Sample report · fictional company, made-up data
The September pack for Bilby Flats Contracting.
A made-up WA mining contractor with about 950 people on site, built exactly the way a client's pack is built. Every number below comes from the same synthetic dataset used in the walkthrough videos.
August: 232,324
Show the working
Sum of hours_worked for September from the hours export. Checked by hand against the payroll total.
2 medical treatment, 0 lost time
Show the working
Count of injuries classed LTI, RWI or MTI, as agreed in the written definitions. First aid is not recordable.
18 recordables / 2,129,400 hours
Show the working
Recordable injuries year to date x 1,000,000 / hours worked year to date. 18 x 1,000,000 / 2,129,400 = 8.45. Target 5.0.
August: 22
Show the working
Events with a potential severity of 4 or 5 on the 5 x 5 matrix, any category.
Target: under 10
Show the working
Actions with a due date on or before 30 September and no closed date.
TRIFR, year to date
Counted correctly, it's 8.45 against a target of 5.0, down from 10.58 in June. If the 71 first aid cases were wrongly counted as recordable, the same chart would show 41.80. The switch above shows both.
From exports to pack
This is the raw material for September, and what it turns into. Slide across.
# incidents_export_20261001_FINAL_v2.csv incident_id,date,shift,area,category,mechanism,severity_potential,injury_class,days_lost,status BFC-IR-2026-0234,2026-09-02,Day,ROM pad,Near Miss,Vehicle interaction,5,,0,Under investigation BFC-IR-2026-0235,2026-09-03,Day,Workshop,Injury,Hand and finger,3,FAI,0,Under investigation BFC-IR-2026-0236,2026-09-03,Night,Civil works,Near Miss,Underground services,5,,0,Closed BFC-IR-2026-0237,2026-09-05,Night,Workshop,Injury,Hand and finger,3,FAI,0,Closed BFC-IR-2026-0238,2026-09-08,Day,Go-line,Property Damage,Vehicle interaction,3,,0,Closed BFC-IR-2026-0239,2026-09-11,Day,Workshop,Injury,Hand and finger,3,MTI,0,Open BFC-IR-2026-0240,2026-09-13,Night,Pit,Injury,Slip trip fall,3,FAI,0,Closed BFC-IR-2026-0241,2026-09-14,Day,Workshop,Near Miss,Energy isolation,5,,0,Under investigation BFC-IR-2026-0242,2026-09-18,Day,Workshop,Injury,Hand and finger,3,MTI,0,Closed BFC-IR-2026-0243,2026-09-20,Night,Go-line,Property Damage,Vehicle interaction,3,,0,Under investigation BFC-IR-2026-0244,2026-09-22,Night,Go-line,Near Miss,Fatigue,5,,0,Under investigation BFC-IR-2026-0245,2026-09-22,Day,Pit,Near Miss,Vehicle interaction,4,,0,Open BFC-IR-2026-0246,2026-09-23,Day,Pit,Near Miss,Vehicle interaction,5,,0,Closed BFC-IR-2026-0247,2026-09-23,Day,Workshop,Injury,Hand and finger,3,FAI,0,Closed # hours_2026.xlsx (sheet: Site hrs) month,hours_worked,headcount_avg 2026-08,232324,929 2026-09,242037,968 # actions_open.csv action_id,due_date,closed_date,owner_role BFC-AC-0003,2026-09-29,,HSE Advisor BFC-AC-0006,2026-02-28,,Pit Supervisor BFC-AC-0008,2026-07-29,,HSE Advisor ...
September 2026 at a glance
Headline. No lost time injuries. Two medical treatment injuries, both hand lacerations from sheet metal in the workshop. TRIFR year to date is 8.45 against a target of 5.0. Overdue actions sit at 40 against a target of under 10.
What the data says. Reported events fell from 33 to 17. Most of that drop is near misses. [CHECK: did reporting drop, or did the hazards?]
The narrative, with its [CHECK] markers
Drafted by AI from the checked KPI summary only. Every marker is something a person had to answer. Tap one to see what happened to it.
1. Headline
- No lost time injuries. Two medical treatment injuries, both hand lacerations from handling sheet metal in the workshop.
- TRIFR year to date is 8.45 against a target of 5.0.
- Overdue actions sit at 40 against a target of under 10. This is the biggest gap right now.
2. What the data is telling us
- Reported events fell from 33 in August to 17 in September, and high-potential events from 22 to 8. Partly answered from the data, then passed to the client. Most of the drop is near misses (19 in August, 7 in September). Injuries barely moved (8 to 6). Water services reported nothing at all in September. Question for the client: was that crew working, and did anything change in how near misses are reported?
- Vehicle interaction is still the most common mechanism (6 events), including two light vehicles entering intersections without a positive radio call.
- Hand and finger injuries rose from 3 to 5, all in the workshop: three pinch points fitting hose clamps and two sheet metal lacerations. Passed to the client. Both laceration reports say cut-rated gloves weren't worn. Neither says whether any were available at the bench. That needs someone who was in the workshop.
- Two night shift operators reported microsleeps at the go-line and were stood down. Reporting it is the right behaviour. Two in one month is a pattern.
3. Decisions or support needed
- Agree an owner and a date to clear the 40 overdue actions. Answered from the data. Yes. 19 of the 40 overdue actions sit with the HSE Advisor role. The next highest is the Pit Supervisor with 10. Owner table added to the pack as evidence.
- Endorse a review of night shift fatigue controls, given two microsleeps in a month.
The three claims the AI made up
The first draft read well and was mostly right. Three lines came from somewhere other than the data it was given. This is why every pack is checked against source before it leaves.
Claim 1
“TRIFR is moving down slowly.”
Not in the summary the AI was given. It happens to be true since June (10.58 to 8.45), but it was a guess. Fix: the monthly trend is now in the pack, so the claim has evidence.
Claim 2
“It will not reach target this year at the current rate.”
A projection with no maths behind it. Fix: removed. If leadership wants a forecast, it gets its own page with the working shown.
Claim 3
“19 actions are owned by the HSE Advisor.”
True, but it came from an earlier query in the same chat, not from the summary. Fix: the owner breakdown is now part of the pack, below.
Overdue actions by owner
At 30 September. Due on or before that date, not closed.
High-potential events, grouped
| Mechanism | Events | What happened |
|---|---|---|
| Vehicle interaction | 4 | lV entered intersection without positive radio call while haul truck approaching (ROM pad, day shift, 2 Sep); haul truck reversed at dump point with reduced bund height (Pit, day shift, 22 Sep); lV entered intersection without positive radio call while haul truck approaching (Pit, day shift, 23 Sep); lV made contact with windrow on haul road bend (Haul road, day shift, 27 Sep) |
| Fatigue | 2 | operator reported microsleep on night shift and was stood down (Go-line, night shift, 22 Sep); operator reported microsleep on night shift and was stood down (Go-line, night shift, 28 Sep) |
| Underground services | 1 | excavator bucket exposed unmarked conduit during culvert trenching (Civil works, night shift, 3 Sep) |
| Energy isolation | 1 | hydraulic line opened before stored pressure was confirmed released (Workshop, day shift, 14 Sep) |
One set of numbers, three formats
Your board wants one layout. Each principal wants another. Same checked figures, rendered for each. Principals here are fictional.
| Measure | September | Target | Status |
|---|---|---|---|
| TRIFR, year to date | 8.45 | 5.0 | Off track |
| Lost time injuries | 0 | 0 | On track |
| Overdue actions | 40 | Under 10 | Off track |
| Contractor statistics, Sep 2026 | Month | YTD |
|---|---|---|
| Exposure hours | 242,037 | 2,129,400 |
| Recordable injuries (LTI + RWI + MTI) | 2 | 18 |
| First aid injuries | 4 | 71 |
| High-potential incidents | 8 | see register |
| Inspections completed / findings | 7 / 11 |
* Principal B asks for a 12-month rolling rate. Bilby Flats only has 2026 data, so year to date is shown and labelled. That decision goes in the change note.
Want yours to look like this?
Bring last month's report to a 30-minute call. You'll see what would change.