Quickstart

Prerequisites

Before you begin, make sure you have:

An active Stackless user account.
Your organization's Stackless URL, usually a customer-specific Stackless subdomain.
Access to the dashboards, sources, and schemas your role is allowed to view.
A modern browser such as Chrome, Edge, Safari, or Firefox.

If you cannot sign in or a dashboard is missing, ask a Stackless administrator to confirm your invite, role, and dashboard access.

Open your Stackless URL and sign in with the email and temporary password from your invitation email; you set your own password the first time. If your organization uses single sign-on, sign in with your company account instead. If you forget your password later, reset it from the sign-in page.

After sign-in, the left navigation shows the product areas available to your role: Overview, Agent, Monitors (where enabled), Dashboards, Operations, and Data (Sources, Catalog, Transformation Models, and Semantic Models), with Support and Settings at the bottom. A short product tour introduces the main areas on your first visit; replay it from ? > Restart product tour.

Step 2: Open the Agent

Go to Agent. The Agent page has three main areas:

Conversations on the left, with your conversations under Owned and conversations others shared under Shared with you.
Chat in the center, where you ask questions and review responses, the Agent's reasoning, and its evidence.
Artifacts on the right, where generated tables, chart cards, briefs, findings, and dashboard drafts appear.

Try one of these first prompts:

"What data do we have available? Show me the key tables and metrics."
"Show me all metric definitions and what they measure"
"Build me a dashboard to track our key performance metrics"
"Query our data warehouse to analyze recent trends"

Step 3: Ask a precise first question

For the first question, include a metric, a time period, and the breakdown you want. For example:

"Show weekly net revenue for the last 90 days."
"Create a table of orders by channel for this month."
"Which locations had the largest revenue increase versus the prior month?"
"Explain how the active customers metric is calculated."

The Agent may search the Catalog, inspect metric definitions, run a read-only warehouse query, or create an artifact. Expand Reasoning & tool calls to see what it did, and Evidence to check the sources, filters, and definitions behind the answer. It respects your role's data access.

Step 4: Save the useful output

When the Agent produces a result you want to keep, ask it to save the result as the right artifact:

"Save this as a table artifact."
"Turn this trend into a chart card."
"Create a brief with the main findings and supporting evidence."
"Add this chart to a new dashboard."

Artifacts stay linked to the conversation that created them. You can also attach existing artifacts to a new conversation when you want the Agent to continue from prior work.

Step 5: Move from artifact to dashboard

If a result should become recurring reporting, ask the Agent to build a dashboard:

"Create a dashboard with this KPI, the weekly trend, and a table by channel."
"Add filters for date range and location."
"Publish this dashboard after I review it."

Draft dashboards are private while you work. Publishing is an explicit action: the publish dialog shows any models the dashboard depends on, and you confirm before anything goes live.

Step 6: Automate and keep it fresh

For recurring work, use:

Dashboards to view published, private, and shared dashboards, and to schedule email exports.
Monitors to run a check or analysis every day, or after data refreshes, and deliver it to Slack or email.
Data > Transformation Models to see whether each model is fresh or stale.
Data > Catalog to search assets and trace lineage.
Operations to review refresh jobs, run history, and failure alerts.

Next steps

Core Concepts explains the main objects in Stackless.
Using the Agent covers conversation workflow in more detail.
Artifacts explains tables, charts, briefs, findings, and dashboards.
Transformations explains managed Transformation Models.
Monitors explains scheduled analyses and checks.