Personal AI Assistant Tab Emerges From Stealth at a $300 Million Valuation

Tab, a start-up building a personal AI assistant that people message much as they would a friend, has emerged from stealth at a valuation of $300 million, joining a fast-growing group of consumer AI products that promise to take everyday chores off their users' hands.

The company announced the valuation on Tuesday, 6 October, TechCrunch reported, but declined to say how much it has raised or on what terms. Its backers include SV Angel, Valar Ventures and American Spirit. Tab was founded by Brennan Erbz, Stafford Schlitt and Ammar Amdani, who began working on the product earlier this year.

That timeline is what makes the number stand out. A company whose founders started building only in 2026 is now valued at $300 million, an indication of how much investors will pay for teams that can turn large language models into agents that complete tasks rather than simply answer questions.

An assistant that lives in your messages

Tab does not ask users to download a new app. People text it through iMessage or WhatsApp with whatever they need, whether that is ordering groceries or finding a birthday present, and Tab carries out the task. According to the company's website, users can also send photos, screenshots, PDFs and videos, from which the assistant can extract details, edit media or create new images, audio and video.

The design choice is deliberate. Messaging threads are where many people already organise their lives, and placing an assistant inside them removes the effort of learning a new interface. Tab's website says the assistant asks clarifying questions, sends short progress updates, returns results to the same thread and keeps working if the user steps away, picking up where it left off.

"You text it like you'd text someone you trust and it takes the weight off," Amdani told TechCrunch. "The bills, the bookings, the calls you keep putting off, remembering the people you love."

Behind the chat window, Tab is set up to act on its own. Amdani said the assistant has its own phone number, computer and wallet, so rather than explaining how to do something, it does the job itself. The website describes it running its own computer and browser to open websites, create files and use accounts a user connects, and placing phone calls on a user's behalf, for example ringing a hotel or vendor, waiting on hold, asking questions and reporting back.

Payments run through the same thread. The company says Tab can place orders, book reservations and pay invoices, but before spending it shows what it is buying, the total and who will be paid. "Nothing is paid until you approve it," the site states. Once approved, it pays, posts the receipt in the conversation and tracks the next steps.

Access is limited for now. Tab is in private beta, and prospective users join a waitlist by entering a mobile number and replying to a text message. The website does not list prices, and TechCrunch's report did not include any.

Selling trust in a crowded race

Tab is entering a busy market. TechCrunch placed it alongside Instinct, which it described as a breakout hit, and Meta's Muse as part of a rising wave of consumer assistants that handle errands such as making reservations and booking travel. Meta brought Muse to the iPad on 7 October, a month after its mobile debut, according to a separate TechCrunch report.

Investor appetite for personal AI has spread to hardware as well. On 5 October, TechCrunch reported that Ghost, whose founder is 19, had raised $11 million to build a $3,499 computer for personal AI.

Amdani acknowledged the competition. "Everyone is racing to make AI more capable," he said, while arguing that the winners will not necessarily be the companies with the most impressive demos. "We think the harder question is whether you should let it in."

His pitch draws on an older model of personal service. "Your grandparents had a family doctor, a banker, someone who knew them and looked out for them," he said. "Most of us don't have that anymore." Tab wants to occupy that role, which in practice means asking users to hand over access to their email, phone calls and payment cards.

The company has made specific promises in response. Amdani said user data is never used to train AI, that card details sit in a vault Tab itself never sees, and that nothing irreversible happens without the user's explicit approval. Tab's website adds that it keeps each user's work separate from other users'. "Capability is table stakes. Trust is the product," Amdani said.

Those commitments go to the central risk of agentic AI. An assistant able to spend money, send messages and speak on the phone in someone's name can cause real harm if it misreads an instruction, is manipulated by a malicious website or is compromised. Approval steps for purchases reduce that risk but add friction, and companies in this market are still deciding where to draw the line between convenience and control.

What comes next

Tab says most of its growth so far has come through word of mouth, and it plans to launch a new product soon. It gave no details on what that product will be or when it will arrive.

The absence of funding figures leaves open questions. Without knowing how much capital Tab has raised, it is hard to judge how long it can afford to compete with Meta and other well-resourced rivals, or how much of the $300 million valuation reflects revenue rather than expectation. The investor list offers some signal: SV Angel is a long-established Silicon Valley seed investor, and Valar Ventures is the firm co-founded by Peter Thiel.

Readers looking up the company should note that it is distinct from an earlier start-up also called Tab, a wearable AI companion developed by Avi Schiffmann, which CB Insights lists separately. The new Tab's website names Terrasoft, Inc. as the company behind the service.

For Indian users, the choice of WhatsApp as one of Tab's two channels is notable, since India is WhatsApp's largest market. Tab has not said which countries its beta covers, and a service that makes calls, places orders and pays bills generally needs local integrations, payment rails and merchant relationships before it can work well in a new country.

The immediate test is reliability. A consumer assistant earns permission to handle bills and bookings one successful errand at a time, and loses it with a single unauthorised charge or botched reservation. Tab's approval-first design is built to limit that downside; whether users will tolerate the extra taps in exchange for peace of mind is what its private beta will now have to show.