Wajo pushes personal AI agent Fo, with human assistants stepping in when AI falls short

Wajo pushes personal AI agent Fo, with human assistants stepping in when AI falls short

N
News Editor
2026-09-29 03:39:40
Wajo, founded by former Google Jules engineering lead Shivani Poddar, is promoting a personal AI agent called Fo that can place phone calls, send emails, book services, and make payments on a user’s behalf. The product is designed for tasks that often break fully automated systems, especially when dealing with businesses that do not offer APIs. In those cases, Fo can hand the job to a human assistant, who then continues the task by calling the merchant, confirming details, and completing the request. According to the company, the human assistant only receives the information needed to finish the task and does not get access to the user’s full chat history. Wajo also shared internal test results. In a self-test covering 104 real-world tasks, with each task repeated three times and no human fallback used during the evaluation, Fo posted a 71% task completion rate, compared with 42% for OpenClaw. That puts Fo roughly 69% higher on a relative basis. In a separate safety test focused on whether the AI would contact others without permission, leak private information, or make unauthorized decisions, Fo recorded a 94% pass rate.

Wajo, founded by former Google Jules engineering lead Shivani Poddar, is promoting a personal AI agent called Fo. The product can place phone calls, send emails, book services, and make payments for users.

When Fo cannot complete a task on its own, it hands the work to a human assistant. That applies to cases such as contacting barbershops, personal trainers, or local merchants that do not have APIs. If Fo cannot get the job done by itself, a human assistant steps in to call the business, communicate the user’s request, and finish the task.

Wajo said the human assistant only receives the information required to complete the task and does not see the user’s full chat history.

Internal testing results

Wajo said it ran a self-test using 104 real-world tasks, with each task repeated three times. No human fallback was used in the test. Fo recorded a 71% task completion rate, versus 42% for OpenClaw, making Fo about 69% higher on a relative basis.

In a separate safety test, the company checked whether the AI would contact other people without authorization, expose private information, or make decisions it was not allowed to make. Fo posted a 94% pass rate.

This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan.
200

Disclaimer:

The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.

Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.