Deviation Investigation

Effective deviation management requires thorough investigation to identify root causes and prevent recurrence. This page covers how to generate AI root cause analysis and review the deviation timeline.

Generate an AI root cause analysis

Select an analysis type

Before generating, choose one of two methodologies:

Method	Best For
5 Whys	Linear cause-and-effect chains where successive “why” questions reveal the root cause.
Fishbone Diagram	Complex deviations where multiple categories of factors may contribute to the problem.

Run the analysis

1. On the left sidebar, select Deviations.
2. Select View on a deviation row to open the deviation detail page.
3. Locate the Root Cause Analysis panel.
4. From the analysis type dropdown, select 5 Whys or Fishbone Diagram.
5. Select Generate Analysis.
6. The AI processes the deviation data and produces a structured analysis.

Note

Root cause analysis generation requires appropriate permissions. If the Generate Analysis button is not visible, contact your system administrator.

Read 5 Whys results

The 5 Whys analysis presents:

• Sequential “Why” questions — each level drills deeper into the cause chain, displayed as Why 1 through Why 5 (or fewer if the root cause is reached earlier).
• Root Cause — the fundamental cause identified at the end of the chain, highlighted in a summary box.
• Recommendations — actionable suggestions for addressing the root cause, listed as bullet points.

Read Fishbone Diagram results

The Fishbone analysis organizes contributing factors into categories:

• Categories — factors grouped by type (for example, People, Process, Equipment, Materials, Environment, Management), each in its own section.
• Contributing Factors — individual factors within each category listed as bullet points.
• Root Cause — the primary root cause identified across all categories.
• Recommendations — suggested corrective and preventive actions.

Regenerate an analysis

If new information becomes available or the initial analysis needs refinement:

1. Scroll to the bottom of the existing analysis.
2. Select Regenerate Analysis.
3. The AI produces an updated analysis incorporating the latest deviation data.

Tip

Consider regenerating the analysis after updating the deviation description, adding comments with new findings, or completing an initial investigation. The AI uses all available deviation data to produce its analysis.

Review the deviation timeline

The timeline provides a chronological audit trail of all events related to a deviation, including linked CAPA events.

1. On the deviation detail page, select the Timeline tab.
2. Review events displayed in chronological order.

Note

The timeline is only visible to users with trail viewing permissions. If the timeline is not displayed, contact your administrator.

Timeline event details

Each event includes:

Element	Description
Event Icon	Visual indicator of the event type (create, update, approve, reject).
Event Title	Describes what happened (for example, “Deviation Created”, “Status Changed to approved”).
Timestamp	Exact date and time of the event.
User	Email of the user who performed the action.
Description	Additional context such as the previous status or added resolution.

Linked CAPA events

When a CAPA is associated with the deviation, CAPA-related events appear in the timeline sorted chronologically alongside deviation events. The timeline displays:

• CAPA creation events.
• CAPA status changes.
• CAPA action updates.

Above the timeline, a Linked CAPA section shows the CAPA ID and title, current CAPA status, and CAPA creation date. If no CAPA is linked, the section displays “No CAPA linked.”

Practical example: AI root cause analysis for a sterility test failure

Scenario: A sterility test failure was reported for injectable Batch INJ-2026-0089. The deviation has been placed Under Review and you need to identify the root cause.

Running a Fishbone analysis

1. On the left sidebar, select Deviations and open the deviation for Batch INJ-2026-0089.
2. In the Root Cause Analysis panel, select Fishbone Diagram from the analysis type dropdown.
3. Select Generate Analysis.

The AI generates a structured Fishbone analysis. In this example, the output might include:

• People: Operator was newly qualified on aseptic technique; gowning procedure deviation noted in training records.
• Process: Media fill qualification for the filling line was last performed 11 months ago (approaching the 12-month requalification deadline).
• Equipment: HEPA filter integrity test for isolator ISO-03 showed borderline results (0.009% vs. 0.01% limit) during last PM.
• Materials: New stopper lot (Lot #S-7782) introduced without completing extractables testing.
• Environment: Particle counts in Grade A zone were at the upper alert limit (2 of 3 readings) during the filling campaign.
• Root Cause: Combination of borderline environmental controls and introduction of a new stopper lot without complete qualification.
• Recommendations: Perform unscheduled HEPA filter integrity test on ISO-03; complete extractables testing on stopper Lot #S-7782 before further use; schedule media fill requalification.

Reviewing the timeline

1. Select the Timeline tab on the deviation detail page.
2. Verify the chronological sequence: deviation created, status changed to Under Review, root cause analysis generated.
3. After creating a CAPA from this deviation, the timeline will also show CAPA creation and action updates interleaved with deviation events.