The short answer
Robotics, drone, autonomous-vehicle and hardware applications get declined because they reach insurers whose appetite has no room for machines, or because a generic application shows the hazard without the engineering that controls it. A decline is a statement about one carrier's box. The same company, described in an application built for machine risk and sent to insurers who want autonomy, is routinely quoted on the second submission.
What a declination actually means
A declination letter reads like a verdict on your company. It almost never is one. When a carrier declines a robotics or AI application, the usual meaning is narrower and duller: this risk does not fit the box this carrier prices. Every insurer has an appetite, the classes of business it wants, and submissions outside it get declined regardless of quality. A robotics company declined by a generalist insurer has learned about that insurer's appetite, and very little about its own insurability.
The declines that sting most are the ones without reasons. Standard market practice is a one-line declination, and founders are left guessing whether the problem was the autonomy, the revenue, a word on the application, or nothing at all. Sometimes the trigger is a single word on the application that an automated rule or a hurried underwriter matched to a risk the insurer avoids.
Standard commercial applications were designed for restaurants and contractors. They have no fields for a supervision model, an operating envelope, geofencing, or an emergency stop. An underwriter reading a robotics company through that form sees only the hazard words, such as autonomy and customer premises, with none of the engineering that controls them. Pricing what you cannot see means padding it or declining it, and generalist carriers mostly decline.
Where each segment trips the wire
This is a presentation problem wearing the costume of an insurability problem. The same operation, described in an application built for machine risk and sent to insurers whose appetite includes robotics, routinely produces quotes where the generic route produced declines. Specialist capacity for robotics and autonomy exists, and a form with no field for what makes the risk good cannot reach it.
The pattern repeats across hard tech, with a different tripwire in each segment. A drone company trips the aircraft exclusion: standard general liability forms exclude injury and damage arising from aircraft you own or operate, and a small unmanned aircraft is an aircraft under that wording. A generalist either declines the whole account or quotes it and silently excludes the flying part, which is worse, because the certificate looks fine until the claim. An autonomous-vehicle company trips two appetites at once. Operation on a public road puts the risk in commercial auto, where few carriers write a vehicle without a driver, and selling the driving stack puts it in products liability, where the underwriter has no loss history to price from.
A hardware company gets declined for being early. Carriers whose appetite is finished goods write products liability for things with a spec sheet, a test report and units in the field, and a policy written that way can exclude pre-commercial products, prototypes and R&D units outright (a specialty broker's deep tech insurance page, accessed 2026-09-12). An AI software company faces the newest pattern. The standard-forms bureau's generative AI exclusions for general liability took effect in January 2026 (Independent Agent magazine, 2025-10-21), and a trade publication counted more than 60 insurance groups filing to adopt them by July (The Insurer, 2026-07-23).
What trips the decline, by segment
| Robotics on customer sites | Machines around people, and no field on the form for the safety case | Places with specialist general liability |
|---|---|---|
| Drones | The aircraft exclusion in standard general liability | Places on an aviation form or an aircraft endorsement |
| Autonomous vehicles | Public-road operation trips both auto and products appetite | Places with autonomy-specific auto and products capacity |
| Hardware and prototypes | Pre-commercial products excluded by finished-goods carriers | Places once the unit, the test plan and the field count are described |
| AI software | Generative AI exclusions, and carriers leaving the class | Places with carriers that cover models explicitly |
What a decline costs you
A decline costs time, and almost never money. Carriers do not charge to decline, and a broker is paid on a bound policy, so the only invoice from a failed submission is the weeks it consumed. Those weeks are the real cost. A pilot waiting on a certificate or a term sheet waiting on D&O turns a dull appetite decision into a deal delay.
We do not publish a number for a second submission, because there is no premium attached to it. What it needs is a better file. A description of the machine and where it operates in plain terms, the supervision model and operating envelope, the emergency stop and the geofence, the count of units in the field, and any incident history including the boring kind. For drones it needs the pilot roster with certificates, the aircraft list with values, and the operations you fly. For AI it needs what the model decides, what a human checks, and how you evaluate it.
Budget the calendar accordingly. Underwriter turnaround is measured in weeks, and running a generalist submission before a specialist one puts those weeks end to end. Starting with the specialist submission is the whole saving.
What to do after a decline
First, get the reason if one exists. A broker can usually learn whether the decline was appetite, a specific exposure, or a missing piece of information, and each has a different fix. Appetite declines mean the submission goes elsewhere. Exposure concerns sometimes negotiate: a carrier nervous about one deployment pattern may quote once it is fenced off or better documented. Information gaps just mean the application answers the question next time.
Second, treat the decline as information about the market. A declination is not a mark on your record that follows you between carriers, and most applications ask whether you have had coverage cancelled or non-renewed, which a declined application is not. What compounds instead is the file: each round of underwriter questions tells you what the market wants documented, and a submission that arrives with the safety case already written skips the round entirely.
Third, read the quotes that come back after the decline as closely as the decline itself. A carrier that quotes a drone company with the aircraft exclusion intact, or an AI company with a generative AI exclusion endorsed onto the general liability form, has declined the part of your business that matters and kept the premium on the rest. Read the exclusions page before you celebrate.
How the second submission gets placed
Start with the market list, since appetite decided the first round. Robotics and autonomy place with a short list of insurers whose appetite includes machines around people. Drones place with aviation markets that write an aviation liability form or a drone endorsement. AI software places with carriers that cover models explicitly, or that have declined to adopt the new exclusions. We keep those lists current and send the file only where it can be written.
Then describe the operation in the terms an underwriter prices. Tell us what you build and where it operates, in plain English, and we turn that into the safety case the generic application had no field for. The submission goes out with the supervision model, the operating envelope, the fleet count and the contract requirements already written, so the underwriter's first question is a pricing question.
This page describes coverage in general terms. It is not an offer of insurance, and carrier appetite, policy wording, licensing, and availability govern every quote and every claim.
Common questions
Why was my robotics company declined for insurance?
Most often because the submission reached an insurer whose appetite excludes machine risk, or because a generic application showed the hazard without the engineering that controls it. Both are placement problems with placement fixes. Neither is a verdict on insurability.
Does a declination go on my record?
No. A declined application is not a cancellation or non-renewal, which are what future applications ask about. Carriers do not share declination files, and the next submission starts fresh.
Can I find out why a carrier declined my application?
Often, through a broker. Carriers rarely volunteer reasons but frequently give them when asked by the broker who submitted the file, and knowing whether the issue was appetite, a specific exposure, or missing information determines the fix.
Are robotics companies insurable at all?
Yes. Specialist markets with appetite for robotics and autonomy exist. The failure is reaching the wrong markets through forms that cannot describe the risk, and the risk itself is writable.
Why did my drone company get declined for general liability?
Standard general liability forms exclude aircraft, and a drone is an aircraft under that wording, so a generalist carrier has nothing to write for the flying part of your business. Drone liability goes on an aviation form or an aircraft endorsement placed with markets that write unmanned aircraft. A quote that arrives with the aircraft exclusion still in it is a decline dressed as a quote.
Does an AI exclusion on my quote count as a decline?
For the part of your business the exclusion reaches, yes. The standard-forms bureau's generative AI exclusions for general liability took effect in January 2026, and a carrier that endorses one onto your quote has removed the claims a model-driven product would generate. Ask for the quote without it, or take the submission to a carrier that covers AI explicitly.
Terms in this guide
Sources
- 01Independent Agent magazine, on the standard-forms bureau's generative AI exclusions for general liability, effective January 2026, 2025-10-21
- 02The Insurer (subscription), "More than 60 P&C insurance groups file to adopt AI exclusions", 2026-07-23
- 03A specialty broker's deep tech insurance page, on pre-commercial product exclusions, accessed 2026-09-12
Read next
- Insurance for robotics companiesThe lines a robotics startup needs before its first customer deployment, with typical contract limits, the segments underwriters treat differently, and where 2026 AI exclusions reach robots.
- Product liability for hardware startupsWhat product liability covers for a company that ships hardware, why policies written for finished goods can exclude prototypes, and what demos, contract manufacturers and enterprise buyers ask you to carry.