Overview report

The Overview report gives you a high-level picture of how Atom is performing across your organization — how many conversations are happening, how they're flowing through the system, and how much time Atom is saving your team. Use it to track adoption, measure deflection, and identify which employee segments are getting the most value.

To access this report, navigate to Analytics > Overview. Only organization admins and organization analysts can access the Overview report.

Report filters

You can scope the Overview report using three filters at the top of the page:

• Workspace: Filter to a specific workspace or view data across all workspaces.

• User segment: Filter by any employee segment configured in your organization (e.g., Employees, Engineering, Managers). Select All users to see the full picture.

• Duration: Choose a predefined time range (Today, This week, This month, This quarter, This year) or set a custom date range.

For example, to understand how Atom is performing for your Engineering team this quarter, set Workspace to your IT workspace, User segment to Engineering, and Duration to This quarter.

Understanding Overview analytics

KPI cards

The four summary cards at the top of the report give you a snapshot of Atom's activity and impact for the selected time period and filters.

|

Metric

|

What it measures

|

Why it matters

|

Total conversations

|

The number of conversations employees had with Atom, with a breakdown by channel (Web, Slack, Portal, Extension, etc.).

|

Shows overall adoption and tells you which channels employees prefer. A low number may indicate adoption gaps; the channel breakdown helps you decide where to focus enablement.

|

Requests created

|

The number of requests (service requests, incidents, and general requests) created from Atom conversations, broken down by source (Portal, API, Email, Others).

|

Tells you how much of the conversation volume is translating into formal requests. A high ratio of requests to conversations may indicate gaps in Atom's knowledge base.

|

Deflection rate

|

The percentage of qualifying conversations that Atom resolved without any request being created. Shown alongside the count of deflected conversations.

|

Your primary measure of Atom's effectiveness. A higher deflection rate means more employees are getting answers without waiting for an agent. For details on how this metric is calculated, see Improve deflection with Recommendations.

|

Hours saved

|

The estimated time Atom saved your team by resolving conversations that would otherwise have required manual handling, calculated based on your actual resolution times per topic.

|

Translates Atom's impact into a metric leadership cares about. Use this to quantify ROI and justify continued investment in knowledge and AI capabilities. For details on how this metric is calculated, see Improve deflection with Recommendations.

Conversations and requests flow

The flow chart in the middle of the report visualizes how conversations move through the system — from where they started, through whether Atom was involved, to how they were resolved. Each conversation maps to a single path through the chart, so the numbers are consistent and you won't see inflated counts from multi-message conversations.

Reading it left to right, the chart has three layers:

Entry channels

The leftmost column shows where conversations originated. The height of each band reflects relative volume. You'll see the following labels:

• Slack or Teams: Conversations initiated through the Atomicwork Slack or Microsoft Teams app, depending on which messaging platform your organization uses.

• Extension: Conversations initiated through the Atomicwork browser extension.

• Portal: Conversations started from the Atomicwork web portal.

• Email: Conversations that originated via email.

• API: Conversations created programmatically through the API.

• Workflow: Conversations triggered by an automated workflow.

• Journey: Conversations triggered by an employee journey (e.g., onboarding or offboarding flows).

Atom involvement

The middle column shows whether Atom played a role in handling the conversation:

• Atom involved: Atom actively participated in the conversation — answering questions, executing skills, surfacing knowledge, or handling access requests.

• Atom not involved: The conversation reached the system but Atom did not participate in the resolution (for example, the request was routed directly to an agent or created through a channel that bypasses Atom).

Outcomes

The rightmost column shows how each conversation ended. The Outcome nodes are expandable by clicking on the right edge, revealing a further breakdown.

• Resolved by Atom: Atom successfully handled the conversation without a request being created. Click this node to expand it and see how Atom resolved these conversations — for example, whether Atom answered from Knowledge articles or handled Small talk.

• Request created: The conversation resulted in a formal request. Click this node to expand it and see the breakdown by request type: Service request, Incident, Request, or Problem.

• Dropped: The conversation ended without a resolution or a request being created. This typically happens when Atom asked a clarifying question and the employee didn't respond, or when Atom encountered an error (in which case the employee sees an "Atom is unavailable" message).

This chart helps you quickly identify patterns. For example, if a large proportion of Portal conversations result in requests while Slack conversations are mostly resolved by Atom, that might indicate that your portal-facing knowledge base needs attention, or that the types of questions coming from each channel differ in complexity.

Usage and deflection analysis

The chart at the bottom of the report breaks down conversation volume and deflection performance by user segment.

Each bar represents a user segment and shows:

• Deflected: Conversations where Atom resolved the employee's question without a request being created.

• Not deflected: Conversations that resulted in a request or were not successfully resolved by Atom.

You can sort segments by Most usage to see which groups are engaging with Atom the most.

This view helps you spot adoption and effectiveness patterns across your organization. For example, if your Engineering segment has high usage but a lower deflection rate than other teams, that may indicate that your knowledge base doesn't cover engineering-specific topics well enough — a signal to review your content or check the Recommendations page for suggested articles to add.

Related resources

• Improve deflection with Recommendations — Understand how Deflection rate and Hours saved are calculated, and get AI-generated suggestions to improve your numbers.

• The admin's deflection guide — Best practices for improving Atom's effectiveness.