Publishing & AI Agents
Multi-Vertical Content Pipeline
End-to-end content pipeline for a digital publisher operating across multiple verticals. The system monitors RSS feeds and news sources, identifies relevant stories, generates SEO-optimised articles using Claude, and publishes them with proper metadata, structured data, and category tagging. Scaled from one vertical to five with zero additional headcount.
The challenge
What we were solving
A digital publisher had proven strong demand across five content verticals but couldn't justify hiring separate editorial teams for each. Traditional content operations require writers, editors, and SEO specialists per vertical - making expansion prohibitively expensive. The client needed a way to scale content production from one vertical to five without proportionally scaling headcount or sacrificing quality.
Our approach
How we built it
We designed a modular content pipeline where each vertical operates as an independent channel with its own RSS sources, topic filters, and tone configuration. Claude processes each story through vertical-specific prompts that maintain editorial consistency. Quality controls at every stage include source attribution verification, fact-checking against original material, and tone analysis. The pipeline architecture is designed for horizontal scaling - adding a new vertical requires only configuration, not code changes.
Results
What we delivered
100+ articles published per week across 5 verticals
90% reduction in editorial time per article
SEO-optimised output with structured data and dynamic OG images
Scaled from 1 to 5 verticals with no additional staff
Quality controls including fact-checking and tone consistency
Tech stack
Built with
Deep dive
The full picture
Digital publishers scaling across multiple verticals traditionally need to hire editorial teams for each new vertical - writers, editors, and SEO specialists. This client had proven the model with one vertical but couldn't justify the headcount to expand into four more niches that showed strong demand.
We built a content pipeline that handles the entire editorial workflow autonomously. The system monitors RSS feeds and news sources across all five verticals, uses AI to identify the most newsworthy stories, and generates SEO-optimised articles through Claude with consistent tone, proper attribution, and structured data. Each article is tagged by category, enriched with relevant keywords, and published with dynamic OG images.
Quality controls are built into every stage of the pipeline. Fact-checking routines cross-reference key claims against source material. Tone consistency is maintained through carefully engineered prompts that match the client's editorial voice. The system scaled from producing 20 articles per week in one vertical to over 100 articles per week across five verticals - with no additional editorial staff.
More work
Explore other projects
PropertyNews.io - Autonomous AI Publishing Platform
Built a fully autonomous property news platform that sources, rewrites, and publishes 16+ articles daily across 8 categories. The platform uses an AI-powered content pipeline - RSS feeds are ingested, processed through Claude, and published to a custom Next.js frontend with zero human intervention. Includes city-level landing pages with live market data, SEO infrastructure, and an admin dashboard for monitoring pipeline health.
Read case study →Custom SaaS & AutomationBloodstone.co.uk - AI-Built Consultancy Website
Designed and built the Bloodstone Projects website itself using AI-assisted development. Full-stack Next.js application with dynamic service pages, programmatic SEO landing pages for 5 industries and London, integrated contact form with Supabase and automated email workflows, 14 blog articles, FAQ schema markup, and dynamic OG image generation. Built and deployed in days, not months.
Read case study →Hospitality & Custom SaaSRestaurant Group Operations Dashboard
Real-time analytics dashboard for a multi-site restaurant group consolidating data from POS systems, delivery platforms (Deliveroo, UberEats, JustEat), and reservation systems into a single view. Replaced three separate tools and eliminated the Monday morning spreadsheet scramble. Automated daily P&L reports delivered to management via email.
Read case study →Got a similar challenge?
We build AI agents, automation workflows, and custom tools that solve real business problems. If this case study resonated, let's talk about what we can build for you.
Start a conversation