Why we built Deputant · Deputant Blog

A new AI agent comes out almost every week. Meta has Muse. xAI has Grok Bot. OpenAI has Dots. Smaller teams ship agents like Instinct and Hark. Each one does some things well, and the smartest one keeps changing.

That is good news. Competition makes agents better, fast. But it has a cost, and you pay it every time you switch.

Every switch starts from zero

Think about the last time you tried a new agent. You connected your email. You added a card. You typed your home address, then your work address. You explained that you hate aisle seats and that you only buy fragrance-free soap.

Then a better agent came out, and you did it all again.

When switching is this much work, two habits form. Some people stay with an agent that is no longer the best, because switching is too much work. Other people give each new agent access to everything, because choosing item by item is too much work. Neither habit is good for you.

Set up your digital life once

We think you should set up your digital life one time. Then any agent you pick should be able to use it, with your permission.

That is what we are building. Deputant will keep these things in one vault:

Passwords for the websites you use
Cards you pay with
Addresses for home, work, and the people you send gifts to
Apps like your calendar, email, and documents
Memories like your sizes, your favorite brands, and what you like

When a new agent comes out, you won't start over. You will choose what it can use. That's the whole setup.

You choose what each agent can use

One vault for every agent only works if you stay in control. So each item in your vault will have its own rules. You choose which agents can use it.

Maybe your travel agent can use your name and home address, but not your work calendar. Maybe a shopping agent can use one card, but not the other. Maybe a new agent you are only testing gets nothing until it earns your trust.

You can change these rules at any time. If an agent stops being useful, you remove it. Your vault stays where it is.

A second set of eyes on your data

Rules help, but agents ask for things in ways that rules can't always predict. So we are also building a privacy filter.

The privacy filter is an AI layer between your vault and your agents. When you turn it on for an agent, it checks each request. Then it hides what that agent does not need for the task.

Say a booking agent wants to hold a table for dinner. It needs your name and your phone number. It does not need your card number. The privacy filter sends the first two and holds back the card.

The privacy filter will make mistakes sometimes. It does not replace your rules. It is a second set of eyes that helps catch what your rules miss.

Why now

Agents can do real tasks today. They shop, book, fill in forms, and sort email. To do that, they need real access to your life. The more useful agents get, the more of your life you hand them.

We want that handoff to be easy. We also want it to be safe, and we want you to decide where the line is.

We also don't want to pick a winner. We don't know which agent will be the best next month. You don't either. You shouldn't have to know before you set things up.

Where we are today

We want to be plain about this. Deputant is not live yet. We are still building the first version. We have no users or results to show you yet.

The agents we named above are made by other companies. We have no partnership with any of them. We name them because they are the agents people talk about this year.

Here are the rules we build by:

You own your data.
You approve what agents and humans do.
The privacy filter is a second set of eyes on your data.
You can use any agent, and switch when a better one comes out.