Wajo's Fo: The AI Agent That Hires Humans for the Calls
Fo is a personal AI agent from Wajo, a San Francisco startup founded by Shivani Poddar, who led engineering at Google DeepMind before this. It opened to the public on September 28, 2026, after a month of invite-only alpha. The homepage pitch is one sentence: Fo calls, emails and coordinates with people to get your everyday to-dos done, with human help when the task needs it.
That last clause is the whole product. Every other personal agent launched this year, Meta’s Muse, Instinct, OpenAI’s Dots, sells the idea that the model does everything. Poddar’s launch post says the opposite: “Other personal AI’s pretend AI can do everything. Fo employs humans to do tasks that AI cannot.” It passed 2.8 million views in two days.
What Fo actually does
The example tasks on Wajo’s site are errands, not work. Book a dentist for Thursday afternoon. Find a barber open on Sunday. Get a plumber out this week. Cancel the gym membership. Send Mom flowers on the 12th. Fight the parking ticket. Order lunch for the team and collect everyone’s order in Slack.
Fo reaches you by email, text, Slack, phone and a web app, and it connects to Gmail, Google Calendar and Google Drive. The loop is: understand the task, search, take an action (a call, an email, a checkout), report back with what it did and what it cost. When the action is a phone call to a business with no API, which describes most barbershops and plumbers, a human executive assistant on Wajo’s staff makes the call. Wajo says the assistant gets only the information needed for that one task, not your chat history.
Three design choices carry the trust argument:
- Single-use cards. Fo checks out with a fresh virtual card for each purchase. Your real card details are never shared with the merchant or seen by Fo.
- Its own email address. Fo emails from its own address and never sends as you.
- No stored passwords. Fo signs in to websites for you without seeing or storing your password.
The numbers Wajo is quoting
Wajo ran 104 real-world tasks and says Fo completed 71 percent of them against 42 percent for OpenClaw, which it rounds to “2x better” and “beats other agents by 69%.” On a separate safety test covering unauthorized contact, information leakage and unauthorized decisions, it reports a 94 percent pass rate, and the launch post claims Fo is four times less likely to leak private information than Muse or Instinct.
Those are Wajo’s tests on Wajo’s task set. The one independent read so far is from assistantbenchmark.com, which ran Fo through six of its fifteen categories in early September, before the public launch. Fo scored 6.2 out of 10 overall. It passed travel search, memory, restaurant recommendations and a weekday 7am briefing that fired on time both mornings. It failed the online booking task: 64 minutes and 24 messages, a non-refundable hotel rate described as cancellable, and no booking at the end. The email test scored a partial, because the reply went out from Fo’s own address rather than the user’s.
Both results can be true at once. Fo is strong at the errand loop the site advertises and weaker the moment a task needs the agent to be you.
Fo pricing
| Tier | Price | What you get |
|---|---|---|
| Fo Mini | Free | A daily task cap (the site lists 1,000 tasks a day, resetting daily) |
| OG Fo | Free, invite only | No daily cap, unlimited call time, group chats with Fo |
| Pro | Contact sales | Heavier workloads, MCP access, custom-built agents |
Wajo has not published a Pro price. The independent benchmark listed it at $99 a month behind a waitlist as of September 9. Funding and investors have not been disclosed.
What the terms say
Wajo’s privacy policy is more conventional than the one Instinct launched with, but three parts matter:
- Calls that Fo or its human assistants make on your behalf may be recorded and transcribed, and Wajo keeps the summaries and outcomes.
- Wajo says it does not use Google Workspace data to train generalized AI models and does not sell it. The wider policy does list developing, training and fine-tuning models, including personalizing the agent, as a use of personal information.
- Disputes go to binding individual arbitration under California law. There is a ten-day refund window on the subscription itself, not on anything Fo buys for you.
Where Fo fits
Fo is the best-argued version of the human-in-the-loop errand agent, and the design is honest about what models cannot do yet. If the job is the dentist, the plumber, the flowers and the hold music, it is worth a free account.
Work is a different job. A client reply has to come from your address, not the agent’s. A calendar change has to land on your calendar and your colleague’s. A lead that arrives at 2am has to trigger something without anyone typing a task. That is what Carly is built for: it works inside your Gmail or Outlook, sends as you, watches for events and runs workflows on them, and connects to thousands of apps through official connections. Free Zapier-style workflows; AI agents from $35/month. Get started here.
For the full field, see Wajo Fo alternatives, and for the head-to-head, Carly vs Wajo Fo.
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
