AI is reshaping how we assess, manage and communicate risk.
AI isn't just automating tasks. It's changing how data is collected, interpreted and acted upon. That shift has real consequences—and opportunities—for risk engineers, brokers, underwriters and facility managers alike.
Across the field, AI is beginning to show up in practical, measurable ways. Risk engineers are using digital twins—virtual models of physical facilities—to simulate fire protection system performance and track when equipment is due for maintenance. These models provide dynamic insights into how systems behave under different conditions, helping teams plan ahead rather than react after something breaks.
At the same time,
Another fast-emerging tool is large language models (LLMs). As part of the natural language processing (NLP) domain, these models scan massive volumes of inspection reports and documentation to identify recurring issues, emerging risks, and trends that might otherwise go unnoticed. Instead of relying solely on human pattern recognition, AI can sift through data at scale and surface insights that help risk managers make more informed decisions.
These applications don't eliminate the need for trained professionals. But they allow teams to see more, move faster, and prioritize actions that reduce the likelihood of a costly event.
What makes a submission stand out
When I was underwriting, the strongest submissions weren't just those with low modeled losses—they were well-organized, rich with reliable data, and clearly documented.
AI is now raising expectations.
Forward-looking risk engineers are now using AI to strengthen those submissions: identifying data gaps, integrating risk data, and ensuring every site profile is as complete as possible.
It all comes back to data. "Garbage in, garbage out" applies here as much as anywhere. The most reliable data doesn't come from generalized assumptions—it comes from independent risk engineering.
When companies outsource risk engineering to their insurers, they often lose direct access to inspection results or risk recommendations. Independent firms, by contrast, provide not just the insights—but the ownership. That includes data on risks that fall below a deductible, helping companies get a fuller view of their exposures.
Data ownership also drives better strategy. It supports more informed internal decisions, more persuasive insurance negotiations, and
AI is a tool, not a replacement
AI can detect anomalies, simulate what-if scenarios, and flag unusual trends—but can't walk through a facility and make critical decisions. Human discretion and engineering judgment are still essential.
It's also important to acknowledge the real risks AI brings: bias and variance in data models, security issues, regulatory pitfalls, and gaps in historical data for new or emerging hazards. Risk managers and brokers need to vet AI tools as carefully as any other vendor—and know how to challenge the results when needed.
Looking ahead, expect AI to help engineers work faster, analyze more data, and identify risk trends sooner. Underwriters will likely use AI models to price risk more efficiently—but human expertise and deep communication with insureds is still mission critical, especially when working with complex occupancies or high-value portfolios.
The organizations that succeed won't be those that hand over decisions to machines. They'll be the ones that use AI to sharpen their understanding without losing sight of the people and expertise that keep facilities safe.
Start with this: own your data, improve the quality of your submissions, and use AI as a tool to support—not override—good engineering.
