Systems design · SEO · AI orchestration

An SEO content engine that runs on AI agents and refuses to trust them.

A multi-brand content system where AI handles every judgment step, research, topical planning, writing, and adversarial review, while deterministic code handles everything that must be exact: spec conversion, style validation, image generation, and publishing.

Two views of the same system. Start wherever you like.

Interactive system map

The full architecture. Hover or tap any node to trace exactly what it reads and what it writes, across every phase from brand research to publishing.

Open the system map → View 2

Phase dependency map

What every phase actually reads, and why. Nodes are colored by kind, engine file, tool, or external input, and badged for whether a file is read in full or only in part.

Open the dependency map →

What the system demonstrates

Systems design. A seven-phase pipeline where every step reads specific files and writes specific files, so the whole thing is auditable end to end.
Multi-agent orchestration. Research, writing, and adversarial review run as separate specialist agents, with a cold handoff so quality is judged independently of who wrote it.
Machine-enforced quality. Output cannot ship unless it proves it followed the rules, on top of a deterministic validator for style, structure, and linking.
Self-healing lifecycle. The live CMS and sitemap continuously reconcile the plan, so tracked state can never quietly drift from reality.

I build AI-powered SEO content systems for early-stage B2B SaaS: full-scale SEO, semantic topical maps, content briefs, content writing, and end-to-end content automation. These diagrams are sanitized; client names and proprietary methodology are intentionally abstracted.

LinkedIn Book a 30-min call salehin.riad96@gmail.com

Both pages are single self-contained files. No build step, no dependencies, no tracking.