Vygard — UK Lone Worker Safety Platform with AI Dispatcher

So no-one works alone. SOS, man-down, check-ins, geofencing, escalation, audit and CQC/HSE/RIDDOR-ready reporting in one console — with Howie, the AI dispatcher.

Vygard is a UK-built, AI-augmented lone-worker safety platform purpose-built for organisations whose workforce operates solo — care providers, construction crews, delivery drivers, security officers, field engineers, agricultural workers, oil & gas operators and many more. The platform combines real-time SOS, man-down, fall detection, geofencing, welfare check-ins, escalation and a tamper-evident audit trail in one tenant. "Howie" — the AI watcher inside Vygard, powered by Claude Opus 4.8 — drafts the dispatcher's situation summary in under a second, surfaces vital signs anomalies, proactively pings workers before they have to ask for help, and forces a mandatory three-step acknowledgement before an incident can be closed. Live in 10 working days, no hardware required, BS 8484-aligned, UK-hosted, with 12 vertical sector packs from care to ATEX oil & gas.

Key features

• AI Dispatcher Copilot — "Howie". Powered by Claude Opus 4.8 with adaptive thinking. Howie composes a sub-second SOS briefing — situation summary, next action and priority call list — inside the dispatcher's SOS modal, then forces a mandatory three-step acknowledgement before the incident can be closed.
• SOS, Man-Down & Fall Detection. Background accelerometer-tuned fall detection with auto-SOS escalation, panic-button SOS on phones, watches and BLE pendants, 30-second confirmation windows, and sector-specific patterns (chainsaw-cut on Apple Watch, quad-bike rollover via 4-second tilt-threshold).
• Welfare-Ping Engine. Eight trigger reasons — vitals drift, livestock zone, weather alert and more — drive proactive check-ins. "Asks workers if they're OK before they have to ask". Scans worker state every 60 seconds.
• 12 Vertical Sector Packs. Adult social care, construction, logistics & driver fleet, security, mountain rescue, forestry, refuse, elderly care, oil & gas (ATEX Zone 1/2), agriculture, NHS community nursing, field engineers. Custom packs in ~6 weeks.
• Geofencing & Risk Overlays. Weather and crime polygons that auto-flag workers inside alert zones. Heat Stress Index (welfare-at-70, evac-at-85) for petrochemical. Permit-to-work register and Working Time Directive monitoring built in.
• Vitals, PPE & Cert Gap Detection. Continuous heart-rate and SpO₂ streaming where the device supports it. PPE compliance checks via selfie gate at sign-in. Certification expiry detection (ppe_overdue / cert_expiring) with sector-aware severity — petro vs retail differ.
• Intrinsically-Safe & Satellite-Fallback. The only lone-worker platform running on intrinsically-safe ATEX handsets (Bartec, Ecom, i.safe, Aegex) for Zone 1/2 work. Iridium satellite-fallback SOS when GSM drops — essential for remote agriculture, forestry and mountain rescue.
• BYOD-Friendly with No MDM. One-tap PWA — workers open a URL, no app-store install, no MDM enrolment. Full uninstall on offboarding. Native iOS, Android, watchOS and Wear OS apps for richer experiences. BLE pendants and rugged handsets (Cat S62, Sonim XP10, CrossCall) supported.
• Mutual-Aid Peer Network. A nearest-neighbour buddy network routes an SOS to the closest available peer responder as well as the control room — in rural work the nearest colleague is often ~4 km away, against a 20-minute-plus rural 999 response. Mutual-aid escalation runs alongside the formal ARC / 999 path, never instead of it.
• Regulator-Ready Audit & Reporting. RIDDOR auto-PDF export, BS 8484-aligned provision, CQC Reg 13 alignment for care, HSE / DVSA / SIA / MREW / FISA 302 alignment per sector. Tamper-evident audit trail on every incident. UK-hosted in eu-west-2.

Why lone-worker safety is a workflow problem first, a hardware problem second

Most lone-worker incidents that end badly do so for one of three reasons: the alert never fired (worker incapacitated before they could hit SOS), the alert fired but landed somewhere nobody was looking (a generic inbox at 02:00), or the alert fired and was acknowledged but nobody actioned the next step ("saw it, assumed someone else had it"). The hardware problem is solvable — phones, watches, pendants and rugged handsets all exist — but the workflow problem is where deployments succeed or fail.

Vygard's design is built around closing those three gaps. The welfare-ping engine catches workers who can't trigger themselves. The mandatory three-step dispatcher acknowledgement (read, actioned, welfare confirmed) is tamper-evident, so the audit trail proves the workflow ran. And Howie, the AI copilot, drafts the next action in sub-second so the dispatcher isn't reading the alert cold at 02:00 — which is exactly when the existing system fails.

Frequently asked questions

How is Vygard different from older lone-worker apps and pager-only setups?

Older lone-worker apps are essentially a panic button on a phone with a control-room number to call. Vygard adds the AI dispatcher copilot (Howie / Claude Opus 4.8) that drafts the incident briefing in under a second, the proactive welfare-ping engine that catches workers in trouble before they hit SOS, the 12-sector vertical packs tuned to specific hazards, and a tamper-evident audit trail that satisfies CQC, HSE and RIDDOR without spreadsheet chasing. Pager-only setups have no audit, no AI, no escalation and no regulator-ready reporting at all.

Does Vygard need special hardware?

No — the standard deployment is BYOD on phones, watches and the web PWA. Native iOS, Android, watchOS and Wear OS apps unlock richer features (vitals, watch SOS, fall detection). For workers in tougher environments, Vygard supports BLE pendants, rugged handsets (Cat S62, Sonim XP10, CrossCall) and — for ATEX zones — intrinsically-safe handsets from Bartec, Ecom, i.safe and Aegex.

What is "Howie" and what does it actually do?

Howie is the AI watcher inside Vygard, powered by Claude Opus 4.8. Three jobs: (1) sub-second SOS briefing — when a worker triggers, Howie composes the situation summary, next action and priority call list inside the dispatcher's SOS modal before the dispatcher's even finished reading the alert; (2) welfare anomaly detection — Howie scans vitals, location, weather, livestock zones and crime overlays every 60 seconds and decides when a proactive check-in is warranted; (3) audit support — structured tool-use output with a strict JSON schema means every Howie action is logged, reviewable and exportable for RIDDOR / HSE.

How fast can it be deployed?

Vygard's standard rollout is live in 10 working days from order. No hardware required for a BYOD launch, no MDM enrolment, no app-store deployment friction — workers open the PWA URL or install the native app and they're on the system. Sector pack templates make this much faster than building lone-worker workflows from scratch.

Is it compliant with UK lone-worker regulations?

Yes. Vygard is BS 8484-aligned for provision, CQC Reg 13 aligned for adult social care, HSE / DVSA / SIA / MREW / FISA 302 aligned per sector pack, and produces RIDDOR-ready auto-PDF exports. UK-hosted in AWS eu-west-2. SOC 2 Type II and ISO 27001 are in audit (targeted Q3-Q4 2026), with Cyber Essentials Plus and NHS DSPT in flight.

Can it handle workers in ATEX hazardous zones?

Yes — Vygard is the only lone-worker platform that runs on intrinsically-safe ATEX-rated handsets (Bartec, Ecom, i.safe, Aegex). The petrochemical sector pack adds a Heat Stress Index (welfare-at-70, evac-at-85), and the oil & gas workflows hook into Iridium satellite-fallback SOS for when workers go beyond cellular coverage on rigs and remote installations.

What does it integrate with?

Outbound: Microsoft Teams and Slack for comms, ServiceNow for ITSM, BambooHR and Workday for HRIS, Twilio for SMS and voice, Resend for email, MQTT for IoT, SIA-IP and Fast Format for ARC escalation, Salesforce for CRM. Vygard's tenant publishes events via webhook so custom integrations are straightforward — most customers wire Vygard to their incident-management and HR systems within the first week.

How quickly does help actually reach a worker?

Vygard's published figures put the median SOS-to-dispatcher time at around 90 seconds — a sub-second AI briefing from Howie, then a 60-second escalation to the alarm-receiving centre if no-one acknowledges. In elder care the median fall-to-family alert is about 8 seconds. The mutual-aid network also routes to the nearest peer responder — typically around 4 km away in rural work, versus a 20-minute-plus rural 999 response — so help can be moving while the formal escalation runs in parallel.

How do we justify the cost — what's the ROI?

Vygard typically pays for itself by retiring three to five legacy tools (pagers, check-in apps, permit systems, spreadsheets) and reducing incidents and insurance exposure. In Vygard's illustrative 150-worker construction scenario — retiring three tools at ~£3.5k/yr each and a broker-typical insurance saving — net annual savings come out around £18,000 with a ~7.8-month payback, roughly £0.62 per worker per day. Against Sentencing Council fines that start around £180,000 for medium-culpability harm, the risk-reduction case usually dwarfs the licence cost.

What does it cost?

Transparent per-worker pricing. Starter is £9/worker/month (core safety, live map, basic audit). Pro is £19/worker/month (the most popular tier — adds AI module, risk scoring, compliance hub and one sector pack). Enterprise is custom from £49/worker/month (custom escalation, dedicated CSM, 24/7 SOS line). 14-day free trial on all plans. No per-supervisor fee.