Adobe, Salesforce, Meta, and Vercel have collectively signalled where enterprise CX infrastructure is heading, and the trajectory reveals both opportunity and significant risk. Adobe's Summit announcements centred on agentic AI orchestration—specifically, how to maintain brand safety and customer accuracy when AI agents operate autonomously on behalf of customers—whilst Salesforce's leadership doubled down on the thesis that AI strengthens rather than threatens SaaS adoption, directly countering investor anxiety about a "SaaS-pocalypse." These moves reflect confidence that AI will deepen dependency on enterprise platforms rather than commoditise them. Yet this optimism sits uncomfortably alongside Meta's revelation that it is harvesting employee keystrokes and mouse movements to train its AI agents, exposing a critical blind spot in how these systems are being built: the data collection methods underpinning agent training remain largely opaque and potentially problematic.
The Vercel breach crystallises the operational risk this opacity creates. An employee's use of Context.ai, a third-party agentic AI tool, became the entry point for attackers to compromise customer data—a scenario that should alarm any CX leader managing sensitive customer information. This is not merely a security incident; it exposes how "shadow AI" tools can bypass organisational controls and create vulnerabilities that traditional security frameworks were not designed to address. For teams already running Agentforce or similar agentic platforms, the question is no longer whether AI agents improve efficiency, but whether your organisation has visibility into every AI tool your team uses and whether your security posture accounts for the expanded attack surface these tools introduce.
The week's announcements collectively suggest that agentic AI in CX is moving from pilot to production at scale, but the infrastructure—both technical and governance—to manage this transition safely is lagging. Adobe's focus on brand visibility and accuracy is necessary but insufficient if employees can introduce unvetted AI tools into their workflows. CX leaders must now treat AI governance as a core operational priority, not an afterthought, and demand transparency from vendors about how their agents are trained and what data they require. The competitive advantage will belong not to those who deploy agents fastest, but to those who deploy them most securely.
From major announcements at Adobe Summit to Meta’s plans to start tracking its employees’ keystrokes, here are extracts from some of this week’s most popular news stories. Five Adobe Summit Announcements Every CX Leader Should See Adobe dropped a raft of major product announcements at its annu