Investor Vinod Khosla has backed Wajo, a personal AI agent startup founded by former Meta, Google and DeepMind engineer Shivani Poddar. In Khosla's view, trust is the decisive factor in this market — because agents act on behalf of users. The company's flagship agent, Fo, works in the iMessage and WhatsApp messengers as well as through a web interface: it calls businesses and people on the user's behalf, books or cancels services, and creates virtual credit cards to pay without exposing payment details. TechCrunch reported the news on October 6, 2026.
Trust Is the Decisive Factor in the Agent Market
Khosla argues that the winner in the agent market will be decided not by technical superiority, but by trust. An agent calls on your behalf, books services and makes payments — such powers demand enormous trust. Khosla said this is exactly the criterion he used when choosing Poddar's team.
According to a Bloomberg report from October 8, 2026, startups like Wajo are entering the personal AI agent space that Google left open. Their target is to compete head-on with Meta's Muse agent in the consumer market. Competition between tech giants and new startups around personal AI agents is thus intensifying.
In practice, this trust is built through several mechanisms: virtual cards that hide payment details, a "trusted contacts" list that limits the agent's calls, and permissions that can be revoked at any time. These measures are aimed at keeping the agent's actions on the user's behalf under control.
What the Fo Agent Does
Fo is different from ordinary chatbots: it doesn't just reply — it moves to real actions to complete the user's tasks. The agent works via iMessage and WhatsApp, as well as in a web interface — the user gives instructions as a plain message, and the agent handles the rest. Unlike traditional chatbots, such agents interact with external systems — they make calls and process payments.
First, Fo calls businesses and people, booking services or canceling existing reservations. That means the agent talks directly with humans over the phone.
Second, for payments it creates virtual credit cards: payment goes through this card, but the user's real payment details are never visible to websites. As a result, users can pay without giving their card details to sites.
Third, when automation can't complete a task, Fo hands it over to human assistants. In other words, complex or ambiguous tasks are passed to a human operator.
Used in 107 Countries
According to Wajo, Fo is already used in 107 countries — a sign of the agent's global reach. The company says it has grown 10x since launching the agent at the end of September 2026. This figure was recorded in the short period after launch: the TechCrunch article was published on October 6.
In addition to Khosla, the startup has been backed by Jeff Dean and Gokul Rajaram. The total amount of funding raised has not been disclosed — the company has not announced the figure.
Patent and Call Controls
Wajo has filed a patent application for a virtual-card system that hides payment details from websites. The application covers specifically the mechanism for protecting user data in agent payments — this system is the technical foundation of Fo's virtual card feature.
The company also plans to introduce a "trusted contacts" list. This list limits whom the agent may call: Fo will only call people on the list. At the same time, the user can revoke the agent's calling permission at any time — control over calls remains fully in the user's hands. Wajo sees these measures as part of the product's architecture: trust is not a layer added later, but a factor considered from the very start of the design.
Poddar's Experience
Poddar is a specialist who worked as an engineer at Meta, Google and DeepMind; she led the Gemini launch. Poddar told TechCrunch that this experience taught her the need to combine deep technology with strong product design. According to her, Wajo is building trust into its architecture and product choices from day one. Her approach draws on her experience at large tech companies: she worked as an engineer at Meta, Google and DeepMind and led the Gemini launch.
In her conversation with TechCrunch, Poddar said leading the Gemini launch taught her the need to combine deep technology with strong product design, and that Wajo is building trust into its architectural and product choices from day one.
