Lone Worker
Safety check-ins and one-tap SOS for frontline workers, with a guaranteed escalation engine that climbs a responder ladder until a human acknowledges.
MangoApps
Meet the agent
This app ships with a production AI agent — permission-aware, tenant-scoped, audit-logged, and governed by the Agent Development Lifecycle.
Lone Worker AI
Timed check-ins, one-tap SOS, tiered escalation — read live, acknowledged on confirm.
Overview
Lone Worker makes safety a property of the shift app frontline workers already carry. Workers on a covered shift get timed 'I'm safe' check-ins and a one-tap SOS; a missed check-in or an SOS opens an alert that climbs a site/team-scoped responder roster via push, SMS, and voice until a responder acknowledges it — with a guaranteed dead-man's switch broadcast to safety admins if nobody does. Location is captured at trigger time only (no continuous tracking). Every state change, notification, acknowledgement, and resolution is written to an immutable, exportable audit trail, and resolving an alert as an incident hands off to Safety Hub. It depends on Shifts & Scheduling for who is on, when, and where, and reuses the platform's notification channels. Sensor-dependent features (man-down, fall detection) are intentionally deferred.
Highlights
Capabilities
Escalation & acknowledgement
-
Tiered responder ladder that climbs until a responder acknowledges
-
Mandatory acknowledgement — an alert is owned, not just sent
-
Per-tier notification channels (push / SMS / voice)
-
Per-tier acknowledgement timeout before the alert climbs (or a tenant default)
-
Dead-man's switch broadcast to safety admins when the ladder is exhausted (non-suppressible)
-
Break-glass fallback to org safety admins when no roster resolves
-
Live safety dashboard — who has an active SOS or overdue check-in right now
Check-ins & SOS
-
One-tap SOS, on or off shift
-
Timed "I'm safe" check-ins auto-armed for workers on a covered shift
-
Missed check-in automatically raises an alert — no manual monitoring
-
"Report a floor issue" — worker-raised equipment / line alert
-
Device location captured once at SOS trigger time
-
Live acknowledgement status — the worker sees the moment a responder takes their SOS
Responder rosters & routing
-
Rosters scoped to a site (Location), team, or specific shift
-
Most-specific-scope routing (shift beats team beats site)
-
Ordered escalation tiers with per-tier members and channels
-
Per-site / per-team check-in schedules (interval + grace)
Resolution, incidents & audit
-
Resolve as safe / false alarm / incident, with a required note on incidents
-
Worker self-cancel of their own alert as a false alarm
-
One-click Safety Hub incident on "incident" resolution (configurable)
-
Immutable, append-only audit trail of every state change and notification
-
CSV export of the alert audit trail for EHS / regulator review
Devices & delivery
-
Signed webhook to ingest wearable / IoT panic buttons & line sensors (RFID badge press, push-to-talk)
-
Emergency SMS that user notification preferences can't silence
-
Automated voice call-out to responders
-
One-tap "I've got this" acknowledgement from the push / in-app alert
-
Live video to the responder on acknowledgement (Live Assist) Requires the Live Assist app + a configured LiveKit provider
AI, API & mobile
-
Ask AI agent: who's at risk now, alert timeline, your status, roster coverage
-
Acknowledge an alert directly from Ask AI (responders only, confirmation-gated)
-
OAuth-scoped worker safety API (status, SOS, floor issue, answer check-in)
-
OAuth-scoped responder API — list, acknowledge, and resolve alerts
-
Mobile worker + responder surfaces
-
Admin analytics: ack rate, ack latency, outcome mix, check-in completion
Limits & Specs
-
Notification channels: Push, SMS, voice (per tier)
-
Active alert triggers: SOS, missed check-in, floor issue, wearable device
-
Check-in interval default: 60 min (5–480, admin-configurable)
-
Check-in grace default: 10 min (1–120, admin-configurable)
-
Tier ack-timeout default: 120 sec (30–1,800, admin-configurable)
-
Escalation / miss-detector cadence: Every minute (Sentry-monitored)
-
Pricing: Requires a license
Use cases
Resources
FAQ
No. Lone Worker escalates to your own responders and shows the worker whether help has been reached; it is not a PSAP and does not dial emergency services on the worker's behalf.
No. Location is captured only at trigger time — when an SOS is tapped. There is no continuous background GPS tracking, and a worker-raised floor issue skips geolocation entirely.
The alert climbs each roster tier as its acknowledgement timeout elapses. If the whole ladder is exhausted with no acknowledgement, a dead-man's switch broadcasts to your safety admins over push, email, SMS, and voice. That guarantee is deliberately not gated by any tenant toggle — an admin can't turn the break-glass off.
It's automatic. Timed "I'm safe" check-ins arm on their own for any worker on a covered shift — one where a responder roster resolves for the site, team, or shift — so there's nothing for the worker to set up. Cadence comes from a site/team check-in schedule or your tenant's defaults (60-minute interval, 10-minute grace, both admin-configurable). If a check-in is missed past its grace window, an alert auto-raises; no one has to watch a board and no worker has to remember.
Yes. A signed (HMAC) webhook accepts events from wearable/IoT devices — RFID badge presses, push-to-talk panic buttons, line sensors — and each raises an alert onto the same responder escalation, acknowledgement, and audit trail as a phone-raised SOS. Devices are attributed to a worker by badge ID, with a configurable fallback user for unbadged sensor events.
They're complementary. Lone Worker is the real-time engine — trigger (SOS, missed check-in, floor issue, wearable), escalate up the responder ladder, acknowledge, resolve — for someone who may be in danger right now. Safety Hub is where incidents live as durable records. When a responder resolves an alert as "incident" (a note is required), Lone Worker hands off and opens a Safety Hub incident automatically — typed as an equipment failure for a floor issue, otherwise a security incident. That handoff is on by default and configurable per tenant.
Optionally. When the Live Assist app is installed with a configured video provider and you turn on the "open live video on SOS acknowledgement" toggle (off by default), acknowledging an SOS opens a live video channel to the worker so the responder can see the scene. Without Live Assist configured, escalation, acknowledgement, and resolution all work exactly the same — video is an add-on, not a dependency.
Yes, both OAuth-scoped. Workers get a safety API to read their status, trigger an SOS, raise a floor issue, and answer the current check-in (native-app parity with the mobile surface). Responders get an alerts API to list, acknowledge, and resolve alerts. Both run on the same escalation and audit machinery as the in-app surfaces.