How Singapore and ASEAN Businesses Can Move Agentic AI Beyond Pilots

For Salesforce ASEAN senior vice-president and general manager Paul Carvouni, the challenge is not simply getting businesses to try agentic AI. It is helping employees use it effectively in ordinary work while putting the safeguards in place for wider adoption.118

Build around the work, not the pilot

Carvouni warns that organisations can remain stuck in small-scale trials if they do not rethink jobs, employee support and technology guardrails.1 Salesforce research reported on workers who moved from initial pilots to deep daily use points to three supporting conditions: role-specific training, AI embedded in existing workflows and strong data security.18

That suggests a practical shift in focus: train people for the tasks they perform, decide where AI fits into those tasks and assess whether it remains useful in day-to-day operations. Giving staff access to a generic tool alone does not address the workflow and trust issues Carvouni identifies.118

Make responsible deployment part of adoption

Security and governance should accompany wider use rather than follow it. Singapore’s guidance describes its Model AI Governance Framework for Agentic AI as practical guidance for deploying agents responsibly and safely, with humans ultimately accountable.4 For businesses, the implication is to design human oversight and data safeguards into the workflow as it expands.14

Put Singapore’s strategy in context

The question’s reference to a “National AI Strategy 2.2” is not confirmed by the official materials provided. Singapore’s Digital Gateway calls the strategy National AI Strategy 2.0, while the Smart Nation site describes a May 2026 update setting out 10 refreshed priorities.34 The relevant business lesson is consistent with Carvouni’s advice: national ambition becomes meaningful at work when employees have the skills, workflows and trust needed to use AI responsibly beyond a trial.1318