Uploaded March 2026 | Updated September 2026, 2 weeks ago
If your safety metrics are improving, your people should be safer.
But what if the numbers are measuring performance, not risk?
I’m joined here by John Dony, CEO and co-founder of the What Works Institute, and Mike Swain, Technical Enablement Manager at Evotix, to examine a stubborn operational problem: why serious injuries and fatalities remain hard to reduce even when reporting, systems, and traditional safety KPIs appear to improve. Their work sits right at the intersection of supply chain resilience, operational risk, human factors, EHS data, and practical AI adoption.
This matters now because supply chains are under more strain, not less. Labour constraints, contractor-heavy operating models, tighter compliance expectations, more fragmented accountability, and growing pressure to digitise decision-making all raise the stakes. If your metrics create a false sense of control, you are not managing resilience. You are managing optics. And in a complex supply chain, that gap becomes a very real business risk.
A few things in this one genuinely sharpened my thinking. First, Mike’s point that incident reporting jumping by 800% can actually be good news if it means reality is finally surfacing. Second, John’s argument that traditional safety metrics can slip into performance theatre, where organisations manage the appearance of safety rather than the real drivers of harm. And third, the AI point: the real constraint is not ambition, it’s data quality, governance, trust, and whether people believe these systems are helping them or policing them.
If you lead supply chain, procurement, operations, EHS, sustainability, or enterprise risk, this one is for you.
If you’re dealing with this on the ground, I’d like to hear how you’re handling it.
🔗 Podcast website: resilientsupplychainpodcast.com
🔔 Follow the Resilient Supply Chain Podcast for weekly conversations on operational resilience, supply chain risk, sustainability, and decision-making.
🎧 Also available wherever you get your podcasts.
Timestamps
00:00 – Cold open: AI, data, and “the monolith”
00:17 – Why improving metrics may still hide serious harm
09:04 – Why serious injuries remain stubbornly high
12:18 – Why the safety triangle breaks down
17:22 – What companies are actually measuring
21:26 – When safety metrics become performance theatre
25:11 – Why no single metric solves this
25:18 – What counts as serious harm
28:30 – Human factors, fatigue, and stress
30:42 – Where AI is useful in safety today
33:48 – Why AI still depends on data quality
35:46 – One thing leaders should remember
If your safety metrics are improving, your people should be safer.
But what if the numbers are measuring performance, not risk?
I’m joined here by John Dony, CEO and co-founder of the What Works Institute, and Mike Swain, Technical Enablement Manager at Evotix, to examine a stubborn operational problem: why serious injuries and fatalities remain hard to reduce even when reporting, systems, and traditional safety KPIs appear to improve. Their work sits right at the intersection of supply chain resilience, operational risk, human factors, EHS data, and practical AI adoption.
This matters now because supply chains are under more strain, not less. Labour constraints, contractor-heavy operating models, tighter compliance expectations, more fragmented accountability, and growing pressure to digitise decision-making all raise the stakes. If your metrics create a false sense of control, you are not managing resilience. You are managing optics. And in a complex supply chain, that gap becomes a very real business risk.
A few things in this one genuinely sharpened my thinking. First, Mike’s point that incident reporting jumping by 800% can actually be good news if it means reality is finally surfacing. Second, John’s argument that traditional safety metrics can slip into performance theatre, where organisations manage the appearance of safety rather than the real drivers of harm. And third, the AI point: the real constraint is not ambition, it’s data quality, governance, trust, and whether people believe these systems are helping them or policing them.
If you lead supply chain, procurement, operations, EHS, sustainability, or enterprise risk, this one is for you.
If you’re dealing with this on the ground, I’d like to hear how you’re handling it.
🔗 Podcast website: resilientsupplychainpodcast.com
🔔 Follow the Resilient Supply Chain Podcast for weekly conversations on operational resilience, supply chain risk, sustainability, and decision-making.
🎧 Also available wherever you get your podcasts.
Timestamps
00:00 – Cold open: AI, data, and “the monolith”
00:17 – Why improving metrics may still hide serious harm
09:04 – Why serious injuries remain stubbornly high
12:18 – Why the safety triangle breaks down
17:22 – What companies are actually measuring
21:26 – When safety metrics become performance theatre
25:11 – Why no single metric solves this
25:18 – What counts as serious harm
28:30 – Human factors, fatigue, and stress
30:42 – Where AI is useful in safety today
33:48 – Why AI still depends on data quality
35:46 – One thing leaders should remember
