Set Up Your Co-Worker

This walkthrough takes an empty account to a co-worker that answers questions using your systems and your organization's knowledge. Nothing here requires a workflow or a line of code.

Before you begin

• A FlowGenX Chat account. Co-Worker Studio is one-to-one with the account, so no provisioning request is needed.
• At least one connected app for the systems you want it to reach, set up in Integration Studio. A co-worker with no connections can still reason, retrieve knowledge, and remember, but it cannot read your task board.
• Optional: documents loaded into the Knowledge Base if you want it to answer policy and process questions.

Step 1 - Open Co-Worker Studio

Goal: provision the co-worker and land on the tab where it gets its name. (~1 min)

1. In FlowGenX Chat, open Co-Worker Studio from the left rail.
2. On a fresh account, a draft co-worker is created for you automatically and the studio opens on the Identity tab.

Step 2 - Give it an identity

Goal: fill the four identity fields that decide who the co-worker is and what it takes on. (~5 min)

Either edit the fields directly on the left, or describe the role in the Identity studio chat on the right and let it draft them.

1. Name. Give it a human name. It appears as the author of every message it posts.
2. Job title. Be specific. The title also drives which skills the library recommends.
3. Personality. How it should sound. Two or three traits is enough.
4. Job description. The most consequential field. Name the systems it owns, the decisions it can make, and the work it should hand back.

The checklist header tracks progress as N of 4 set.

Step 3 - Equip its first skills

Goal: give it the saved procedures that match its role. (~4 min)

1. Open the Skills tab. The top card is Equipped skills, the set the co-worker uses on its own.
2. In Add skills below it, browse the library. Filter by tag, or search by name.
3. If the job title matches a known role, a Recommended banner offers a small bundle you can add in one click.
4. Equip only what the role needs. Skills are capability, and unequipping is the cheapest way to keep a co-worker inside its lane.

Step 4 - Check what it can already retrieve

Goal: confirm knowledge retrieval works before anyone depends on it. (~3 min)

1. Open the Context tab.
2. In the Context playground, ask a question your knowledge base should answer, for example a policy or pricing question.
3. Read the Retrieved column on the right. Each result carries its store, its relevance score, and its text, so you can see which source produced the answer.
4. Switch the lower card to Knowledgebase to browse what is synced and which collections are in scope.

Step 5 - Enable it

Goal: publish the persona so it goes live. (~1 min)

1. Flip the Enable switch in the studio header.
2. A confirmation appears naming the co-worker as live.

Enabling is what activates the persona, its managed skills, the activity digest, and access to conversation history and the knowledge base. While it is off, the account answers as the default assistant without any of them. The switch takes effect on the very next message, so it also works as a stop control. See Governance and control.

Step 6 - Give it real work

Goal: confirm it can act in your systems, not just talk about them. (~5 min)

Ask it something that requires a tool call against a connected system. A good first request names the system, the scope, and the shape of the answer you want:

Look through our ClickUp workspace and find all tasks that are overdue or look
blocked. For each one tell me the task name, who owns it, its status and the
due date.

Watch what comes back. A healthy first run loads a skill, runs the scan, and returns a structured answer with owners and reasons rather than a summary.

Where to go next