follows a repeatable pattern, freeing your human writers to create the thought leadership and narrative content that AI cannot match.