AI Spend Limits

Why spend governance matters

Without explicit spend controls, AI usage can grow in ways that are hard to predict or attribute. LLM calls are billed by tokens and model tier; many users, personas, and projects can run in parallel. That often leads to surprise invoices, pilots that exceed their intended budget, and no clear line between “allowed” and “too much” usage for a given team, user, or model.

AI Spend Rules addresses that gap by letting administrators define who or what is subject to a limit (user, group, project, persona, or model), over which period spend is measured (day, week, or month), how much is allowed, and what happens when the limit is reached (notify or block). The platform tracks current spend against those rules so organizations can enforce budgets, get early warnings, and keep AI consumption aligned with policy.

How it works

Administrators open AI Administration → AI Spend Rules, review existing rules in the grid, and use Add (or Add Default for a baseline policy) to create a new rule. Each rule ties a subject (selected after choosing subject type) to a duration, spend limit in dollars, and an action. The system accumulates spend for that subject during the active window; when the limit is reached, it either notifies (usage can continue) or blocks further AI usage for that subject until the cycle resets or the rule is adjusted.

---

1. Spend Rules Overview

The AI Spend Rules page displays all configured spend rules in a tabular view. Each row represents one spend policy applied to a specific subject.

Grid columns

Column	Description
Rule ID	Unique identifier of the spend rule.
Subject Type	Scope of the rule, such as User, Project, Persona, or AI Model.
Subject ID	The specific subject selected for the rule.
Duration	The time window for spend tracking: Day, Week, or Month.
Spend Limit ($)	Maximum allowed spend for the selected period.
Current Spend ($)	Current accumulated spend for that subject in the active period.
Action	The action triggered when the limit is reached: Notify or Block.
Block Clearance	How a blocked state is cleared. The value is Automatic.
Status	Shows whether the rule is active.
Created At	Timestamp when the rule was created.
Updated At	Timestamp when the rule was last modified.

2. Add Spend Rule

The Add action opens the Add Spend Rule form, where administrators define a new spend control policy.

Fields in the Add Spend Rule form

Field	Description
Limit Rule ID	A unique ID or name for the spend rule.
Subject Type	Defines the scope where the rule will apply.
Subject Selection	The actual user, group, project, persona, or AI model selected for the rule.
Duration	The billing/control window for the rule: Day, Week, or Month.
Spend Limit Amount ($)	The maximum spend allowed during the selected duration.
Action When Spend Limit is Reached	Determines whether the system should Notify or Block once the limit is crossed.

3. Subject Type

Subject Type is the most important field because it defines where the spend rule will be applied.

The available values are:

Subject Type	Description
User	Applies the spend rule to a single user account.
User Group	Applies the rule to all users in a selected group.
Project	Applies the rule to a specific AI or business project.
Persona	Applies the rule to a selected AI persona or copilot configuration.
AI Model	Applies the rule to one or more selected LLMs or AI models.

Subject selection behavior

After choosing the Subject Type, the administrator uses the + selector to choose the target subject.

For example:

• If User is selected, the system opens a Users selection grid.
• The Users selector displays fields such as User ID, First Name, Last Name, Email ID, User Role, User Group, Status, and Local User.

Other subject types (Persona, Project, AI Model) use the same selection pattern.

This design ensures that the rule is tied to an exact object instead of relying on free-text entry.

4. Duration

The Duration field defines the time period over which spend is calculated.

The options are:

• Day
• Week
• Month

This means the spend limit can be enforced on a daily, weekly, or monthly basis. At the end of each duration cycle, the spend tracking window resets and a new cycle begins.

5. Actions: Notify or Block

AI Spend Rules are enforced automatically when the defined spend threshold is reached. Depending on the configured action, the system either alerts the user or restricts further usage.

Enforcement applies wherever usage counts toward the rule—for example Fabrix.ai Copilot, automated workflows, and invocations of external agents—not only a single interface or entry point.

Notify

When the spend limit is reached and the rule is set to Notify, usage continues, but the user is informed that the threshold has been reached or exceeded.

• AI usage continues without interruption.
• A warning is shown indicating that the spend limit has been reached or exceeded.
• The message includes usage details and guidance to contact an administrator to adjust limits if needed.
• Use this mode to monitor trends, get early warnings before spend becomes critical, and keep flexibility for pilots and teams while staying visible to governance.

Block

When the spend limit is reached and the rule is set to Block, the system prevents further AI usage for the selected subject until usage can resume under the rule. That includes workflow steps and external agent calls when those flows incur usage against the governed subject.

• AI requests are not processed after the limit is reached, including requests from workflows that invoke external agents when they count toward the same subject.
• A blocking message is displayed stating that the spend limit has been exceeded.
• The message includes reset timing, rule reference, and when the user can try again after the duration cycle resets.
• Access is restored when the duration cycle resets (Day, Week, or Month), or when an administrator updates, raises, or removes the spend rule.
• Use this mode for strict cost control, preventing budget overruns on expensive models or high-volume personas, and enforcing governance where usage must stop at the limit.

6. Default Spend Rules

The page includes an Add Default option.

Use it to create a default spend rule that serves as a baseline policy when a subject does not already have a more specific rule.

Typical use of a default rule:

• Apply a standard daily or monthly spend policy across the environment
• Establish a minimum governance layer before custom rules are added
• Create fallback controls for new users or newly created AI entities

7. How to Create a Spend Rule

1. Open AI Administration.
2. Navigate to AI Spend Rules.
3. Click Add.
4. Enter a Limit Rule ID.
5. Select a Subject Type.
6. Use the + selector to choose the relevant subject.
7. Select the Duration as Day, Week, or Month.
8. Enter the Spend Limit Amount ($).
9. Choose the action: Notify or Block.
10. Click Save.

8. Example Use Cases

Scenario	Description
User-level limit	Set a daily spend limit for an individual user to monitor trial or pilot usage.
User group limit	Apply a shared spending rule to a specific team or access group.
Project-level limit	Control the total AI budget consumed by a specific project.
Persona-level limit	Restrict how much a given persona can consume within a day, week, or month.
AI model limit	Cap usage for higher-cost models such as premium LLMs to prevent unexpected cost spikes.

9. Business Value

AI Spend Rules provides several operational benefits:

• Cost visibility through current spend tracking
• Budget enforcement across multiple AI scopes
• Policy-driven governance for personas, users, projects, and models
• Operational safety by preventing uncontrolled AI usage
• Administrative control with clear audit timestamps and active status tracking

10. Summary

AI Spend Rules enables administrators to define spending thresholds by user, user group, project, persona, or AI model, with configurable duration windows and automated actions such as notification or usage blocking.