The shifts calendar view
The Shifts section displays a calendar showing shift activity across your fleet. Each day cell shows a count of shifts for that day, colour-coded by their collective status — green if all shifts completed cleanly, amber if some need attention, red if there are unresolved issues.
- Navigate to Shifts from the sidebar.
- Use the month/week toggle to switch between views. The monthly view is best for spotting patterns; the weekly view is better for day-to-day management.
- Click on any day to drill into that day's shift list.
Daily shift list
When you drill into a specific day, you see every shift scheduled or completed for that date. Each shift row displays:
- Driver name and vehicle plate
- Start time and end time (or "In progress" for active shifts)
- Status badge — ACTIVE, COMPLETED, or CANCELLED
- Photo review status — whether the shift's inspection photos have been reviewed
Filtering shifts
Use the filter bar at the top of the daily list to narrow results by:
- Status — ACTIVE, COMPLETED, or CANCELLED
- Driver — select a specific driver to see only their shifts
- Vehicle — select a specific vehicle to see all shifts involving it
Combine filters to answer specific questions, such as "Show me all completed shifts for Van 12 this week" — useful when investigating a damage claim.
Shift statuses
| Status | Meaning |
|---|---|
| ACTIVE | The driver has started the shift and it is currently in progress. Start-of-shift photos have been submitted. |
| COMPLETED | The driver has ended the shift and submitted end-of-shift photos. The shift is ready for dispatcher review. |
| CANCELLED | The shift was cancelled before completion — by the dispatcher, the driver, or the system (e.g., vehicle pulled from service). |
Viewing shift inspection photos
Each shift includes up to 12 photos — 6 taken at the start and 6 at the end, covering the front, rear, left side, right side, interior, and dashboard/mileage. See the six-photo inspection guide for what each angle should capture. These photos form the vehicle condition record.
- Click on a shift row to open its detail page.
- The photo grid shows start-of-shift photos on the left and end-of-shift photos on the right, paired by angle (e.g., front-start next to front-end).
- Click any photo to open it full-size. Metadata is displayed below: timestamp, GPS coordinates (if available), and the device that captured it.
The photo review workflow
Photo review is the dispatcher's core daily task, but as an owner you need to understand and oversee this process. Here is how it works:
How dispatchers review photos
After each shift is completed, its photos enter the review queue. The dispatcher opens Shifts and works through the queue, examining each photo pair to identify any new damage that appeared during the shift.
For each photo, the dispatcher selects a tag:
| Tag | When to use |
|---|---|
| OK | No visible damage or change compared to the previous inspection |
| Suspicious | Something looks off but is not conclusive — flagged for a second look |
| Damage | Clear new damage visible — this triggers the incident creation flow |
A shift cannot leave the review queue until every photo has been tagged.
AI-assisted vehicle condition assessment
When the ai_photo_analysis_enabled module is active, Fleet reviews the completed shift's inspection photos and provides advisory context:
- Review bucket — Clear, Inconclusive / retake, or Possible damage
- Confidence score — how confident the model is in the bucket; it is not a damage probability
- AI reasoning — a short explanation of what the model observed or why the evidence is inconclusive
- Photo-level context — affected photo, damage type, severity and vehicle area where the evidence supports it. Fleet deliberately does not draw bounding boxes.
AI analysis is advisory. It never clears an inspection, confirms damage, creates an incident, or blocks a workflow. If analysis is temporarily unavailable, manual review remains available and the failed attempt is not counted as a scan.
Previous shift comparison
The review interface automatically loads the matching photo from the previous shift for side-by-side comparison. This makes it straightforward to spot new damage — if the left side of the vehicle was clean yesterday and shows a scratch today, the damage occurred during the intervening shift.
Verdict system
After reviewing all photos (with or without AI assistance), the dispatcher assigns a shift-level verdict:
| Verdict | Meaning |
|---|---|
| OK | The vehicle appears unchanged and no new concern is visible. |
| DAMAGE | You identified possible new damage or a significant discrepancy. Create an incident separately when appropriate. |
| SUSPICIOUS | The evidence is inconclusive. Request a retake or physical inspection. |
Neither an AI bucket nor a dispatcher verdict automatically creates an incident or blocks the vehicle. Follow your normal incident and safety process when action is required.
Overseeing the process as an owner
While dispatchers handle the day-to-day photo reviews, you should periodically:
- Check the Inspection Compliance KPI on your dashboard to confirm drivers are submitting all six required photos
- Look at the Unreviewed shifts count in the Action Queue — a backlog here means your dispatcher team may be understaffed
- Spot-check reviewed shifts to ensure your dispatchers are being thorough, not just bulk-approving photos
- Review the AI confidence scores on flagged photos to calibrate how well the AI is performing for your specific vehicle types
If you notice the AI is consistently flagging false positives for a particular vehicle (e.g., a permanent mark being flagged every shift), document it with your dispatchers so they know to expect and override that flag.
Why inspections protect your fleet
- Pre-trip and post-trip photos create a continuous visual record of every vehicle's condition. This record is your strongest evidence during damage disputes — you can prove exactly when damage appeared and which driver had the vehicle.
- Consistent inspections change driver behavior. When drivers know every shift starts and ends with a documented vehicle check, they treat vehicles more carefully.
- For Amazon DSP operations, inspection compliance is a key scorecard metric. High compliance = fewer surprises, fewer claims, lower insurance costs over time.
Recommendations
- Review the Inspection Compliance KPI on your dashboard daily. If it drops below 90%, investigate which drivers or shifts are missing required photos.
- Spot-check 2-3 reviewed shifts per week yourself. This keeps dispatchers thorough and helps you calibrate whether your AI photo analysis settings are tuned correctly.
- If the AI consistently flags the same spot on a vehicle (e.g., a permanent scratch), document it as a known pre-existing condition so your team can override confidently.
- Use your normal incident and safety process whenever a human review identifies a concern; Fleet keeps the AI output advisory.