The industry consensus around agentic AI in customer experience has shifted from pursuing maximum autonomy to designing systems that augment human agents rather than replace them. SYKES's positioning reflects a broader recognition that the goal isn't to remove humans from the loop but to create AI that handles routine tasks whilst escalating complex issues intelligently. This reframing matters because it directly challenges the automation-at-all-costs mentality that dominated vendor messaging over the past two years. For teams already running Agentforce or similar agentic platforms, this means the ROI calculation changes: success is no longer measured by deflection rates alone, but by whether AI is genuinely reducing cognitive load on your best agents and improving first-contact resolution on nuanced cases. The related consolidation activity—such as Automation Anywhere acquiring Boost.ai—suggests vendors are racing to build this human-centric orchestration layer, recognising that pure autonomy creates liability and customer dissatisfaction.
What's emerging is a tension between vendor capability and customer expectation. Whilst platforms can now handle increasingly complex interactions autonomously, brands like Hotel Bethlehem are explicitly rejecting AI agents as a differentiator, signalling that autonomy itself has become a liability in certain segments. This creates a strategic question for CX leaders: should your team be optimising for agent efficiency through AI augmentation, or using human-first service as a competitive moat? The answer likely depends on your customer base and margin structure, but the shift away from "autonomous" framing suggests that mid-market and enterprise teams should be stress-testing their current automation strategies against customer sentiment, not just cost savings. Vendors pushing pure autonomy without robust human handoff mechanisms are increasingly out of step with where the market is moving.
Agentic AI and customer experience: Why autonomy isn’t the goal SYKES