How it works

You don't need to know the machinery to use the Associate — but a little of it explains why its answers are grounded and why you stay in control.

It knows where it is — and who you are

Every time you send a message, the Associate is given fresh context:

Who you are — your name, your role, and your company, taken from your secure login. This comes only from how you signed in, never from the chat itself, so the Associate can't be talked into thinking it's someone else.
Where you are — the project you're viewing (its tasks, timeline, documents, and team), or your personal context on the dashboard.

This context is rebuilt on every message, so a long conversation never works from a stale snapshot — it always reflects the current state of your project.

Two model tiers

You can run the Associate on one of two tiers, depending on the job:

Tier	Best for
Standard	Fast, everyday questions and quick changes.
Max	Deep analysis — longer reasoning, more steps per request, and the only tier that can plan timelines or generate documents.

Which tiers you can pick depends on your role (see Permissions & data).

The Associate streams its answer as it goes, and shows its reasoning as it works through a problem — so you can follow how it reached a conclusion rather than getting a black-box result. You can stop it at any time.

Memory

The Associate remembers the current conversation, and you can give it notes to hold onto — a "Create Memory" action and the /memory save command let you record context you want it to keep in mind. It can also pull in relevant details from earlier conversations when they help. You're always in control of what it remembers.

Specialists behind the scenes

The Associate is the conversational front, but it's backed by a set of specialist agents that each do one job well — for example the Daily Brief on your dashboard, plus behind-the-scenes helpers for task analysis, time estimates, snapshot reviews, and productivity insights. You talk to the Associate; it draws on the right specialist when needed.

Last updated: 2026-05-31

PreviousWhat it can do Next Review & approve

How it works

You don't need to know the machinery to use the Associate — but a little of it explains why its answers are grounded and why you stay in control.

It knows where it is — and who you are

Every time you send a message, the Associate is given fresh context:

Who you are — your name, your role, and your company, taken from your secure login. This comes only from how you signed in, never from the chat itself, so the Associate can't be talked into thinking it's someone else.
Where you are — the project you're viewing (its tasks, timeline, documents, and team), or your personal context on the dashboard.

This context is rebuilt on every message, so a long conversation never works from a stale snapshot — it always reflects the current state of your project.

Two model tiers

You can run the Associate on one of two tiers, depending on the job:

Tier	Best for
Standard	Fast, everyday questions and quick changes.
Max	Deep analysis — longer reasoning, more steps per request, and the only tier that can plan timelines or generate documents.

Which tiers you can pick depends on your role (see Permissions & data).

How it works

It knows where it is — and who you are

Two model tiers

You can watch it think

Memory

Specialists behind the scenes

Next

How it works

It knows where it is — and who you are

Two model tiers

You can watch it think

Memory

Specialists behind the scenes

Next