Wajo is positioning trust and privacy as key differentiators for its Fo personal AI agent as competition grows from products such as Meta’s Muse, OpenAI’s Dot, and Instinct. The startup says Fo is now being used in 107 countries and has grown tenfold since its launch in late September, though it has not disclosed user numbers.
Fo can work across iMessage and WhatsApp and also offers a web interface for tasks such as scheduling, gifting, and life administration. Wajo says the agent can call businesses, send emails, make purchases, coordinate with people, and bring in a human assistant when a task cannot be completed automatically.
Wajo Builds Fo Around Privacy Controls
Wajo founder Shivani Poddar said trust needs to be considered from the beginning because agents can gain access to sensitive personal information while acting for users. The company has published its own research on trust and task completion, examining how agents can fail either by taking actions users did not approve or by abandoning difficult tasks.
Fo uses one-time virtual payment cards so merchants do not receive a user’s real card details. Wajo also says Fo sends messages from its own email address, does not expose stored passwords when signing into websites, and does not use customer data to train third-party AI models.
The company has applied for a patent covering its virtual card system. It also plans a trusted-contact feature for phone calls, while people receiving calls can tell Fo not to contact them again and have that preference recorded.
Fo Can Hand Tasks To Humans
Wajo introduced Fo as an agent designed to complete real-world tasks rather than only answer questions. The company says early uses have included booking appointments, ordering food, purchasing gifts, checking flights, comparing hotels, and arranging home services.
TechCrunch testing found that Fo successfully called an airline to ask about upgrades, although another test involving a personal phone call exposed shortcomings around call handling and disclosure. Wajo and other agent developers are still refining how AI systems should identify themselves when calling people on a user’s behalf.
Khosla Backs Wajo’s Trust-First Approach
Investor Vinod Khosla told TechCrunch that he backed Wajo because he considers trust especially important for agents that act on users’ behalf. Wajo has also secured backing from Google DeepMind chief scientist Jeff Dean and former Google and Meta executive Gokul Rajaram, although the company has not disclosed how much it has raised.
Poddar previously worked at Meta, Google, and DeepMind, including on conversational and multimodal AI systems. Wajo says it plans to continue developing Fo for social and multi-user situations while trying to preserve privacy between participants.
Featured image credits: Wajo
For more stories like it, click the +Follow button at the top of this page to follow us.