Appearance
Your First 30 Minutes as an Admin
This walkthrough gets you from a fresh DMOps.ai tenant to your first agent running a task. Follow these 5 steps in order.
Time required: ~30 minutes
What you'll have at the end: One AI agent fully configured and running your first task.
Step 1 — Sign in and explore (3 min)
- Open your tenant URL (e.g.
https://acme.app.dmops.ai) in your browser. - Sign in with the admin credentials provided during provisioning.
- Take a quick look around:
- Task Board — this is where all work lives
- Settings — everything you'll configure is here
- Observability — cost and execution data appears here after your first task
Step 2 — Set up your org chart (10 min)
Your org chart defines who's in your organisation. Start with one agent.
Go to Settings → Org Chart → + Add Member.
Set Type to Agent.
Give it a name and role — keep it simple for now (e.g. Assistant, role: General).
Write a basic system prompt:
You are a helpful AI assistant for [your company name]. Your job is to complete tasks clearly and thoroughly. When you're done, summarise what you did in 2-3 sentences.Leave all other settings at their defaults.
Click Save.
✅ You now have one agent in your org chart.
Step 3 — Connect a model provider (5 min)
Your agent needs an LLM to think with.
- Go to Settings → Model Providers → + Add Provider.
- Choose your provider:
- AWS Bedrock — recommended if you're running in AWS
- Anthropic — easiest to get started (API key from console.anthropic.com)
- OpenRouter — access to many models with one key
- Enter the API key and click Save & Test.
- Click Set as Default and choose a model (e.g. Claude 3.5 Haiku for cost-effectiveness).
✅ Your agent now has a brain.
Step 4 — Run your first task (5 min)
Let's test the whole pipeline.
- Go back to the Task Board.
- Click + New Task.
- Enter a simple title: Introduce yourself and describe what you can help me with.
- Leave the assignee blank (the dispatcher will route it to your new agent).
- Click Create Task.
- Watch the task move: Triage → Ready → Running.
The task usually completes within 30–60 seconds.
- Once status is Done, click the task card → open the Trace tab to see what the agent did step by step.
✅ Your first agent task completed successfully.
Step 5 — Review results and next steps (7 min)
- Open Observability → Spending to see the token cost of your first task.
- Open Observability → Execution Logs to see the full run log.
You're ready to go deeper:
| Next step | Guide |
|---|---|
| Add more agents with real roles | Org Chart |
| Connect your real systems (ERP, Jira, etc.) | Integrations |
| Write skills for your business workflows | Skills & Tools |
| Invite your team | Access Control |
| Set budget limits before going live | Model Providers & Budgets |
TIP
The fastest way to learn DMOps.ai is to run real tasks. Start with low-stakes workflows (research, summarisation, data formatting) before moving to tasks that touch your live systems.