Wajo wants its AI agent to handle errands, calls and payments

Wajo is building an AI agent that can contact people and businesses, arrange services and pay on a user’s behalf. The startup’s pitch is that trust should be built into the product from the start, which matters when an assistant is being given access to your inbox, contacts and wallet.

Watch Desk analysis

What happened

TechCrunch reports that Wajo’s Fo agent works across iMessage and WhatsApp and offers a web interface for tasks such as scheduling, gifting and life admin. It can call businesses or other people, book or cancel services, create a virtual credit card and use it to pay without sharing the user’s card details. The company says the agent can also hire a person to complete a task when needed.

Wajo founder Shivani Poddar told TechCrunch that the company is designing around trust and safety. Wajo says Fo has an email account of its own and can start a group with a user and a business when a service blocks AI agents. The company also says it is working on a trusted-contact list for calls, and that recipients can tell the agent not to call again.

TechCrunch describes early tests in which Fo successfully called an airline about upgrades, but a call to remind the writer’s partner to feed the cats did not go smoothly, with unclear disclosures. Wajo said its agent is used in 107 countries and has grown tenfold since launching in late September. Those are company-reported figures; the article does not give a user count.

Why it matters

This is a more consequential proposition than an assistant that drafts a shopping list. Fo is meant to act in the world: make calls, communicate by email and spend money. That could save time, but it also makes consent, disclosure and the ability to revoke access central product features, not fine print for a quiet afternoon.

Wajo’s trust-first positioning is a claim, not proof that the agent will behave reliably. TechCrunch’s reported test results offer a useful reminder that a slick task list and a dependable representative are not quite the same thing.

Our read

The interesting question is not whether an AI can place a call. It is whether the person on the other end knows what is happening, and whether the user can keep the agent’s authority tightly bounded. Wajo is addressing some of the right problems, but readers should treat its privacy and growth claims as the company’s account, not an independent audit. If you try an agent like this, start with low-stakes tasks and permissions you can easily withdraw.

What to watch

• Whether Fo clearly identifies itself when calling people or businesses.
• How users approve payments and limit the agent’s access to personal information.
• Whether Wajo publishes independent evidence for its privacy and safety claims.
• How its human-task service works when an agent cannot complete a request.

Discussion spark: Would you let an AI agent make calls and payments for you if you could set limits, or should those tasks always require your direct involvement?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.