Workflow Intelligence - Analyzing Workflows

Once your apps are configured and WorkflowIDs are enabled, Workflow Intelligence provides a detailed, process-oriented view of your team's work. This article covers how to analyze workflows even before you add Key Activities: scoping a question, reading the canvas, measuring effort between any two events, reviewing individual instances, spotting inefficiencies, and generating opportunity reports.

This article covers the following:

• Asking and scoping a question
• Reading the canvas
• Exploring a workflow and measuring effort
• Reviewing individual instances
• How Workforce builds a workflow: the collection of pages
• Spotting inefficiency signals
• Watching session replay from a node
• Generating an AI opportunity report (Beta)
• Frequently Asked Questions

Asking and scoping a question

Analysis starts with a question about how your team performs, such as "How effectively is my team handling tickets in Zendesk?" You frame and scope that question using the filters in the left panel, which narrow the workflow down to the work you care about.

• Employee: focus on the work of one or more specific people.
• Activity: focus on instances where a particular activity occurred.
• Time: focus on a specific date range.

The filters in the left panel apply to both the aggregate view and the list of individual instances, so you can move between the big picture and specific examples without losing your scope.

Reading the canvas

The canvas in the center of the screen is a generalized process map. It shows how work flows from a first touch through to completion, and it is a view of how well your team adheres to the process you intend for them to follow.

Exploring a workflow and measuring effort

The Workflow Insights panel on the right measures the active working time associated with a workflow, including cross-app activity.

With Explore a Workflow, you search directly by an ID field, such as zendeskTicketId, and measure effort between any two events. You don't need to configure Key Activities first.

Choosing From and To events

Use the From and To fields in Workflow Insights to measure the time between two events. By default, the measurement runs From First Touch To Last Touch, covering the whole instance.

You can also align From and To with Key Activities to measure part of a process. For example, set From to when a case was assigned and To to when it was solved.

Interpreting Duration and Effort

For the selected From and To events, Workflow Insights shows:

• Duration: the median wall-clock time between the From and To events.
• Effort: the hands-on active time between the From and To events. The Active time field shows the average active time per instance, along with the total active time across all instances in scope.

Effort is also broken down by application, so you can see how much active time was spent in each app.

Leaders use this effort measurement to:

• Understand the opportunity size of fully automating a process, which helps prioritize where investment will have the highest economic impact.
• Measure the impact of an investment made to reduce the human effort a process requires.
• Understand the tool mix and application footprint of a process, answering "How is employee time distributed across the stack?" and revealing opportunities to consolidate work into fewer applications.

Reviewing individual instances

Beyond the aggregate view, Workforce keeps a detailed record of every individual instance of a workflow it observes. Reviewing individual instances supports several use cases:

• Root cause analysis: for example, a poor CSAT score prompts a support leader to investigate how a case was handled, understand where it went wrong, and course-correct through coaching.
• Process auditing: for example, a loan discrepancy prompts a claims manager to confirm whether a field was modified by the agent or the customer.
• Quality assurance: for example, a QA analyst reviews randomly selected tickets to measure response quality and identify coaching opportunities.
• Process observation and shadowing: for example, a process improvement leader reviews instances to research best practices and update Standard Operating Procedures (SOPs) based on actual behavior.

The Instances table

The Instances view displays all instances that match the filters selected in the left panel. All columns are sortable, and each row can be clicked into to open the individual instance.

The individual instance view

Clicking into an individual instance displays all activity for a particular WorkflowID, from the first observation to the last. The canvas populates a node each time employee activity related to the instance is captured, and any additional tracked applications the employee used during that period appear within the same node.

How Workforce builds a workflow: the collection of pages

Workforce does not work in terms of sessions the way Fullstory Analytics does. Instead, the activity it shows is based on a collection of pages assembled for each unit of work:

• Any pages from the primary application that carry the WorkflowID page property.
• Any pages from supporting applications that were accessed in the same timeframe.

Activity accumulates within a node until 60 minutes have elapsed since the employee last interacted with a page containing the WorkflowID. Additional nodes appear on the canvas whenever the same instance is accessed by a different employee, or revisited later by the same employee.

Case Swivels

A case swivel is when an employee switches to another ID of the same type of WorkflowID. For example, if a user is working on one Zendesk ticket and switches to a different Zendesk ticket, that's a case swivel.

Because a workflow is a collection of pages, you can toggle Case Swivels on or off in the left panel. Turning Case Swivels off removes pages with a different WorkflowID that were worked at the same time, so you see only the case in question. Some teams prefer to leave them on—especially in linear workflows where researching related cases is considered part of the investigation.

For details on how a case swivel affects node activity and session replay, depending on whether the employee returns to the original case within the 60-minute node window, see Workflow Intelligence - Case Swivel Behavior.

Spotting inefficiency signals

As you review workflows, several signals help you identify bottlenecks and repetitive work that may be ripe for automation.

• Copy-and-paste signals: copy-paste activity tends to happen in bursts during working periods. These "time drips" add up to meaningful inefficiency and often point to manual steps that could be automated.
• Swivel chair behavior: this shows how often an employee works multiple cases at the same time. Many SOPs limit agents to one or two concurrent cases to avoid spreading cognitive load too thin, so this signal helps you see whether swivel-chairing is slowing your team down.
• Time between key activities: shift-select two steps on the canvas to measure the time that elapses between them. This gives quick answers to questions about how long work sits between key moments in a process.
• Participants and handoffs: the number of people involved in an instance correlates with how long it takes. In one example, tickets handled by a single person averaged about 1.7 hours, while tickets touched by two or more people took considerably longer—likely due to gaps between handoffs.

Watching session replay from a node

Clicking the play button on any node opens a cross-domain session replay at that moment, right on the canvas. This is distinct from replay in Fullstory Analytics: it represents all activity in that node, providing a ground-truth recreation of what the employee did while active in the workflow, both within and across applications.

To go further, click View in Analytics. This opens the same moment in the full Fullstory Analytics experience, where you can add notes and share the session.

The replay experience also includes:

• An automatic summary of the node: plain-language text describing what the employee was doing at that step, such as "Priya Castellano navigated between two Zendesk tickets..."
• An events list, where you can see where copy-and-paste blocks and handoff delays occur.

Generating an AI opportunity report (Beta)

Workforce can generate an in-app report that identifies automation opportunities discovered by AI and Skills for the specific workflow you're viewing.

To generate a report:

1. In the middle panel, click the Reports tab, to the right of Flows and Instances.
2. Click Generate Opportunity Report (BETA).

Below the button, a table lists previously generated reports, showing the Generated date, Status, and Author for each. Use the delete control on a row to remove a report you no longer need.

Report generation runs asynchronously and typically takes a few minutes. You can navigate away from the page or close your browser while the report generates—it keeps running in the background, and the finished report appears in the UI when you return.

Each report covers the workflow you were viewing when you generated it. It estimates the hours your team could save through automation, ranks the candidate workflows by impact, and links to session replay moments that show the manual work behind each recommendation. A Methodology tab explains how the estimates are calculated, so you can verify each number.

Frequently Asked Questions

Why do I see cases that seem unrelated to the one I'm investigating?

A workflow is built from a collection of pages, which can include other cases an employee worked at the same time. Turn off Case Swivels in the left panel to remove pages with a different WorkflowID and focus only on the case in question.

Why does one workflow instance have multiple nodes?

Activity accumulates in a node until 60 minutes pass without interaction on a page containing the WorkflowID. A new node appears whenever the instance is revisited later, or when a different employee works on it.

How is this different from active time in Fullstory Analytics?

Effort in Workflow Insights measures active working time across all the applications involved in a workflow. For the underlying definition of active, focused, and visible-tab time, see Workforce Active Time.