``` Feature Anomaly Detection | TrackDots – AI Productivity & Time Tracking Software
Start Free Trial
Anomaly Detection

When something changes, you'll know

TrackDots runs 7 behaviour detectors across your entire team every day. When a work pattern deviates significantly from an employee's own baseline, you get flagged — before it becomes a missed deadline or a lost team member.

Anomaly Detection — Last 30 Days 24 flags raised
0
Critical
18
Warning
6
Info
0
All Clear
James Wilson
Head of Technology
Warning
Sudden hour drop Erratic start times Unusual working hours
Sarah Mitchell
Sr. Software Engineer
Warning
Sudden hour drop No-activity streak
Robert Chen
Software Engineer
All Clear
7
Behaviour detectors running daily
3
Severity levels — Critical, Warning, Info
30d
Sparkline history per employee
Personal
Baseline — flags deviate from each person's own pattern
The 7 Detectors

What TrackDots watches for — automatically

Each detector compares an employee's recent behaviour against their own historical baseline — not a generic threshold. Context-aware anomaly detection.

📉

Sudden Productivity Drop

Productivity score falls significantly below the employee's personal average over recent days. Not one bad day — a sustained shift in output quality.

Critical potential

Consistently Late Start

Employee repeatedly begins work significantly later than their own typical start pattern. Signals disengagement, personal issues, or timezone drift on remote teams.

Warning level
🏃

Unusually Short Days

Total tracked hours consistently below the employee's own baseline — not just below a company threshold. Personalised detection catches subtle changes others miss.

Warning level

No-Activity Streak

Employee has logged zero activity for multiple consecutive days. Immediate flag — may indicate health issues, resignation risk, or system error worth investigating.

Critical potential
💤

Excessive Idle Time

A significant increase in idle time as a proportion of total tracked hours. Indicates distraction, blocked progress, or waiting on dependencies.

Info level
🌙

Unusual Working Hours

Employee working before 6 AM or after 11 PM on multiple days. Distinguishes from normal late nights — flags extreme hours that indicate either overwork or role confusion.

Warning level
📊

Erratic Schedule

Start times vary by several hours across active days — no consistent work rhythm. Signals unclear expectations, timezone issues, or personal instability requiring a conversation.

Info level
Anomaly vs Burnout Detection

Two different systems for two different problems

People often ask how Anomaly Detection differs from Burnout Detection. The answer is the direction of the signal.

Burnout Detection monitors overwork — too many hours, too many late nights, declining output despite effort. Anomaly Detection monitors pattern deviation — sudden drops, inconsistency, absence patterns that are unusual for that specific person.

  • Burnout — catches people working too hard for too long
  • Anomaly — catches people whose pattern has suddenly shifted
  • Both run simultaneously — different signals, different actions
  • Combined view shows the full picture per employee
  • Available on Growth and Business plans
30-Day Activity Pattern Erratic schedule detected
James Wilson — Start time varies by 9h 23m across last 14 active days
Apr 18
9:02 AM
Apr 21
11:47 AM
Apr 22
6:24 PM
Apr 23
9:31 AM
Erratic Start Times
Start times vary by 9h 23m across last 14 active days. Consider a conversation about work schedule expectations.

Know when something shifts before it becomes a problem

Seven detectors. Personalised baselines. Daily updates. TrackDots anomaly detection runs silently so you're never caught off guard.