Ask your own assistant about your team
Connect your Claude or ChatGPT account to Worktivity and ask in plain language: what the team did last week, which project ate the most time, who is on leave.
The connection belongs to the person, not to the workspace, so everyone links their own account. And it never widens what anyone can see: an employee gets their own data, a manager their own team, an owner the whole workspace. Exactly the same limits as the screens they already use.
Included in Growth and Pro, and in the free trial.
Authorised once, with your own account
Worktivity speaks MCP, the protocol assistants use to reach outside tools. Nothing is pasted anywhere: you approve the connection by signing in, and you can end it whenever you like.
An owner opens the door
The feature ships switched off. An owner turns it on in workspace settings and picks who may connect: owners only, owners and co-owners, managers as well, everyone, or a named list of people.
You link your assistant
Add Worktivity to Claude or ChatGPT and sign in with your Worktivity account. Authorisation runs over OAuth, and one connection covers every workspace you belong to. You name the workspace in each question.
You ask, in your own words
The answer is built from your workspace within the limits of your role, and the call is written to the audit log the owner can read.
One question, three different answers
Not because the assistant is guessing, but because every answer is cut to the role of the person asking before it is built.
- As an employee
How many hours did I work last week?
Only your own data comes back. No colleague appears in the answer, because the question never reaches theirs.
- As a manager
How is my team doing compared with last month?
Your own team, and nothing outside it. The same boundary the panel draws for you.
- As an owner
Who is on leave next week?
The whole workspace, because that is what you already see when you open Worktivity.
This is not a promise about wording, it is how the connection is built. Adding somebody to the allowed list decides who may connect an assistant. It never decides what they get to see.
Nothing opens by itself
This is an employee monitoring product, so the connection carries data the workspace cares about. Every switch below sits with the owner.
Off until you turn it on
No workspace gets this by surprise. Until an owner enables it, nobody can connect anything at all.
You choose who can connect
Owners only, owners and co-owners, managers as well, everyone, or a handful of named people. The choice can be widened or narrowed any day.
Cut a connection in one click
Owners see the connections approved in their workspace: who linked which assistant, with which role, and when it was last used. One button ends it. And every question asked about the workspace reaches the audit log, whichever connection it came from.
Every question leaves a trace
Each call is written to an audit log the owner reads in the panel: who asked, when, and which tool ran. The wording of the question itself is not stored.
Reading is the default
Writing is a separate switch and it also ships off. Switched on, the assistant can create and update projects, customers, tasks and task comments.
No deleting, no HR decisions
There is no delete tool anywhere on the surface. Employee records, leave decisions, payroll and billing cannot be touched through the connection either.
Or just ask inside Worktivity
There is a chat screen in the panel too. Same questions, same role limits, nothing to link and nothing to install.
In Growth and Pro, and in the free trial
Starter does not include AI connections. Every other plan does, and so does the trial, so you can try it before you decide.
See the comparisonQuestions people actually ask
Which assistants can I connect?
Claude and ChatGPT, using your own account on either. The link is made over MCP, the protocol those assistants use to reach outside tools, so any client that speaks MCP can sign in the same way.
Can the assistant see more than I can?
No. The answer is narrowed to your role before it is built: an employee gets their own data, a manager their own team, an owner the whole workspace. Being on the allowed list changes who can connect, never what they can see.
Does it use Worktivity AI credits?
No. The model is yours, so asking through the connection runs on your own Claude or ChatGPT account. That also means the relationship with the provider is your relationship, not something Worktivity resells.
Do screenshots go to the assistant?
No. Screenshots, timelapse videos, visited addresses and window titles are not on the surface at all. Aggregated figures travel; row-by-row browsing history does not.
Can the assistant change things in Worktivity?
Only if an owner turns on the separate write switch. With it on, projects, customers, tasks and task comments can be created and updated, and each change is recorded in the person's own name. There is no delete tool, and employees, leave, payroll and billing stay out of reach.
Which plans include it?
Growth and Pro, and the free trial. Starter does not include it. If a workspace moves down to Starter the existing connections stop answering, but they are not deleted.
Bring your own assistant
Create a workspace, switch the connection on and ask your first question the same day.
Included in Growth and Pro, and in the free trial.