The problem: burnout noticed too late
Status meetings look green. Boards fill with Done. Yet Slack stays active late, PTO goes unused, and "just this sprint" repeats. Burnout often stays unspoken until resignation, medical leave, or quality collapse.
Managers want early warning. Vendors promise "AI burnout detection." Responsible tools offer risk signals from workload and time patterns — then require human conversation, not algorithmic labels on people.
What AI can and cannot do here
Can help:
- Surface chronic overload via
workload_heatmap(remaining hours, open task count) - Show sustained high
time_summarylogged hours vs estimates - Highlight
flow_agingon items stuck on the same overloaded assignees - Cross-project
portfolio_overviewfor matrixed "always on" people
Cannot do:
- Clinically diagnose burnout or mental health conditions
- Replace 1:1 trust and HR processes
- Score individuals for punishment in shared channels
- Read sentiment from private life outside work systems
WKFGo is a project management platform, not a wellness surveillance product. Frame AI as pattern detection for managers who already care.
Framework: signals → review → act → recover
1. Signals (system) — Weekly scan: persistent red heatmap cells, same names on every critical path, rising logged hours in time_summary, missed PTO patterns visible in calendar integration where used.
2. Review (manager) — Private 1:1: blockers, scope pressure, skill gaps, personal context AI does not see.
3. Act (team) — Rebalance tasks (assign_task), WIP limits, scope cut, temporary deadline relief — logged decisions.
4. Recover (org) — Cool-down sprint after crunch; retro on why overtime happened; adjust estimates.
| Signal | Possible meaning | Wrong reaction |
|---|---|---|
| Red heatmap 3+ weeks | Structural overload | Public callout |
| High time_summary hours | Underestimation or hero culture | Surveillance memo |
| Same owner every critical task | Key-person risk | More critical tasks |
| Thin review before release | Rush from overload | Blame individual |
Where WKFGo fits: honest early warning
workload_heatmap
See who carries too much open work before deadlines slip. Pair with policy: red cell → no new assign without trade-off.
time_summary
Compare estimated vs logged time for calibration, not ranking. Sustained logged hours above sustainable capacity (after subtracting meetings) warrant conversation — not automated "burnout score."
flow_aging and reports
Stuck work on overloaded people may indicate context-switching, not laziness. get_project_report and aging views feed weekly manager digest — steering, not surveillance.
MCP synthesis (careful prompts)
"List users with red workload_heatmap cells for two consecutive weeks on Project Atlas — task IDs only, no commentary on character."
Manager interprets; HR involved when appropriate.
Breaking the burnout cycle
Week one: critical task slips — volunteer overtime.
Week two: same volunteer default owner.
Week three: review quality drops.
Week four: hotfix overtime — "just this once."
Without heatmap the cycle stays invisible until exit interview.
Break it: red-cell policy, recovery sprint, praise sustainable pace — not only rescue stories.
Getting started this week
Add five-minute heatmap scan to weekly manager sync. Flag anyone red two weeks running — private 1:1 within 48 hours, not channel callout. Log rebalance actions on tasks. After major release, schedule cool-down sprint with real capacity. Track whether time_summary hours normalize — learning metric, not surveillance.
Document team policy: what happens when heatmap is red (no new work without drop or defer). Enforcement matters more than tooling.
Anti-patterns
- "Burnout score" leaderboard from AI
- Alerting executives on individual names without manager context
- Using signals for layoff justification alone
- Ignoring signals because "they're a hero"
Burnout signals are lagging—act on workload and sustained overtime patterns, not pop psychology scores. Combine workload_heatmap, time_summary, and aging on owned tasks. Follow up with humans; WKFGo does not diagnose medical conditions.
Sustained weekend work logged across three sprints is a staffing signal, not a badge of honor. Discuss load in 1:1 before retrospective blame games.
Run a monthly audit: do AI answers cite task IDs from MCP? If not, tighten prompts to require citations before any stakeholder-facing draft leaves the team.
FAQ
Does WKFGo have dedicated burnout ML?
No dedicated sentiment or burnout ML product — use workload, time, and flow signals plus human review.
Is time tracking required?
Heatmaps work with remaining estimates; time_summary improves overtime pattern visibility when teams log honestly.
Can AI recommend PTO?
AI can note load patterns; PTO decisions stay with people and managers.
How is this related to sentiment analysis?
Sentiment from comments/forms is a separate, weaker signal — see our AI team sentiment analysis piece; never substitute for conversation.
Flag overload early; lead with humans
Use workload_heatmap, time_summary, and flow reports to notice burnout risk patterns — then act with rebalance, scope honesty, and recovery time.
Try it now
Put these patterns on live project data—not slide decks.