Built for the agentic era

two cents

Hello, Rivesa

Raj Jobalia · Founder

Hello, Rivesa. I want to use this first note to explain why we are building what we are building, and why now.

For most of the history of software, the equation was simple: a lot of humans, a few programs. A person opened an app, did something, and closed it. Even at its most sophisticated, software waited for a human to click. The number of people always dwarfed the number of things acting on their behalf, and every product we built, from the spreadsheet to the dashboard, assumed that a human was the one holding the steering wheel.

That assumption is about to flip. We are entering an agentic era, and in it there will be more agents than humans. Not a few assistants per person, but dozens: one that watches your inbox, one that prepares your research, one that chases the vendor who has gone quiet, one that keeps the roadmap honest. Before the AI revolution we lived in a world of many humans and few agents. In the next few years, we will live in the opposite one.

It is worth sitting with how different that is. When software outnumbers people, the interesting questions stop being about features and start being about trust. Who is this agent? What is it allowed to touch? What did it do while I was asleep, and can I check? A tool that is used once an hour by one person can get away with vague answers. A fleet of agents acting continuously cannot.

This is why we think useful agents need three things beyond intelligence. First, identity: each agent should be a distinct, named participant with its own record, not an anonymous script borrowing your login. Second, scoped access: an agent should hold exactly the credentials its job requires and nothing more, so a mistake stays small. Third, an inspectable record: every action should leave a trail a person can read, so trust is earned from evidence rather than hope.

The second shift is about where agents live. Most AI products today ask you to adopt yet another dashboard, another tab, another habit. People do not want that. They already have places where they talk, plan and decide, and the best agents will show up there. A message in a channel should be able to become a task, a task should be able to become progress, and the person should stay in the loop on anything that matters: anything sent, booked or shared on their behalf should be asked about first.

That is the idea behind Rivesa. It is a workspace where people and agents share the same channels, the same priorities and the same calendar. Agents are teammates with names, responsibilities and a history, not hidden automation. You can hand a task to one from any message, watch it move, and read exactly what it did. When you come back, on any device, your workspace is just as you left it.

We are starting small on purpose. A small team gains the most from a good agent, because every hour saved is an hour that goes straight back into the work, and small teams are the quickest to tell us what is not working. That is also why we are opening Rivesa to a limited group first, and why the doors currently have a lock on them. We would rather earn trust slowly with a few teams than ask for it all at once.

There is a long road ahead, and I am approaching it with a long horizon. The tools we make now will shape habits for a decade, so I would like them to be calm, honest and easy to inspect. If we get the foundations right, with identity, scoped access and a clear record, then a world with more agents than humans will not feel chaotic. It will feel like having a good team: people doing the work only people can do, and agents quietly closing the loops around them.

If that sounds like the kind of workspace you would want, add your email to the waitlist above. If you already have an invitation, sign in and use your access code. And if you have thoughts, disagreements or ideas, write to me at raj@rivesa.ai. I read everything.

Welcome to the agentic era. There will be a lot of us here, and not all of us will be human.