Recent comments from Carl Pei, founder and CEO of Nothing, suggesting that “apps will disappear” have reignited industry debate around the future of smartphone interfaces. The idea of replacing today’s app-centric experience with AI-driven, intent-based interactions is gaining traction across the ecosystem, as vendors explore what a truly agentic UI could look like.
From Smart Analytics Global (SAG)’s perspective, this vision represents a clear long-term direction for the industry. However, the transition from apps to AI agents will be significantly more gradual than current narratives suggest. While technological capabilities are advancing rapidly, consumer trust, privacy expectations, and ecosystem readiness remain critical constraints.
Early implementations already highlight this gap. On one end, aggressive approaches such as the Doubao-powered ZTE Nubia M153 attempt to bypass traditional app workflows entirely. On the other, more measured strategies—such as Samsung’s incremental integration of AI-driven app automation—reflect a more pragmatic path aligned with current user expectations.
As a result, SAG believes the industry is entering a multi-phase transition, where apps will not disappear overnight, but instead become increasingly invisible as AI layers evolve on top of existing ecosystems.
For nearly two decades, the smartphone experience has remained largely unchanged.
Agentic UI aims to replace this with intent-based interaction, where AI orchestrates actions across multiple services without requiring users to manually switch between apps.
However, it is a fundamental ecosystem shift, requiring deep integration across apps and services, new developer framework and APIs, and a redefinition of user control and trust.
Case Study 1: Doubao Phone (ZTE Nubia M153): Bold but Aggressive
At MWC 2026, SAG had hands-on experience with the Doubao-powered ZTE Nubia M153, one of the most aggressive attempts at AI-native smartphones to date.
The device demonstrated four core AI-native scenarios: dining, vacation planning, sharing and chatting.
These features aim to bypass traditional apps and enable AI to directly complete user tasks.
However, the real-world experience highlighted several challenges:
- Limited overseas compatibility:
Many services are tightly integrated with China-based ecosystems, leading to fragmented experiences in international markets - Ecosystem resistance:
Some apps in China have restricted or banned integration with Doubao due to security and data concerns - User trust gap:
The level of automation can feel intrusive, raising concerns about control and unintended actions
From SAG’s perspective, Doubao represents an important directional experiment. But its approach may be too aggressive for current market readiness.
Case Study 2: Samsung Galaxy S26 — A Measured Approach
In contrast, Samsung Galaxy S26 takes a more incremental and pragmatic path toward agentic experiences.
One example is automated app actions (starting with Uber):users can initiate a ride via voice by stating a destination. The system launches Uber and connects to drivers automatically. Final payment still requires user confirmation, preserving user control
This design reflects a key principle: assist, don’t override
From SAG’s perspective, this is an encouraging early step. Samsung is testing agentic workflows within controlled boundaries, maintaining user trust through confirmation layers, and gradually expanding ecosystem integration.
Platform Shift: Google and Android Are Quietly Enabling Agentic UI
Beyond OEM-level experimentation, the transition toward agentic UI is also being shaped at the platform level — most notably by Google and the broader Android ecosystem.
Google’s AI strategy signals a clear move toward system-level intelligence, where AI is no longer confined to individual apps but acts as a coordination layer across services.
Key developments include deeper system integration of AI assistants (e.g., Gemini), cross-app task orchestration, and developer ecosystem enablement. From SAG’s perspective, this platform-layer evolution is critical. Unlike device-specific implementations, OS-level integration provides the scalability needed for agentic UI to reach mass adoption.
However, the same constraints apply due to the fragmentation across Android vendors and regions, inconsistent app-level support and APIs. Ongoing concerns around data privacy and permission control.
As a result, even with Google pushing the ecosystem forward, the transition will remain gradual and uneven across markets and devices.
The Real Bottleneck: Trust, Not Technology
While AI capabilities are advancing rapidly, consumer trust remains the primary barrier to adoption. The vision of an AI that can anticipate needs, make decisions, and act across services is compelling, but it also challenges deeply rooted expectations around control and autonomy. Many users are not yet comfortable delegating high-value or sensitive actions to AI, particularly when the decision-making process is not fully visible or easily reversible.
At the same time, there is a fine balance between assistance and overreach. If AI begins to act ahead of user intent, it risks creating friction. For agentic UI to scale, users must feel confident that they remain in control, with clear boundaries, transparency, and the ability to intervene at any point. Until that trust framework is firmly established, adoption will lag behind technological progress, indicating a gradual rather than immediate transition.