Most retail enterprises in the region are no longer short of AI ambition, or AI pilots. What they lack is AI they can trust to touch live operations: to reprice, reorder, resolve a customer, or release stock in the moment, across fragmented markets and channels, without eroding margin or customer trust.

That is the real threshold between an impressive demo and AI that runs the business, and it is far less about models than about foundations. It comes down to data that is trusted, unified and fast enough for an agent to act on with confidence, and an operating model redesigned around agents rather than one that simply bolts them onto old workflows.

This invitation-only roundtable brought together senior retail and technology leaders from across Singapore and Southeast Asia to discuss what it actually takes to cross that line.

The conversation moved past the pilot-versus-production to the harder questions: where agentic systems are already earning their place in retail operations, what real-time data foundation makes that possible, how accountability and control work when software starts making operational decisions, and how the retail operating model itself changes when humans and agents share the work.

The goal was a candid exchange of what is working, what has broken, and where peers are placing their next bets.