← Blog

The rise of the personal assistant for everybody else

One entity, one number, one relationship.

Lawrence Coburn
4 MIN

For the last few weeks in SF, a lot of the conversation has been dominated by a couple of surging products in the Personal Assistant space: Instinct, Muse, and Dots.

We have written a lot on this blog about how the mental model of how you work with AI matters, and we see three clear patterns emerging.

The first is the magic bag of tools. ChatGPT, Claude, skills, scheduled jobs, self-contained chat threads. You remain the vessel of the work, and you pull the right tool for the job. AI makes you faster but you are still the one doing everything.

The second is a personified AI teammate with a specific job. They likely have a Slack account, a phone number, an email address. You communicate with it like you are communicating with a person. You do your best to train them up, and fire them if they go off the rails.

The third mental model is what is happening right now. Call it the personal assistant or AI Chief of Staff for everybody. It’s one personified entity. It has a phone number you can text. A WhatsApp channel. Its mission is not to help your company ship faster, it is to help you navigate your life. Zuckerberg calls his version a personal superintelligence. OpenAI describes theirs as an AI helper that always has your back. Instinct describes itself as the best personal agent that can handle the deeply personal nuances of everyday life.

These are consumer offerings with ambitions that match the scale of ChatGPT or Claude themselves.

Here are the three products in the space that are surging.

Instinct is an independent company now valued at $10 billion that is in full on rocketship mode. They raised $1.3B in their Series A, Series B, and Series C, all of which occurred over three months. They are growing at a reported 10% a day. The product is invite-only, so all of that growth is referral. You communicate with Instinct primarily via text. The founder is 23 years old, came out of Sierra, a well regarded AI shop, and was the first author of the 2023 Reflection Paper.

Muse launched September 8 out of Meta. It hit 5 million downloads in 22 days. For context, ChatGPT took 56 days to reach that number. Claude took 492. You communicate with Muse primarily via WhatsApp or Facebook Messenger. The allstar team behind it includes Nat Friedman, longtime entrepreneur and former CEO of GitHub, and Alexander Wang, Meta’s Chief AI Officer.

Dots launched September 29. No growth numbers yet but OpenAI has enormous reach. Altman describes it as an AI helper that always has your back. You can delegate ambitious pieces of work the way you would to a high-agency engineer.

I have been testing Instinct for a while, and here’s how I have been using it.

I connected my calendar, inbox, and a few other personal systems. I had it pull all the birthdays of family and friends from my calendar and set up notifications. I had it order flowers for my wife. It now has my credit card. Flowers arrived as expected. It proactively flags things in my inbox I have not responded to. Helped me file some tax paperwork in San Francisco related to Airbnb. Scheduled roofers via Yelp after we developed a leak. Restaurant reservations. Travel bookings.

I am impressed. The product is still free, which will obviously change. But for handling emails from my daughter’s school and things that have been sitting on my to-do list for weeks, it is already earning its place.

What does this mean if you are following the space?

The magic bag of tools works great for one-off projects. You invoke AI when you need it. Each task can live in its own thread. That model is subsidized when you are going direct to the foundation model.

The AI teammate model, which we have written about here at length, works well if you have managed people before. Different agents do different things. You hire them, train them, give feedback. It is intuitive if hiring and delegation are already in your muscle memory.

The personal assistant is something else. It is closer to having a talented EA. One entity, one number, one relationship. It handles the personal and the professional, sometimes at the same time.

This is a category to watch. A year from now I think most people will have some version of this running. The question is which one they trust with their calendar, their inbox, and their credit card.

Lawrence Coburn
Founding Partner at Legible. Works with executives to set AI strategy.