Separating signal from noise for safety technology in 2026.
Every December, we sift through a year's worth of client engagements, conferences, and technology evaluations to identify what's genuinely poised to shift how organisations manage health and safety risk. Not what's being hyped. What's actually going to matter.
We recently delivered the final session of a four-part Innovation Spark series commissioned by CLP Group, one of Asia's leading power utilities. The focus was a contextualised 2026 forecast for their sector—but the themes have implications well beyond energy and utilities.
The shift from assistants to agents
If you've used Microsoft Copilot to draft an email or summarise meeting notes, you've experienced AI as an assistant. Useful, but limited. What's coming in 2026 is different: AI agents that complete workflows, not just tasks.
The major EHS platforms are embedding these capabilities now. The implications for incident investigation, assurance processes, and administrative burden are significant. But there's a catch—and it's the part that tends to dampen enthusiasm in boardrooms.
The foundation most organisations are missing
We're running a research project with electrical utilities across Australia and the US, examining how they're applying AI to health and safety. The pattern is consistent: organisations start with an exciting use case, invest time and resources, then stall. The reason is almost always the same.
It's not a technology problem. It's a data problem. And the solution requires rethinking how we capture information in the first place—not just how we analyse it.
What this means for your planning
If we were engaged to advise a safety team on technology priorities for 2026, we'd focus on a few things:
- Get serious about data strategy. Not technology strategy—data strategy. What problems are you trying to solve? What data would you need? What are you capturing today that's noise, and what's missing that would be signal?
- Map your personas and workflows first. Before any AI deployment, understand who needs what information, in what format, and at what point in their workflow. A field supervisor, an executive, and an assurance lead all need different things from the same underlying data. The organisations getting value from AI have done this groundwork.
- Understand what your platforms are releasing. The AI capabilities coming to EHS software won't require you to build anything—they'll appear as features. Taking time to understand what's available and what your data needs to look like to take advantage of it is a capability-building exercise. It also clarifies the build-versus-buy question before you invest in custom development you may not need.
This post draws on insights from a quarterly strategic foresight session delivered to CLP Group as part of PKG Safety Innovation's tailored intelligence service. If your organisation would benefit from contextualised technology forecasting and strategic guidance, get in touch to discuss commissioning a similar engagement.
The comprehensive PKG Safety Innovation 2026–2028 Strategic Intelligence Report—covering the full landscape of signals shaping health and safety leadership—will be released in early 2026. Subscribe to our newsletter to be notified.