Admin: Monitor AI Usage and Credits
Background
The Reporting tab within the Control Tower area of Datagrid helps account administrators monitor agent usage and credit consumption. Whether you’re a Datagrid customer or a Procore AI (Digital Coworker) customer, these reports provide account-level visibility into user credit consumption trends, data storage credit costs, agent usage, and user engagement metrics.
This tutorial helps you access and make use of your account’s usage data, which can help you set optional guardrails in the Budgeting tab or make account-level decisions about AI expansion and usage goals.
Things to Consider
Required User Permissions:
‘Admin’ permissions in Datagrid are required for accessing the Control Tower features.
Procore’s ‘Company Admin’ role maps to the ‘Admin’ role in Datagrid, while Procore’s ‘Project Admin’ role maps to ‘Member’ in Datagrid. Therefore, ‘Project Admin’ cannot automatically access Control Tower features in Datagrid, such as these budgeting guardrails.
Procore AI customers:
All tiers of the Digital Coworker packages (Starter, Pro, and Enterprise) have access to Control Tower, including budgeting and reports.
Prerequisites: Since the Control Tower features are currently only accessible via the standalone Datagrid product and not within Procore, Company Admin must perform the credentialing sequence detailed in Enable Procore AI before they can access Control Tower features in Datagrid.
Steps
Log into
app.datagrid.ai
Note: Procore customers with an active Digital Coworker Procore AI subscription should log into Datagrid via Procore SSO.Make sure you're in the correct teamspace. If your company has more than one teamspace, select the one you want from the dropdown at the top of the left navigation panel.
In the bottom-left of the landing page, click on Control Tower.
Click on the Reports tab.
Locate the date range dropdown filter in the upper-right corner, and adjust the view parameters as needed: Last 7 days, Last 30 days, Last 60 days, or Last 90 days.
Analyze the data according to your goals. Explore below for more information about our report types.
Credit Consumption by Category
The Credit Consumption by Category bar chart visualizes credit consumption by macro-level operation: Intelligence and Generation, Data Ingestion and Setup, Activity and Infrastructure, and Data Storage.
Credit Category | Operations Monitored |
|---|---|
Intelligence and Generation | Real-time chat execution, multi-step agent reasoning, and text payload generation. |
Data Ingestion and Setup | Initial file processing, document parsing, and vector rendering of historical project files. |
Activity and Infrastructure | Backend webhooks, connection upkeep, and active orchestration traffic. |
Multi-modal Indexing | The ongoing flat-rate daily ledger charge calculated at 3 credits per GB/day. |
Global Metrics
The Global Metrics section has cards monitoring the amount of chats, teamspaces, agents, and users. These numbers give you a broad view of your AI ecosystem landscape and adoption.
In the corner of some cards is a red or green number showing the net changes for the chosen date range. For example: With the date range filter set to ‘7 days’ in the top right of the Reports window, you may see that the total AI Agents for the account is 106, with a green +10. This means 10 new agents were created in the past 7 days, bringing the total number to 106. If you change the range to ‘30 days’, you may see a higher green number, but the total will remain 106.
Total chats: Shows the cumulative volume of real-time conversational prompts executed by your teams.
Each time a user types a message into the prompt composer and presses ENTER or clicks the send arrow, this counter increases by one. For Procore AI users: Interactions in Datagrid and Procore AI side panel count towards this datapoint.
Teamspaces: Displays the total number of teamspaces (projects) actively mapped and running on the account.
AI Agents: Tracks the total number of pre-built or custom agent templates running across your projects.
Active users: The number of unique team members using the AI tools across the account.
Activity Ranking
In the Activity Ranking section, leaderboards help you pinpoint where AI adoption is highest among the top three teamspaces (projects), agents, and users across the account. At this time, you can only see usage totals for the top three users in each area, not for all users.
Most Active Teamspaces: Ranks your project workspaces by overall chat volume. Use this card to see which projects are relying most heavily on real-time agent utility.
Most Used Agents: Identifies which specific agent templates (e.g., Submittal Reviewer, Deep Search Agent) are running the most background queries.
Most Engaged Users: Top three users of the tool, along with their total chats and how much this quantity is higher or lower than it was during the previous period.
Next to many of the totals is a red or green number showing the net changes for the chosen date range. For example: With the date range set to ‘7 days’ in the top right of the Reports window, you could see a user with 72 chats and a green +20 next to it. This means the user has sent 72 chats to the AI in total, and 20 of them were in the last seven days. If you change the range to ‘30 days’, you may see a higher green number, but the total will remain 72.
Usage Breakdown
For a more granular audit of your ledger, scroll to the Usage Breakdown section and select one of the following dataset types. For each of these reports, you can sort by a particular column to identify patterns.
For example, sort by Agent Name in the ‘User Prompts & Activity’ report to identify which specialized automation tools are receiving the highest user adoption, allowing you to pinpoint high-performance project workflows and standardize successful enablement strategies across other active teamspaces.
User Prompts & Activity
This report captures the actual text used each time a user interacted with the AI in any teamspace (project) on the account. These raw user queries along with back-end semantic processing and keywords can show trends across your organization or support compliance auditing.
Choose User Prompts & Activity from the dropdown in the ‘Usage Breakdown’ section.
Scroll to view the report’s data, which includes:
Date: Calendar date of the user interaction.
User: Name of the person who submitted the prompt.
Agent Name: Name of the provided or custom agent called during the interaction. Blank fields indicate the standard platform assistant.
Teamspace: The specific project where the active session occurred.
User Prompt: The raw, unfiltered text string entered into the chat by the end user.
Task: The classification label assigned by the system orchestration layer (e.g., action-planning, semantic-retrieval, sql-retrieval) to route the request based on detected user intent.
Keywords: Core search terms and nouns isolated from the user's input by the system to look up files within the project database.
Query Rewritten: The semantic search string generated by the AI’s reasoning model to maximize retrieval accuracy from project data.
Optional: Click the Export button so the report downloads to your device and you can view the data in full, use it for external analysis, or create presentations.
Users Credit Consumption
This report shows an itemized transaction log of every fraction of a credit consumed by user activity along with the user name, system action, and more.
Choose Users Credit Consumption from the dropdown in the ‘Usage Breakdown’ section.
Scroll to view the itemized transaction log tracking every fraction of a credit consumed by user activity. Review the columns to audit specific items:
Credits: Exact credit cost deducted (displayed as a negative decimal value).
User: Team member who initiated the system call.
Action: What the system was doing (e.g., Agent Interaction, Run Import Pipeline, Dataset Processing).
Resource Type: Technical asset layer used to perform that action (e.g., Agent, Data Pipeline, Dataset, Knowledge).
Resource: Specific connector or asset name used (e.g., Procore Connector, Site Safety Agent).
Optional: Click the Export button so the report downloads to your device and you can view the data in full, use it for external analysis, or create presentations.
Teamspaces Credit Consumption
This report maps infrastructure costs directly to specific project environments. It exposes the automated Dataset Storage line items, allowing you to track the cost of the data volume each teamspace (project) is maintaining in its active vector index.
Choose Teamspaces Credit Consumption from the dropdown in the ‘Usage Breakdown’ section.
Scroll to view ongoing infrastructure costs mapped directly to specific project environments. The columns include:
Credits: Shows the exact credit cost deducted (displayed as a negative decimal value).
Teamspace: Identifies the team member who initiated the action.
Action: Identifies what the system was doing (e.g., Agent Learning (Vectorization), Agent Interaction, Dataset Processing).
Resource Type: The technical asset layer used to perform that action (e.g., Agent, Data Pipeline, Dataset, Knowledge).
Resource: The application that did the task.
Date: Calendar date the action took place.
Optional: Click the Export button so the report downloads to your device and you can view the data in full, use it for external analysis, or create presentations.
See Also
Loading related articles...