Everyone will get their own AI assistant

In the Nebula Lab we have been testing OpenClaw, an open-source personal AI assistant, since its first release.

After a good month of daily use — and near-daily updates from a very lively community — the direction of AI assistants for personal use is clear.

An assistant that lives with you

OpenClaw is a local-first, open-source personal assistant: it runs on your own machine, keeps its workspace and configuration on your side, and meets you in the channels you already use — messages, files, calendars. You choose which model providers and services it may reach, and a clear security boundary separates the assistant from the outside world.

It is early software and it moves fast — new versions have been arriving almost every day. That pace is itself the signal: this category is being built in the open, and it is maturing in months, not years.

OpenClaw — a personal AI assistant →

The direction is clear

In the not-so-distant future most of us will have our own AI helper for everyday, general use. What remains genuinely open is how and where these assistants will be embedded — in the phone, the operating system, the browser, the messenger, the home — and exactly that question currently occupies a good part of the research community.

What it means for teams and companies

How such assistants can help teams and companies is a question time will answer. Our working assumption in the Lab: we are very likely heading towards hybrid organisations, where selected business functions or roles are delegated to expert agents — companions in our regular work, not replacements for it.

For businesses this raises the questions we keep returning to across our AI work: access and permissions, auditability, data sovereignty — and a person deciding wherever it matters.

A closing thought

Personal assistants are being born in the open, in front of everyone.

The organisations that learn to work with them early will set the pace for everyone else.

Nebula AI →