Yes, for the paperwork and the patterns, not the investigation. AI can turn a worker's voice note, photo or message into a structured near-miss or incident report in your form, in Arabic or English, and group reports across sites to show repeat hazards. The HSE officer verifies the facts, leads the investigation, decides root cause and corrective actions, and signs.
#Key takeaways
- Near misses go unreported mostly because reporting is slow and awkward, so the first win is making it easy to report.
- AI turns rough inputs into a complete report in your form and asks for the fields that are missing.
- The investigation, root cause and corrective actions stay with the HSE officer and the site manager.
- Pattern-spotting across many reports is where AI adds something a busy HSE team rarely has time for.
- Anything that must go to a regulator or a client is checked and sent by a person.
#Why do near misses go unreported on site?
A steel fixer sees a hook load swing close to a crew. Nobody is hurt. Reporting means finding the HSE officer, filling a form in English, and naming people, so most of the time it becomes a story in the canteen instead of a record. The HSE officer, meanwhile, spends evenings retyping handwritten forms into Excel for the monthly statistics.
That matters, because near-miss reporting is linked to better safety results. Construction Industry Institute, checked november 2026 reports that near-miss reporting was found to lower the OSHA total recordable incident rate and to increase communication and trust about safety on construction sites.
#What can AI do in the HSE reporting process?
- intake in the worker's language. a voice note or message in Arabic, English or another site language becomes a draft report.
- complete the form. location, activity, equipment, people involved, immediate actions, with gaps flagged as questions back to the reporter.
- classify. near miss, first aid, property damage or recordable, as a suggestion for the HSE officer to confirm.
- spot patterns. group reports by hazard, trade, subcontractor and area across sites.
- track actions. remind owners of open corrective actions and summarise status for the monthly HSE report.
#What must stay with the HSE officer?
The facts of the event, the classification, the investigation, the root cause, the corrective actions and any notification to authorities or the client. Many projects add their own rules on top of local law: Aldar, checked november 2026, for example, requires contractors on its developments to follow the Abu Dhabi OSHAD framework and a hazard, near miss and incident reporting and investigation procedure. The person accountable under those rules signs the report.
AI should never soften wording, decide blame, or close an action. The NIST AI risk management framework treats human oversight and clear accountability as part of managing AI risk, and safety records are where that matters most.
#When is a simpler fix enough?
On a single site with a strong HSE culture, a short mobile form with pictures, available in Arabic and English, a QR code at each gate and a toolbox talk on why near misses matter may lift reporting more than any AI. Many HSE software products also include mobile reporting out of the box.
AI earns its place when you run several sites with a multilingual workforce, reports arrive in many formats, and the HSE team has no time to look for patterns across them.
#Where does Stennir fit?
Stennir's AI consulting audit looks at HSE reporting next to your other workflows and ranks five, each with a cost and a payback case. If it is the right first build, AI implementation puts the intake and pattern workflow into the channels your workforce already uses, with the HSE officer approving every report. You pay for finished work.
AI training then works with HSE officers and supervisors on the real workflow, including how to check a draft and when to go back to the reporter.
We are
Stennir builds HSE reporting workflows that make reporting easy and patterns visible, with the HSE officer approving every report and action.
We aren't
Stennir is not an HSE consultancy and does not investigate incidents, assign blame or notify regulators.
Related reading: AI for SOP automation.
#What should an HSE manager do next?
Take a sample of recent near-miss and incident reports, note which fields are usually missing and how they arrived, and ask your supervisors why some events are never reported. See AI for operations, take the free readiness assessment or book a 30-minute call.
#Common questions
What do you need from us?
Your incident and near-miss forms, a sample of past reports in whatever format they arrived, your classification rules, the channels your workforce uses, and the HSE officer who would approve drafts. Stennir maps the reporting path with them and agrees which steps stay strictly with a person.
Can workers report in their own language?
Arabic and English are fully supported, and other site languages can be added on request after testing on real voice notes and messages. The report keeps the original words next to the structured version, so the HSE officer can check what the worker actually said.
Where is incident data stored?
In the region your rules require. Incident reports can contain health and personal data, so Stennir hosts the workflow and runs the AI where your data protection obligations point, limits who can see names, logs every action, and leaves you the accounts, the data and the off switch.
What could go wrong?
A wrong classification, a missing detail filled in by guesswork, or a pattern that is really a reporting habit. That is why the AI asks the reporter instead of guessing, classifications are suggestions until the HSE officer confirms them, and every report links back to the original message or photo.
