AI is becoming a new competition for smartphone manufacturers. TechCrunch's test of the British luxury mobile phone brand Vertu's latest folding screen, Alphafold, shows that the 6880-dollar device does not focus on conventional hardware parameters, but rather focuses on AI smartware services for executive users.
Hermes, the core.
Vertu uses pre-packaged Hermes Agent as a core selling point. According to the company, this intelligent body can analyse documents, cross-apply missions, remember the content of the dialogue and, if necessary, transfer them to manual protocol services for processing. Unlike most of the question-and-answer cell phone assistants, Hermes is located closer to a working agent who can perform multi-step tasks.
Hardware follows a luxury route
From the looks of it, Alphaford has continued Vertu's consistent high-end route. The calibers are decorated with calves and titanium metal, and are also packed closer to luxury gift boxes rather than ordinary mobile phone boxes. Vertu is trying to sell not just the phone itself, but a set of high-end use experiences.
However, TechCrunch found that there were significant similarities between the mobile phone and the Nubia folding screen product, including chain design, body size, and the location of speakers, microphones and fingerprint identifiers. Also in the system information are signs associated with the build-up.
In response, Vertu stated that Alphafold had been developed in cooperation with the supply chain associated with the ZTE/Nubia hardware platform, where hardware platforms, component integration and production works were provided; Vertu was responsible for luxury materials, software experience, quality control and after-sales services.
Document analysis has advantages
In a few days of testing, TechCrunch uses Alphaford as its primary mobile phone and focuses on document processing, table analysis, travel arrangements and cross-application automation, rather than taking photographs or recreational experiences.
The test results show that Hermes performed well in the analysis of local documents and forms. It was mentioned that it had correctly summarized data for the second quarter of a sales form. At this point, Gemini, which is running on the Samsung Galaxy ZFold 7, still needs to upload the document manually.
According to the article, Hermes is more proactive in the implementation of multi-step workflows and more willing to proceed directly across applications. This makes it more like an “agent” on some missions than just a chat assistant.
Multiple error in process and alarm
In a mock-up of the airport, the tester asked Hermes to send a message to the contact that "will be 20 minutes late", open the navigation to the airport, open the no-disturbing mode and alert the hotel in 15 minutes. Hermes completes the message, opens no disturbance and opens the map, but does not automatically start the navigation and sets the alarm time error at 9:08 p.m. that night instead of 15 minutes later.
After the same request was sent to Gemini, the latter asked about the airport destination and which application the reminder should be created. While more steps have been taken, they have resulted in a reminder of the right time.
In another business trip test, Hermes was asked to arrange travel from Mumbai to Pune, including early flight, hotel advice and calendar additions. It indicated that it did not have a suitable direct early flight and provided a button to transfer the need to the Vertu protocol service, while at the same time setting the calendar date wrongly at 7 July instead of 18-19 July as requested by the user.
Besides, Hermes' memory of historical documents is unstable. It was mentioned that the sales forms, which had been uploaded and analysed a few days earlier, could not be re-identified in the same follow-up dialogue and that the system had instead requested that the documents be re-published.
Overall, this high-priced AI mobile phone displays some capability in document analysis and cross-application automation, but there are still significant errors in high-frequency workspaces such as time, date and context continuity. The real question for the target user is not whether it is luxuriant enough, but whether this type of AI intelligence is able to perform its daily tasks with stability.