Track MCP Client Usage for SN MCP Servers with the Agentic Usage Overview Dashboard (Action Fabric)

Ashley Snyder · 15m ago

What is the Agentic Usage Overview Dashboard?

The Agentic Usage Overview Dashboard is where platform admins go to monitor usage from inbound agentic connections to MCP servers in MCP Server Console. It contains visualizations that help you assess how Model Context Protocol (MCP) clients are calling your MCP servers, broken down by client, table, flow name, and time period, so admins can monitor usage trends and report on activity over time.

When an external AI agent connects to your ServiceNow MCP server, it can read and write records, trigger flows, and invoke tools across your instance. The Agentic Usage Overview Dashboard is where admins go to see what is actually happening: which clients are active, which ServiceNow tables and flows they are touching, and how that activity trends over time.

This article explains what each part of the dashboard shows and how to use it for usage monitoring and reporting.

In this article: Why this dashboard exists Record Access tab Flow Execution tab Reading the filters Using this for reporting Getting started

Why this dashboard exists

As organizations connect external AI agents to ServiceNow through MCP, platform admins need a way to see what those agents are actually doing. The Agentic Usage Overview Dashboard provides that visibility. It surfaces two categories of activity: records that MCP clients read or wrote, and flows that MCP clients triggered.

Both categories are tracked separately because they represent different types of platform activity: direct data operations versus orchestrated multi-step processes. Monitoring both gives admins a complete view of how external agents are interacting with the instance.

What the dashboard tracks

The dashboard reflects all MCP client activity on your instance. Use the MCP Client filter to isolate activity from specific external agents, or view aggregate totals across all connected clients.

Record Access tab

The Record Access tab shows every ServiceNow record that was read or written by an MCP client. A record access event occurs when an MCP tool reads, creates, or updates a record in a ServiceNow table.

What the tab shows

Daily: a bar chart showing the number of records accessed for each day over the last 30 days, across all MCP clients and tables. Use this to spot spikes in activity and correlate traffic patterns against integrations you have enabled. Selecting a bar drills down to the records accessed, with any currently applied filters retained.

Monthly: a bar chart showing the number of records accessed for each month over the last 13 months — the right view for tracking whether usage is growing, plateauing, or concentrated in specific months.

By MCP Client (last 30 days): a trend line of records accessed by MCP client. Each line corresponds to the client name registered in your MCP Server Console OAuth configuration — for example, a client named "claude" represents connections from a Claude agent configured with your instance credentials. Use this to see whether one client accounts for a disproportionate share of activity.

By Table (last 30 days): a trend line of records accessed by table. Use this to understand which data objects are being queried most heavily — for example, whether agents are primarily reading incident records, task records, or CMDB environment data — helping you assess both usage patterns and data exposure scope.

Agentic Usage Overview Dashboard (1).png

Agentic Usage Overview Dashboard Record Access Tab

Key Takeaway: The Record Access tab tells you which external agents accessed which ServiceNow tables, and how much. Together, the daily/monthly totals, client breakdown, and table breakdown let you move from macro trend to attribution to scope.

Flow Execution tab

The Flow Execution tab shows the number of flows executed through MCP servers. A flow execution event occurs when an MCP tool invokes a ServiceNow subflow or action on behalf of an external agent — tracked separately from record access because it represents orchestrated multi-step processes rather than individual record operations.

What the tab shows

Daily: a bar chart showing the number of flows executed for each day over the last 30 days. Selecting a bar drills down to the flows executed, with any currently applied filters retained.

Monthly: a bar chart showing the number of flows executed for each month over the last 13 months — use this for monthly reporting and to track whether agent-triggered workflow volume is growing over time.

By MCP Client (last 30 days): a trend line of flows executed by MCP client. Use this alongside the Record Access by Client view to build a complete picture of what each external agent is doing on your instance.

CleanShot 2026-08-14 at 11.36.08.png

Agentic Usage Overview Dashboard Flow Execution Tab

Key Takeaway: The Flow Execution tab complements Record Access. Together they cover the two main categories of Action Fabric activity: data operations (records) and process operations (flows).

Reading the filters

Both tabs have filters that scope every chart on that tab simultaneously. Selecting a filter value refreshes all charts to show only data matching that selection.

Filter by MCP Client

The MCP Client dropdown lists every OAuth client registered on your instance that has generated at least one usage event. Selecting a client scopes all charts to that client only. Leave it set to ALL to see aggregate totals across all clients.

For usage attribution: use this filter to isolate one client at a time and compare their record access and flow execution volumes. If your organization has multiple teams using different AI agents, each with its own registered OAuth client, this is how you attribute activity per team or integration.

Filter by Table (Record Access tab only)

On the Record Access tab, a second filter lets you select one or more ServiceNow tables. The available options are populated from tables that have actually been accessed; you will only see tables your MCP clients have touched. Selecting a table scopes all Record Access charts to events involving that table.

Example: If you want to know how often agents are reading incident records specifically, select "Incident" from the Table filter. The daily and monthly charts will then reflect only incident-related access events.

Filter by Flow Name (Flow Execution tab only)

On the Flow Execution tab, a second filter lets you select a specific flow by its system name. Like the Table filter, the options are populated from flows that have actually been triggered by MCP clients. Use this to understand how often a specific subflow or action is being invoked by external agents, and by which client.

Using this for reporting

The dashboard is designed for ongoing monitoring and periodic reporting on MCP client activity. The most useful reporting workflows combine both tabs and the client filter.

Recommended approach for monthly reporting:

1

Read the monthly Record Access and Flow Execution totals

Each monthly chart always shows the full last 13 months. Locate the bar for your reporting month on each tab and note its total, these represent the volume of activity your MCP clients generated that month.

2

Break down by MCP client.

Use the MCP Client filter on each tab to view totals per client. Record these per-client figures to attribute activity to specific integrations or teams.

3

Identify any unexpected spikes or new clients.

Review the daily charts for activity spikes and check the MCP Client filter for any client names you do not recognize. Unexpected clients may indicate a misconfiguration or an integration that has not been formally reviewed.

Tip

Use this dashboard for trend analysis and client-level activity reporting. Cross-reference with the MCP Server Console Monitoring Dashboard when you need a precise, point-in-time count.

Who this affects: Platform Admin Compliance & Security

Getting started

The Agentic Usage Overview Dashboard requires the Agentic Usage Overview Dashboard application (sn_agentic_usage) and is available on Australia Patch 6 (AP6) or later. Once the application is active, navigate to All > Agentic Usage > Agentic Usage Overview Dashboard to open it.

Forward-looking: support for Zurich Patch 13 (ZP13) and BP1 is planned as of the October 8 release. This information is subject to change, is provided for planning purposes only, and is not a commitment, promise, or legal obligation to deliver any functionality.

Detail	Value
Application	Agentic Usage Overview Dashboard (`sn_agentic_usage`)
Minimum release (current)	Australia Patch 6 (AP6)
Planned support (as of Oct 8 release)	Zurich Patch 13 (ZP13), BP1
Navigation	All > Agentic Usage > Agentic Usage Overview Dashboard
Required role	`sn_agentic_usage.user` or `sn_agentic_usage.admin`
Data populated from	MCP Server Console usage events (record access and flow execution)

For more on MCP Server Console, including setup, authentication, and tool configuration, see the MCP Server Console product documentation →