Agentic AI adoption continues to climb, but a closer look at where it is landing tells a more complicated story. Infobip's 2026 Customer Experience Maturity Report finds that 53 per cent of enterprises globally, and 44 per cent in Europe, have already deployed agentic AI. On the surface, it looks like another adoption milestone, but underneath the report shows that AI is being deployed for the easiest parts of the customer journey, while the complex, high-value interactions that drive revenue and retention remain largely manual.
Adoption Is Concentrated in Low-Value Tasks
Infobip's findings expose a significant divide between where agentic AI has taken hold and where it would deliver the most value. Feedback collection has the highest adoption at 56 per cent, followed by reminders and notifications at 52 per cent and authentication at 45 per cent. These are simple, low-friction tasks.
Higher-value journeys tell the opposite story, however. Loyalty management sits at 35 per cent adoption, delivery management at 28 per cent, customer onboarding at 26 per cent, and product returns and refunds at just 15 per cent. These are the interactions with the greatest potential to affect customer loyalty, and the ones still leaning hardest on human effort.
Fragmented Systems Are the Real Bottleneck
For years, the conversation centred on what AI could do. Increasingly, the limiting factor is what the surrounding systems will allow it to do.
Infobip's report points to fragmented data and disconnected systems as the root cause. Just 27 per cent of brands currently use a communications orchestration platform, and only 58 per cent say their channels are fully synchronised. Without centralised, real-time data and the ability to act consistently across channels such as WhatsApp, SMS and email, AI agents lack the access and context needed to resolve issues end-to-end. The fight for control of the agent layer is becoming as important to enterprise AI strategy as the agents themselves, with vendors such as Kore.ai now positioning orchestration and governance as the differentiator rather than the model.
Why This Matters for CX Leaders
Ante Pamuković, Chief Revenue Officer at Infobip, said the report marks “a turning point for global brands” this year, with CX maturity set to separate brands that scale agentic AI successfully from those that struggle. He added that moving from basic automated responses to deep, seamless journeys requires brands to overcome the barriers created by fragmented systems.
The findings echo a wider pattern emerging throughout 2026. Research by Sinch has found that 73 per cent of organisations have rolled back AI customer agents after running into governance problems once deployments moved beyond pilots. AI governance in customer experience has faced issues including fragmented data, unclear ownership and processes built for human handoffs rather than autonomous execution.
Deploying an AI agent is not necessarily the same as transforming a customer journey. As organisations move past experimentation, attention is shifting towards orchestration, governance, integration and data accessibility. These are the capabilities that determine whether AI delivers measurable outcomes or simply becomes another layer on top of existing processes.