Dashboard Tour and Entry Points

The Dashboard is the operating surface for administrators and developers. This chapter covers what's in it — and why it is not your only option.

The Dashboard is not mandatory

Position it correctly: the Dashboard is one of four equal entry points, not the platform itself.

Everything it does — query data, edit data, create tables, set permissions, trigger business actions — goes through the same service layer that also backs the REST API and the Custom App SDK. Which means:

• Need to feed data into an existing system? Use the API. Don't open the Dashboard.
• Need staff to do daily work? Use a Custom App. Don't grant Dashboard access.
• Need to create records in bulk? Use the API or data import. Don't click through one at a time.

What the Dashboard is genuinely irreplaceable for: initial configuration (permission model, approval workflows, business rules), the development environment (App Builder), and audit and observability (logs, usage, resources).

The six navigation sections

1. Overview

• Home — a data lineage canvas showing how data modules and applications connect. Pan, zoom, and click a node to open a side panel with live statistics and recent changes.
• AI Usage War Room — an analytics dashboard for AI consumption: call volume per capability, AI credit spend, most-retrieved knowledge files, Builder development activity, and staff engagement. This is the cost and usage observability surface.

2. AI Applications

• App Builder — the Custom App development environment (Chapter 7)
• Integrations — third-party integrations and API key management (Chapter 11)
• Compute Resources — quota and consumption view

3. AI Agent

• Cowork — the AI assistant conversation interface (Chapter 16)
• Cowork Settings — AI skill and MCP server management

4. Organization & Access Control

• Team Information — organization settings
• Member Management — invite, deactivate, assign roles
• Roles & Permissions — RBAC role definitions and permission tags
• Single Sign-On — Google SSO configuration
• My Approvals — your personal approval queue

5. System & Operations

• Usage Analytics — consumption and quota per metered dimension
• Approval Workflow Engine — approval definitions (Chapter 13)
• Audit Log — complete record of sensitive operations
• App Schedules — organization-wide view of Custom App scheduled tasks (Chapter 9)
• Core Dictionaries — system parameters, reference data, numbering rules

6. Data Center

Where operational data lives — CRM, sales, purchasing, inventory and warehousing, manufacturing (MRP), accounting, projects, and HR, plus data import, exports, announcements, and external data connectors. See Chapter 4.

The Knowledge Center

The Knowledge Center (/dashboard/files) is a top-level item alongside the Data Center, not a child of it.

It stores enterprise documents and doubles as the source for AI retrieval. Any file can be marked as knowledge, which indexes it into your organization's dedicated vector store for retrieval by the AI assistant and by Custom Apps. Each file carries its own access policy. See Chapter 6.

Data Center tables: editable, but usually shouldn't be edited by hand

Every table under the Data Center offers a full grid view with editing. These are real database tables, not reports.

But routine operations should not go through the grid. The reason isn't permissions — it's process integrity:

• Writing through a Custom App or the API passes through form validation, approval interception, and business-engine cascades
• Editing the grid directly bypasses some of those cascades, which can leave the books out of step with reality

The right use: treat the grid views as verification and audit tools, and route writes through Custom Apps, the API, or the AI assistant.

Permissions determine what you see

Every navigation item is bound to a permission tag. If you can't see something, your role lacks that permission — the feature exists.

Common tags:

To change access, have an administrator edit the role definition under Roles & Permissions rather than granting individuals one at a time.