Dashboard

The Dashboard is the operational control center of your Karyam workspace.

It provides a real-time view into the health, activity, usage, and cost of your AI systems.

Rather than navigating through individual agents, flows, or runs, the Dashboard gives teams a high-level understanding of how AI is being used across the organization.

Why the Dashboard Exists

Production AI systems generate a large amount of operational data:

Agent executions
Workflow runs
Token consumption
Model costs
Retrieval activity
Human approvals
Failures and retries

The Dashboard aggregates this information into a single operational view.

What You Can Monitor

The Dashboard surfaces information across four major areas:

Usage

Understand how AI resources are being consumed.

Examples include:

Total token consumption
Model usage trends
Most active workflows
Most active agents

Cost

Monitor spending across models and workloads.

Examples include:

Monthly model costs
Daily token costs
Cost trends over time
Most expensive workflows

Activity

Track execution activity across the workspace.

Examples include:

Recent runs
Workflow executions
Agent activity
Retrieval operations

Performance

Identify bottlenecks and operational issues.

Examples include:

Failed runs
Long-running workflows
Approval bottlenecks
Expensive retrieval operations

Typical Dashboard Widgets

Depending on your workspace configuration, the dashboard may include widgets such as:

Monthly Total Token Consumption
Monthly Cost Trend by Model
Monthly Total Model Cost Overview
Daily Token Consumption
Daily Token Cost
Highest Token Consuming Flows
Top Expensive Workflows

Operational Visibility

The Dashboard helps answer operational questions such as:

Which workflows consume the most tokens?
Which models generate the highest cost?
Which systems are most heavily used?
Are there unusual usage spikes?
Are workflows waiting for approvals?
Are retrieval operations behaving as expected?

Relationship to Other Concepts

The Dashboard aggregates information from multiple Karyam systems:

Runs
    +
RAG Runs
    +
Models
    +
Approvals
    +
Audit Logs
        ↓
Dashboard

The Dashboard does not generate or execute workloads.

Instead, it provides visibility into workloads already running within the platform.

The Karyam Philosophy

Building AI systems is only the first step.

Operating them reliably requires visibility.

The Dashboard provides the information teams need to understand usage, optimize costs, and operate AI systems confidently in production.