The Swiss news media industry faces challenges like never before. Decreasing revenues and readership from print, a saturated market, and AI disrupting how content is created, edited, and consumed, warrant the need for more governance. At Tamedia, various AI tools are already in use, in different capacities and stages. Yet, governance, feedback loops, and training regiments remain unsatisfactory. The aim of further implementing AI in workflows poses risks, financially, legally, and ethically. Also, resistance from editorial teams remains a challenge for upper management.

First, a literature review of academic literature was conducted. The findings were structured along the content lifecycle described by Caramiaux et al. (2019), namely creation, production, and consumption. These findings were then synthesized with company-internal documentation. The gaps between theory and practice were consolidated into a gap analysis, which then formed the basis of an adapted Balanced Scorecard, which is divided into five perspectives. This holistic framework in turn provided the basis for a Decision Tree, which translates strategy into decision-making.

The analysis revealed numerous systemic issues, the main ones being tool fragmentation, missing feedback loops, lack of role-specific AI trainings, insufficient AI governance, and uneven AI usage across the three phases. The Balanced Scorecard, being informed by the gap analysis, highlights these gaps in five perspectives, namely: Customer & Audience, Internal Workflows & AI Integration, Learning & Growth, Financial Contribution, and Editorial Compliance & Ethical AI. The Decision Tree aims to give management and project leads a step-by-step walkthrough when conceptualizing new AI content verticals in five steps: Audience & Strategic Fit, Workflow & Automation Readiness, Editorial Oversight & Capacity, Learning Infrastructure & Feedback Loops, and Risk Governance & Accountability. The viability check and proof of concept for the Balanced Scorecard were done retrospectively with a planned AI content vertical. Both frameworks were created specifically for the client, based on real, tangible, internal documentation. To apply these frameworks in general, both require reevaluation and must be adapted and tested for future industry trends and developments.