Operationalising AI – A Playbook for Value, Governance & Control
Sofia Silvestre, AI Delivery Manager, Galp
In this structured and strategic session, Sofia Silvestre, AI Delivery Manager at Galp, presented a practical playbook for turning AI from a promising technology into a reliable, value-generating capability within the enterprise. Recognising that successful AI implementation doesn’t stop at experimentation, Sofia Silvestre outlined how organisations must operationalise AI through deliberate focus on business impact, governance, and control.
Tracing the evolution from machine learning to generative AI, she explained why AI initiatives must be grounded in strong governance frameworks that prioritise transparency, trust, and compliance—especially as AI capabilities scale and regulatory expectations grow. She walked the audience through Galp’s approach to responsible AI development, including how to anticipate and manage critical risks such as data drift, model drift, and business misalignment.
Sofia Silvestre’s playbook emphasised three core pillars:
- Value: Define business outcomes early and continuously measure impact throughout the AI lifecycle.
- Governance: Build trust through transparency, ethical use of data, and adherence to evolving compliance standards.
- Control: Maintain reliability and accountability by monitoring AI systems for performance, change, and security threats.
The presentation also highlighted the importance of aligning AI initiatives with strategic organisational goals and ensuring that development processes are not only technically sound but fully integrated into operational and IT ecosystems.
Ultimately, Sofia Silvestre made the case that AI should not be a siloed innovation project—it must be treated as a strategic capability with built-in controls, measurable value, and clear ownership. Her session provided a real-world guide for organisations seeking to go beyond hype and embed AI responsibly and effectively into core business operations.
