Skip to main content

​
Get started in three steps

Set up your Utari workspace, configure your first AI agents, and run your first automated task.

​
Step 1: Create your workspace

  1. Visit utari.ai in your browser
  2. Click “Get started” and sign up using Google, GitHub, or email
  3. Complete email verification if required
  4. Wait while Utari sets up your workspace automatically
Your workspace is your private environment where agents, data, and settings live.
Follow the guided setup to personalize your AI workforce:
  1. Choose your usage type: Individual or company/team
  2. Select starter agents: Pick from research, operations, content, or technical agents
  3. Configure agent settings (optional):
    • Set names and roles
    • Define primary goals
    • Choose tone and style (formal, concise, friendly)
  4. Click “Start Working” to reach your dashboard
You can refine or add agents anytime—this setup just gets you productive fast.

​
Step 2: Run your first task

  1. From your dashboard, click “New Thread” or open an existing thread
  2. Select your primary agent (e.g., your generalist agent)
  3. Describe your goal clearly:
    • What you want: “Summarize this market report into a 1-page executive brief”
    • Context: Who it’s for, length, tone, deadline
Attach files or links using the upload button to give your agent source material.
Your agent will:
  • Read your files and analyze content
  • Browse or research (if tools are enabled)
  • Generate structured output (summaries, plans, drafts, checklists)
Refine the output with follow-ups:
  • “Make this shorter and more executive-friendly”
  • “Turn this into an email for my team”
  • “Highlight the top 3 risks and mitigations”
All context stays in the thread—agents remember everything!

​
Step 3: Customize your AI workforce

Edit existing agents:
  1. Go to Agents / Workforce section
  2. Open an agent and update:
    • Description and expertise areas
    • Preferred output formats
    • Tone and communication style
Create specialized agents for recurring workflows:
  • Research Analyst: Market and competitor research
  • Operations Coordinator: SOPs, process docs, checklists
  • Client Delivery Assistant: Client-ready deliverables
When an agent produces great output, save it as an example to guide future work.

​
Core concepts

Workspace

Your private environment where agents, data, and settings live. Invite teammates when team features are enabled.

AI Agents

Specialized AI workers configured with goals, tone, and tool access to handle specific workflows.

Threads

Persistent conversations that keep history, context, and files so agents can pick up where they left off.

Tools

External capabilities like web browsing, data APIs, and sandboxes that let agents act, not just chat.

​
Common workflows

Example: “Research the top 5 vendors for X, summarize their offerings, pricing ranges, and pros/cons in a table.”Agents can browse the web, pull data, and create comprehensive comparison reports.
Example: “Take these meeting notes and turn them into an action plan with owners and deadlines.”Transform raw notes into structured workflows, SOPs, and trackable deliverables.
Example: “Convert this technical analysis into a client-ready summary in simple language.”Repackage internal work into polished, stakeholder-appropriate documents.
Example: “Help me plan my week based on these tasks; prioritize and group them, then create a schedule.”Let agents handle planning, prioritization, and time management.

​
Advanced features

​
Connect tools and data

Unlock powerful capabilities by connecting external tools:
1

Enable web & data tools

Allow agents to research the web, pull public data, and enrich information with search and crawling capabilities.
2

Add execution tools

Enable code execution and sandboxes for data analysis, transformations, and automation tasks.
3

Integrate external systems

Connect your existing tools so agents can read/write data, sync projects, and trigger automations.
Start with view-only / low-risk access, validate agent behavior, then gradually expand permissions.

​
Next steps

Run real tasks

Complete 3–5 real tasks with your primary agent (research, planning, writing).

Create specialized agents

Build one agent for a recurring workflow you care about.

Connect tools

Enable web, data, or sandbox tools to unlock powerful workflows.

Invite your team

Share threads and agents to make your AI workforce a shared asset.
Treat agents like teammates: Give them clear roles, good context, and feedback—then let them own entire workflows, not just single prompts.

​
Troubleshooting

Try refreshing your dashboard. If the issue persists, check your workspace settings.
Continue to the dashboard and retry setup later from your settings.
Use the in-app support channels or join our community (Discord/email) linked from the Utari interface.