Analyst(s): Olivier Blanchard
Publication Date: October 27, 2025

AI PCs have, thus far, largely failed to deliver the proverbial “killer app” that would bring on-device AI capabilities to the forefront of the category’s most underrealized productivity-boosting value proposition. But perhaps the answer has been hiding in plain sight all along: adding a new layer of verbal, context-aware, natural language, conversational agentic user interfaces, similar to Amazon devices’ Alexa+, to the PC experience.

Key Points:

• The shift to frictionless interfaces (hands-free, voice-enabled, natural language) should be central to every AI-enabled PC vendor’s UX strategy for both consumer and commercial segments.
• As users become more comfortable with voice-first interfaces and agentic solutions proliferate, delivering frictionless, reliable, and remarkable voice-first user experiences will be crucial.
• For mobile handset vendors, the transition will be relatively smooth, as platforms already incorporate wake words and assistant-enabled prompts. However, the increased compute and inference requirements for verbal interactions will necessitate pushing SOC and battery performance beyond current trajectories.
• PC vendors specifically need to prioritize on-device capabilities for voice-enabled use cases, including improved microphone and speaker quality, and enhanced compatibility with Bluetooth peripherals and accessories.

Overview:

A significant trend in AI assistant interfaces is the shift from keyboard and screen-based interactions to verbal, hands-free interfaces. While many popular AI assistants already offer verbal interfaces, the default UI design remains rooted in typed interactions. This is expected to change within the next 24 months.

Amazon’s Alexa+ ecosystem, which includes the Alexa+ assistant platform, Alexa-enabled edge devices (Fire TVs, Echo speakers, Ring cameras), and Cloud services, has pioneered the voice-first approach. This approach minimizes friction points, making interactions with assistants more natural, organic, and ambient, offering insights into why agentic interfaces will gravitate toward voice-based interactions. The momentum behind smart glasses, equipped with microphones, speakers, and cameras for contextual awareness, further supports this shift toward hands-free, voice-based, natural language interactions.

As AI assistants and agentic capabilities in natural language processing improve, and hardware leverages these capabilities across various segments (smart home, smart glasses, headphones, software-defined vehicles), users will expect similar functionality in their mobile devices, tablets, and PCs.

Amazon’s Alexa+, the latest iteration of its generative AI-powered assistant, emphasizes voice and presence over text. This represents a different philosophy for AI assistants: while many focus on work, creation, or information retrieval on computers and phones, Alexa+ focuses on more personal, simple, and hands-free tasks, particularly within the home environment.

The report clarifies that this is not an “either/or” situation. Different agentic tasks and use cases may require different types of interfaces, suggesting a multimodal approach. Amazon’s belief in voice-driven, context-aware interaction works well in the home, where users are not always at a desk. Agentic solutions need to function seamlessly in various rooms and during different activities, requiring an ambient, hands-free interface.

Alexa+ now enables multi-step tasks using natural speech, interpreting requests that might seem ambiguous to productivity-focused AI assistants. Combining on-device processing with cloud AI, this intelligence aims to make Alexa+ feel like a built-in, ambient home utility. While this approach may not be as suitable for business environments with PC-constrained workflows, keyboard-based and voice-based agentic interactions can complement each other.

For a PC user, multimodal capabilities could involve keyboard-based agents for tasks such as email management, report drafting, and research. In contrast, a voice-based agent could find and open documents, perform quick searches, summarize results, and offer links without interrupting the user’s flow. This document retrieval capability provides significant time-saving and productivity benefits by allowing users to verbally command an agent to find elusive documents while focusing on critical, focus-intensive tasks.

This type of agentic workflow, primarily enabled by verbal commands, works best when handled locally on the device due to lower latency and orchestration efficiency, and the ability to free up network bandwidth for more compute-intensive tasks, reducing dependence on networks and cloud-based applications for basic, everyday, agentic tasks.

The report also considers scenarios where PC users monopolize GPU and CPU resources with heavy workloads. In such cases, verbally prompting an NPU-powered agent on the device to manage meetings, schedule calls, reply to emails, reorganize documents, summarize discussions, or create PDFs highlights the value of current AI PC SOC architecture.

In essence, building a multimodal agentic architecture directly into PC hardware is a “cheat code” to maximize the utility of a multi-agent environment. This could unlock new value for NPUs, which have been underutilized in the AI PC ecosystem, and potentially provide the “killer app” that the AI PC segment has been seeking. Peripherals such as microphones, speakers, and hearables can further enhance verbal agentic performance, with PC vendors either incorporating voice-enabled utility into their own peripheral catalogs or improving third-party device compatibility.

What to Watch:

PC vendors beginning to leverage multimodal agentic capabilities, including contextually-aware, conversational, voice-enabled interactions with agents and AI assistants, as (1) a fresh productivity-centric value proposition for the AI PC segment (which still struggles to deliver an on-device AI-powered “killer app”), (2) a competitive differentiator against PC vendors who are slow to seize on the opportunity, and (3) a clear value-add (and possible demand driver) for AI PCs in higher price bands. The latter may be especially valuable to PC vendors who remain frustrated with how much AI PC adoption thus far, especially in the commercial segment, has prioritized lower price and lower system performance bands instead of premium, cutting-edge AI PC systems.

The full report is available via subscription to the Intelligent Devices IQ service from Futurum Intelligence—click here for inquiry and access.

Futurum clients can read more in the Futurum Intelligence Platform, and nonclients can learn more here: AI Devices Intelligence Practice.

About the Futurum AI Devices Practice

The Futurum AI Devices Practice provides actionable, objective insights for market leaders and their teams so they can respond to emerging opportunities and innovate. Public access to our coverage can be seen here. Follow news and updates from the Futurum Practice on LinkedIn and X. Visit the Futurum Newsroom for more information and insights.