Firms weigh agent autonomy and execution limits in production
An InfoQ China article promoting a live-streamed discussion brings together frontline practitioners from Fliggy, Huolala and China Mobile Jiangsu to talk about enterprise agent autonomy, execution boundaries and production-grade deployment. Its framing question is that once agents start placing orders, scheduling and changing systems, the issue is no longer whether it can be done but whether it should be. The piece also promotes QCon Shanghai on October 22–24, whose tracks cover AI-native architecture, agent runtime, AI infrastructure, agent security and observability, loop engineering, embodied intelligence and world models.
It signals that enterprises now put governance and permission boundaries ahead of raw agent capability once agents touch production workflows.