Agentic Content Ops — AI Workflow System for miran.at

About the Project

This project defines and operationalizes a structured content workflow for miran.at, focused on reducing manual repetition in multilingual publishing and improving consistency across business-critical pages.

Instead of treating every page as a one-off copy task, the system uses shared data models, translation keys, and clear boundaries between editorial content and code-driven surfaces.

Problem

Before hardening the workflow, core positioning and service messaging drifted between pages, languages, and sections. AI-related claims were present, but proof and page hierarchy were uneven.

Approach

I implemented an AI-first content operations model with three layers:

1. Canonical identity layer for role, offer, and positioning statements
2. Structured profile/service data layer for reuse across routes and components
3. Localized editorial layer for language-specific pages, posts, and landing pages

Measurable Outcomes

• 3 active locales aligned in the same content model: DE, EN, HR
• 76 editorial content files managed in structured collections (blog, projects, landing)
• 6 service pillars mapped to dedicated localized service routes
• 186 tools catalog entries retained as code-driven surfaces outside CMS boundary

Why It Matters

The system creates a scalable foundation for AI-assisted publishing and faster iteration. New content can be added without breaking message consistency across profile, services, and SEO pages.

Next Iteration

The next step is adding outcome-led AI case studies tied to client workflows (support, internal knowledge, and content operations) with explicit baseline and post-launch metrics.