Vertuが幹部向けに68万円のAIエージェントを販売——実際の性能を検証Vertu wants executives to pay $6,880 for an AI agent — here’s how it actually performs

Vertu, the British brand best known for handcrafted luxury smartphones wrapped in titanium, sapphire crystal and exotic leather, is now selling an artificial intelligence agent aimed squarely at corporate executives. The subscription, priced at roughly $6,880 a year (about 680,000 yen), positions itself at the very top of a fast-crowding market for AI assistants. It raises a pointed question that TechCrunch set out to examine: does paying elite money for an AI agent actually buy meaningfully better output than far cheaper, or even free, alternatives?

The pitch centers on the idea of a digital concierge for time-pressed leaders. Marketed as an autonomous assistant capable of handling scheduling, correspondence, research and other routine knowledge work, Vertu's offering follows the broader industry shift from chatbots that merely answer questions toward "agents" that can take multi-step actions on a user's behalf. In practice, that usually means an AI system connected to calendars, email, documents and web services, given the ability to plan a task, carry out several steps, and return a finished result rather than a single reply.

The crucial technical detail is what sits underneath. Most commercial agents today are built on a small number of foundation models from providers such as OpenAI, Anthropic and Google, layered with orchestration software, memory and tool integrations. Unless Vertu has trained or fine-tuned its own model at significant expense, its agent is likely drawing on the same class of underlying intelligence available to services costing a fraction of the price. That makes the performance gap, if any, dependent less on raw model capability and more on the quality of the surrounding workflow, the curation of tools, and the human support wrapped around the product.

That is where luxury branding meets a genuine engineering challenge. Enterprise-grade agents are difficult to make reliable, and the well-documented problems of the category apply regardless of price. Large language models can hallucinate, misread instructions, or take incorrect actions when granted access to real accounts and data. Executives handling sensitive commercial information also face heightened privacy, security and compliance considerations, which a premium vendor would need to address convincingly to justify the cost. Any assessment of whether the service is "worth it" therefore hinges on accuracy, safety guardrails, data handling and the degree of white-glove service, not on the badge alone.

Vertu's move fits a recognizable pattern for the company, which has long sold ordinary phone functionality at extraordinary prices by emphasizing exclusivity, materials and concierge services rather than specifications. Extending that formula to AI is a logical brand play at a moment when agentic products are among the most hyped categories in technology. It also mirrors a wider trend: enterprise software firms, consultancies and startups are all racing to package agents for high-value professional users, betting that businesses will pay a premium for tools that promise to reclaim expensive executive time.

The broader context is a market still searching for proof of value. OpenAI, Anthropic, Microsoft, Google and Salesforce have all pushed agent frameworks and assistants over the past two years, while open-source projects and developer tools such as LangChain and AutoGPT-style systems have made it possible for smaller players to assemble comparable capabilities. Independent benchmarks for agent reliability remain immature, and real-world performance often falls short of demonstrations. Against that backdrop, a five-figure annual subscription invites scrutiny over whether the premium reflects superior technology, superior service, or simply positioning for a clientele accustomed to paying more.

For prospective buyers, the practical takeaway appears to be that price is a poor proxy for AI quality. A luxury wrapper can add value through concierge support, tailored integrations, prioritized human oversight and status signaling, but it does not automatically confer smarter models or fewer errors. Anyone evaluating the Vertu agent, or any comparable high-end offering, would be well served by testing it against mainstream tools on their own tasks before committing.

Vertu's experiment will likely serve as an early data point on whether a distinct luxury tier can hold up in AI the way it has in hardware. If premium agents cannot demonstrate measurable advantages over accessible competitors, the segment may struggle to justify its pricing as the underlying technology continues to commoditize.