Short answer
When a technology has no loss history, underwriters price it by analogy to similar risks, by studying the company's controls and exposure, and by limiting how much they will commit through sublimits, exclusions and shorter terms. Specialist and surplus lines markets usually lead, and the company can improve its position by presenting a clear, honest picture of how failure could actually happen.
Key takeaways
- Underwriters borrow from adjacent risks, so the better a company explains how its technology resembles familiar exposures, the easier it is to price.
- Controls and governance can substitute for loss data, because they show how likely and how severe a failure could be.
- Policies on new categories often carry sublimits, exclusions and conditions, so the wording matters more than usual.
- Surplus lines and specialist markets tend to write novel risks first, so access to them matters.
Why no history makes pricing hard
Insurance pricing normally relies on years of claims data that show how often losses happen and how large they become. A new technology category has none of that, so underwriters cannot simply apply a rate from a table, and they face the risk that a single unexpected event could be far worse than any model predicted.
WTW's discussion of physical AI illustrates the point, describing how software errors that once caused incorrect outputs can now cause bodily injury and property damage, and how responsibility spreads across manufacturers, developers, operators and maintenance providers. When the allocation of responsibility is unsettled, underwriters respond with caution.
Reasoning by analogy
The most common substitute for data is analogy. An underwriter asks which existing risks the new technology resembles, such as a software platform, a medical device or an industrial machine, and borrows relevant loss patterns from them while adjusting for what is different.
A company can help by supplying the analogy itself. Explaining which familiar category the product most resembles, where it differs and why those differences reduce or increase risk gives the underwriter a starting point and signals that the management team understands its own exposure.
Describe the risk, and we will find the market
Tell us what you are building and a broker will approach markets that write new categories.
Controls and governance carry extra weight
With no history to point to, evidence of how the company manages risk becomes the main underwriting input. WTW's recommended steps for physical AI include risk assessments before deployment that cover failure modes, safety engineering such as geofencing and shutoff systems, human oversight and monitoring of regulatory changes, and it notes that insurers increasingly ask about AI controls.
The same logic applies across emerging categories. Written procedures, testing records, incident response plans and independent reviews give an underwriter something concrete to weigh, and they often influence terms as much as the premium.
How underwriters limit their commitment
Underwriters facing uncertainty manage it through policy design. Common tools include sublimits for the least understood exposures, exclusions for scenarios they cannot yet evaluate, shorter policy terms that allow repricing, and conditions requiring the company to maintain specific controls.
Zywave notes that the market is also adding AI-specific wording, in some cases to exclude and in others to cover expressly. Companies should expect new-category policies to include more custom language than a standard form and should read it closely.
Why specialist markets lead
Standard admitted markets work from filed forms and rates, which makes it hard for them to respond quickly to something new. Surplus lines and specialist carriers have more freedom to craft custom terms, and they usually write novel categories first.
Reaching those markets typically means working with a broker that has relationships with specialist underwriters and understands how to present a new risk. A clear submission, with a plain description of the technology, the controls and the realistic scenarios, shortens the path to a quote.
Frequently asked questions
Can a brand-new technology category be insured at all?
Usually yes, through specialist or surplus lines markets that write custom terms, though the policy may include sublimits and exclusions.
What helps most in an underwriting submission for a new category?
A plain description of how the technology fails, the controls in place, and a familiar analogy that helps the underwriter place the risk.
Why do new-category policies have so many conditions?
Conditions let the insurer manage uncertainty by requiring specific controls to stay in place.
Will premiums fall as loss history develops?
Often they do, because data replaces assumptions, though this depends on how the category performs.
Sources
- WTW: Insuring physical AI, how robotics is reshaping risk and liability (August 2026)
- Zywave: How AI is creating new commercial insurance coverage gaps
This article is general information about insurance markets and is not legal or insurance advice. Coverage terms vary by carrier and policy form, so read the policy itself.
Related
Get a quote for a category that does not fit a box
Request access and a broker will follow up by email.