AI Utilities
Overview
The AI Utilities module is your organization's control center for managing all AI-powered capabilities within Orbit Resolve. Orbit leverages advanced cloud-based AI models (such as Google Vision, DeepSeek, and specialized LLMs) to process data intelligently across the platform—from extracting text from receipts to analyzing support emails.
Because these models consume computational resources, they operate on a Credit system. The AI Utilities settings page allows administrators to monitor real-time credit usage, enforce monthly spending limits per utility, and seamlessly toggle offline (cost-free) fallbacks when necessary.
Accessing AI Utilities
To access the AI Utilities configuration, follow this navigation path: Cockpit → Settings (Organization Settings) → AI & Intelligence → AI Utilities
(Image: AI Utilities Overview)
How AI Utilities Works
The configuration operates on a simple principle: Cloud vs. Local execution.
- Load: Upon opening the page, Orbit fetches your organization's active utilities, your current configuration, your global credit usage, and granular per-utility usage stats.
- Configure: Administrators toggle which utilities are allowed to consume credits (using the Cloud APIs) and set limits on how many credits they can burn.
- Execution: When a feature (like the Smart Receipt Scanner) runs, it checks this configuration. If the utility is enabled and permitted to use the Cloud, it deducts a fractional credit from your balance based on the complexity of the execution. If Cloud APIs are disabled (or if credits are exhausted), the system falls back to a free, local offline mode.
- Save: Toggles and limits are saved directly to the organization's configuration profile via the
/org/ai-utilities/settingsendpoint.
Global Controls
Credit Balance
At the top right of the page, the system displays your Credit Balance (e.g., 850.50 / 1000.00).
- Value Source: Fetched live from the
/org/usageAPI. - Meaning: Represents your remaining credits (including any top-ups) versus your total allocated limit.
Enable AI Utilities (Master Toggle)
- What it is: A global kill-switch for all cloud-based AI processing.
- When ON: Individual utilities behave according to their specific configurations.
- When OFF: All cloud communication is instantly suspended. Every utility across the platform will immediately fall back to its offline, zero-cost mode (regardless of its individual settings).
- Backend Value: Updates the root
enabledboolean in the configuration payload.
Available Utilities
To the left of the main utility list, you will see a Category Filtering Sidebar.
- Dynamic Categories: The categories (e.g.,
Orbit Papers,Service Desk,Image Generation) and their respective counts are completely dynamic. They are automatically generated based on thecategoryproperty of the active utilities returned by the Orbit backend. - Clicking on a category will instantly filter the cards on the right to only show utilities within that domain.
Each AI feature (e.g., Smart Receipt Scanner, Support Ticket Classifier) is rendered as an individual card.
(Image: Utility Details Configuration)
Utility Metadata & Badges
- Title & Description: Explains exactly what the utility does.
- Beta Badge: Indicates the underlying model is still experimental or being actively trained.
- Package Badge (Lock Icon): Indicates the utility is exclusively unlocked by your current subscription package.
- Category Badge: Shows the domain it operates in (e.g.,
FINANCE,SERVICE DESK).
Usage Analytics
To help administrators budget effectively, each card displays live statistics fetched from /org/ai-utilities/usage-stats:
- This Month: Shows raw executions and exact credits burned in the current billing cycle.
- Total Usage: Shows lifetime executions and lifetime credits burned.
- Avg Cost: The historical average credit cost per execution for your organization.
- Max Cost: The theoretical maximum limit an execution could cost (e.g., processing a massive 50-page PDF).
Per-Utility Controls
1. Enabled (Toggle)
- Purpose: Turns the utility on or off entirely.
- When ON: The feature is active for end-users.
- When OFF: The feature is completely disabled and hidden from end-users across Orbit.
2. USE AI (Toggle)
- Purpose: Dictates how the utility processes data.
- When ON: The utility communicates with premium Cloud APIs, consuming credits for higher accuracy.
- When OFF (Offline Mode): The utility relies entirely on local fallback methods (such as Tesseract OCR for document reading).
The Offline Mode Warning: If "USE AI" is disabled, a yellow warning box appears. It explicitly informs the administrator that while local fallback methods are less accurate, they won't consume credits.
3. Monthly Execution Limit (Credit Usage)
- Field Type: Number Input.
- Purpose: Acts as a financial safeguard.
- Behaviour: By setting a number (e.g.,
500), you cap how many credits this specific utility is allowed to burn in a single month. If left blank, the utility has unrestricted access to your organization's total credit pool. If the utility hits this limit, it automatically flips into Offline Mode until the next month.
Configuration Workflow
- Review Analytics: Analyze the "This Month" and "Avg Cost" metrics on your heaviest utilities to understand your burn rate.
- Set Safeguards: Input a "Monthly Execution Limit" on high-volume utilities (like Receipt Scanning) to prevent them from draining the credit pool meant for critical utilities (like Legal Document Analysis).
- Toggle Non-Essentials: For internal or low-priority workflows, toggle "USE AI" to OFF to rely on free, local Offline Mode.
- Save: Changes to limits are saved when you click off the input box (on blur), and toggles are saved instantly upon clicking.
How Configuration Is Applied
When any setting is modified, the frontend immediately fires a PUT request to /org/ai-utilities/settings with the updated JSON payload:
{
"enabled": true,
"utilities": {
"receipt_scanner": {
"enabled": true,
"use_cloud": false,
"monthly_execution_limit": 500
}
}
}
Because AI models are invoked on-demand by backend services, these configuration changes take effect immediately. Any execution triggered across the platform a second after saving will respect the new routing (Cloud vs. Local) and credit limits.
Related Features
AI Utilities configuration directly impacts the performance and cost of:
- Expense Management: Receipt OCR and line-item extraction.
- Service Desk: AI Catalog Generation and ticket routing classification.
- Documents (Papers): Smart Field data extraction and Regulatory Drift sweeps.
Best Practices
- Never leave high-volume utilities uncapped: Always set a Monthly Execution Limit on utilities exposed to end-users (like Receipt Scanners) to prevent accidental credit exhaustion.
- Use Offline Mode for testing: If you are running tests or bulk-importing historical data, toggle "USE AI" off temporarily so you don't burn credits on data you don't need high-accuracy processing for.
Troubleshooting
- "Utility is suddenly inaccurate or failing to extract data." Check the AI Utilities page. If your organization's Credit Balance hit zero, or the utility hit its Monthly Execution Limit, Orbit automatically forced it into Offline Mode. Offline fallback methods (like basic OCR) are significantly less accurate than the Cloud APIs.
- "I cannot toggle or edit limits."
Ensure you have the
org.config.aiutility.changeusagepermission. Users with onlyorg.config.aiutility.viewwill see the page in a read-only, disabled state.