Secure, Cost-Controlled GenAI at Group Scale

Deploying generative AI across a whole group raises three questions at once: security, adoption and cost.

A global leader in glass packaging, operating in 12 countries, set out to deploy an enterprise generative AI platform across the group.

The group needed to:

• Deploy generative AI securely, with enterprise-grade identity and access controls.
• Drive adoption department by department, based on real use cases.
• Keep spending under control from day one.

Sia supported the deployment through a phased pilot (technical setup, activation, industrialization) followed by hypercare:

• Zero-trust security by design: single sign-on, user provisioning, role-based access control, audit logging and controlled execution of agentic and coding tools.
• Use-case-driven adoption: adoption per department, powered by business champions and reusable libraries of skills and code.
• FinOps from day one: spend controls, alert thresholds and cost versus value dashboards.
• Pilot scope: 150 pilot users: 125 business users and 25 developers.

Key results:

• Cybersecurity foundations validated: target architecture, connector scope and cyber action plan.
• Enterprise tenant live and domain validated, with the configuration of single sign-on and per-country access groups started.
• Lightweight governance in place: monthly steering committee, weekly operational points, FinOps and security reviews.

The result was a secured, cost-controlled foundation that enabled the group to scale generative AI, with autonomous teams, reusable assets and a group-wide roll-out roadmap.

By combining security, adoption and cost discipline from the start, Sia helps groups turn generative AI into a scalable capability.