We are delighted to announce the launch of COMET AI Assistant, the first and only AI capability built specifically for incident investigation and root cause analysis.
This has been a significant undertaking. The development of COMET AI Assistant represents a six-figure investment in purpose-built AI capability, and it reflects something we feel strongly about: there is a meaningful difference between AI that has been applied to investigations and AI that was built for them. What we are launching today is the latter.
Unlike tools adapted from generic AI solutions, COMET AI Assistant is built from the ground up on the investigative expertise, methodology, and HSEQ data knowledge that has made COMET the trusted name in investigation and RCA, protecting more than 350,000 people across 190 or more global locations worldwide.
“There is a difference between AI that has been applied to investigations and AI that was built for them. COMET AI Assistant is built on the process, the methodology, and the safety data expertise that drives meaningful prevention. This is not a feature update, it is a category shift in what investigation software can do. We built this because our customers deserved better than a generic AI bolted on to their workflow, and we had the expertise and the methodology to do it properly.”
Mark Rushton, CEO, COMET
Why we built it
Even in organisations with mature investigation programmes, three challenges tend to persist.
The first is the documentation burden. Before any meaningful analysis begins, investigators are rebuilding timelines from memory, drafting summaries from scratch, and writing actions without a consistent framework. That time belongs to understanding why the incident happened, not to administration.
The second is inconsistency in investigation quality. When the depth of an investigation is shaped by who is conducting it and how much time they have, organisations accept a level of variability that carries real risk. Barrier failures go unrecorded. Human factors get reduced to conclusions that don’t lead to prevention. Actions are written at a level that makes recurrence likely.
The third is the intelligence gap. Most organisations have years of investigation data. What they rarely have is a practical way to interrogate it at portfolio scale. The systemic patterns, recurring causal factors, and early warning signals are present in that data. Getting to them has always required resource that most teams cannot spare.
We built COMET AI Assistant to address all three.
What we built
COMET AI Assistant is structured around two distinct but complementary capabilities, each addressing a different stage of the investigation lifecycle.
Within the investigation
The first capability is embedded directly into the COMET Investigate workflow and activates at the point it is most useful. It reduces the documentation burden that delays investigations before they begin, elevates the quality of analysis, and ensures nothing gets missed regardless of investigator experience or time pressure.
The design principle we held throughout is that the AI supports, and the investigator decides. Every output is a suggestion that requires human review and approval before any change is committed. Investigative judgement cannot be automated. What we have significantly reduced is the blank-page effort that consumes time before any judgement is applied.
Across the portfolio
The second capability, delivered through COMET Companion Plus, builds on our long-standing capability to surface trends and patterns through investigation dashboards and analytics, and takes it significantly further.
It operates across the entire investigation portfolio through natural language conversation, with no technical expertise required. It surfaces the systemic patterns, recurring factors, and emerging risks that manual review cannot reach, enabling organisations to move from reactive investigation to proactive risk intelligence.
All you have to do is ask.
Together, the two capabilities represent end-to-end AI support across the investigation lifecycle that no other RCA platform currently offers.
Independent validation
The launch follows an independent briefing with Verdantix, the leading analyst network for HSE technology, whose research is used by safety leaders and procurement teams across the industry. Their assessment reinforced what we have always believed about the importance of building on genuine investigative expertise.
How we built it to be trusted
Our customers operate in some of the most regulated, high-risk industries in the world. We took that seriously throughout the development process.
COMET AI Assistant does not use OpenAI or any public large language model. The AI is purpose-built and contained within our own architecture. Client data is fully segregated, does not leave the client environment, and does not contribute to any shared model. The platform is GDPR-compliant throughout. Legally privileged investigations are excluded from portfolio analysis by design.
We have also built COMET AI Assistant for phased adoption. Organisations can enable individual features progressively, building confidence at each stage, with no requirement to deploy everything at once.
See it for yourself
To see COMET AI Assistant in action, book a personalised demo or speak directly to your account manager at COMET.
COMET AI Assistant is a decision-support tool designed to assist COMET users throughout the investigation workflow and across the investigation portfolio. Final judgement and decision-making responsibility remain with the individual.
.webp)
.webp)
