Skip to content

Sessions

The Sessions view is the starting point for investigating agentic activity in your application. It lists every detected agentic session and lets you open a session drawer to understand the agent, its environment, and the path it took through your application.

This page explains the available session data and how Agent Intelligence turns collected signals into session-level data. For integration guidance, see Integration options.

Key concepts

The following terms describe how Agent Intelligence groups and classifies activity in the Sessions view.

TermDefinition
Agent platformA specific tool that executes actions on a user's behalf, such as ChatGPT Agent, Manus AI, or Operator.
Agent categoryThe class an agent platform belongs to, assigned automatically by Mosaic. Categories are not customer-configurable and are extended as new platform types emerge.

Current categories include:
• AI-native agentic browser (examples: Browserbase, Dia, Arc)
• AI-augmented mainstream browser (examples: Chrome with Gemini, Edge with Copilot)
• OS-level / Computer-use agent (examples: Claude Computer Use, Operator, Devin)
• Headless / developer framework (examples: Playwright, Puppeteer, Selenium, LangChain)
• Browser extension agent (examples: Riskified, HubSpot, Grammarly)
Browsing sessionA continuous span of activity from a single agent or user, bounded by a period of inactivity. This is the unit Agent Intelligence uses to group actions and pages into one visit.
UnclassifiedSessions confidently detected as agentic but not yet attributed to a named agent platform.

How session data is produced

At a high level, Agent Intelligence follows this flow:

  • The Platform SDK is loaded into the application, either directly in the app or through an edge/CDN integration.
  • As the application runs in the browser, the SDK captures browser and action signals in the background.
  • Those signals are sent to Mosaic for backend processing.
  • Mosaic correlates the activity, analyzes agentic patterns, and turns the processed results into session-level data.
  • Agent Intelligence then exposes that data as session volume, per-session summary fields, and deeper session and path details.

Session data

Agent Intelligence exposes session data at three levels: aggregate volume, per-session summary fields, and additional session and path details for each individual session.

Agent Intelligence Sessions experience in Mosaic
Click to open the GIF in a dedicated tab.

Detected vs unclassified sessions

The aggregated session data is available over the selected date range and is split into two series:

  • Detected platforms — sessions attributed to a named agent platform.
  • Unclassified — sessions confidently detected as agentic but not yet attributed to a named agent platform.

Comparing these two series shows both how much agentic traffic your application receives and how much of it Mosaic can already attribute to a known platform.

Session list

For each session, Agent Intelligence provides the following summary fields:

ColumnDescription
Session IDUnique identifier of the session.
Activity SurfaceApplication surface where the session was detected, such as login or /.
Start TimeWhen the session started.
Agent PlatformAgent platform detected for the session, such as ChatGPT Agent or Manus AI. Sessions that couldn't be attributed appear as Unclassified.
IP AddressIP address observed for the session (IPv4 or IPv6).
CountryCountry resolved from the IP address.
OSOperating system reported for the session.
BrowserBrowser or runtime used by the agent, such as Chrome or Electron.
Origin TypeWhere the session ran: Cloud/Hosted for agents running on hosted infrastructure, End-User Device for agents running on the user's own device.
AnonymizationAnonymization technique detected on the connection. Shows when none is detected.
Self-DeclaredVerified when the agent declared its identity and Mosaic confirmed the declaration, Observed when the session was classified from observed signals only.

Session drawer

Select a session to open its drawer. The drawer brings together the information needed to move from "an agent was here" to what happened during that visit.

Overview tab

The Overview tab gives you the essential context for the session.

  • Session summary provides an AI-generated account of the session. It is available after a session ends; while a session is still active, return to the Sessions list to follow its activity.
  • General info shows the session's landing page, its actor type (Agent, Human, or both), origin type, and whether the agent identity was self-declared and verified or detected from observed signals.
  • Device describes the device or browser environment and includes identifiers such as the device key ID, device ID, and device fingerprint. Together, these identifiers help correlate activity to a device across sessions; they should be considered as a set rather than in isolation. For more background, see Device identity.
  • Network provides the connection context, such as IP address, location, and detected anonymization or hosting signals. Use it to understand where the activity originated and whether the connection may be masking its source.

Journey path tab

The Journey path tab shows the pages the agent navigated during the session, in order. For each page, it records the observed interaction type, such as page load (opening a page), focus (bringing a page or browser window to the foreground), blur (moving away from it), or click. Use this tab to reconstruct how the agent moved through a flow and to see whether a human also took part in it.

Insights tab

The Insights tab summarizes the session in plain language. It highlights relevant observations about:

  • Device & Browser profile
  • Security concerns
  • Human involvement
  • Agent journey
  • Activity breakdown, which gives a short account of the session as a whole

Use these insights as a quick starting point, then validate the relevant detail in the other tabs.

Agent platform tab

The Agent platform tab describes the detected agent behind the session. It provides quick facts such as the developer, agent type, and where it runs, followed by concise descriptions of what the agent can do and how it typically behaves. It also clarifies whose activity you are reviewing, so you can distinguish activity performed by the agent from activity associated with the end user.

Next steps

Choose the integration that best fits your project. Options include: