Whose policy answers when the app is on

The useful question is not whether the driver was working. It is which phase the app recorded at the moment of impact, because cover for a gig vehicle is a range rather than a state, and the range has an end where nobody disputes anything, a middle where claims are refused, and another end where the answer is again clear.

The phases below are the general pattern that app-based passenger and delivery work has settled into. They are not any particular company’s terms, and this piece does not assert any company’s terms. What travels across arrangements is the shape: a driver who is offline, a driver who is online and waiting, a driver on the way to collect, and a driver carrying the passenger or the parcel.

The end where nothing is contested

App closed, no job open, the car being driven to a supermarket. This is private use of a private vehicle, and an ordinary personal motor policy is the policy that answers. Nobody in the chain has an argument to make.

It matters only because it is the phase drivers generalise from. A driver whose policy has always answered for the driving they do concludes that the policy answers for the driving they do, and the conclusion holds until the app is open.

The other clear end, and why it is clear

Passenger in the car, or parcel in the back, a fare or a fee being earned. Almost every personal motor policy excludes carrying passengers or goods for hire or reward. The exclusion is old, it predates the applications by decades, and it is not an oversight: an insurer that priced a commute did not price a vehicle in commercial service for ten hours a day.

So a claim arising in this phase is not a grey area on a personal policy. It is outside it. Where cover exists it comes from somewhere written for the purpose — a commercial or hire-and-reward policy the driver bought, or cover the platform arranges for its drivers while a job is live, where such an arrangement exists in that market. Both are real mechanisms, and neither is guaranteed to exist where a particular driver works: this piece does not say, of any market, that platform-arranged cover is present there, or mandatory, or absent. Where neither route is available, the honest answer to the question in the title, for this phase, is that nothing responds. Which one applies to any given driver is in their documents and in the platform’s terms, and reading both is the whole of the work.

The contested middle

Between those ends sit two phases, and they are where the refusals come from.

The driver is logged in and waiting for a job. No passenger, no parcel, nothing being carried. Is that private use, or is it commercial availability? A personal insurer can argue that a vehicle held ready for hire is in commercial use regardless of whether anyone is aboard; a driver can argue with equal plausibility that driving around an empty car is the same activity it was ten minutes earlier. Where a platform’s cover exists, this is typically the phase in which it is thinnest, and in some arrangements it addresses injury to other people while leaving damage to the driver’s own vehicle uncovered.

Then the driver accepts a job and drives to collect. Still nobody aboard, but now unambiguously on a commercial errand, on a route chosen by the platform, being paid for by the platform’s fee structure. The commercial character is harder to dispute here, which sounds like an improvement and often is not, because the personal policy’s exclusion bites more cleanly at the same time.

The verdict, stated plainly: the middle is not an area where reasonable people disagree about a well-understood boundary. It is a structural gap, and it exists because a personal policy and a commercial policy were designed for different activities, while the activity of gig driving passes through both several times an hour. A driver working in the middle without cover bought for it is uninsured for part of every shift, whatever the two policies each say about their own halves.

What makes the gap expensive rather than merely untidy

The discovery happens at claim time. A personal policy is not cancelled the moment a driver logs in; it sits there, paid, apparently in force, and it declines the claim afterwards. In the meantime the driver has a damaged vehicle they earn with, and possibly an injured third party whose claim has to be answered by somebody.

It is also the phase structure that decides what the injured third party can recover. What an injured third party can recover is bounded by the compulsory minimum limits where the collision happened, and whether an uninsured or underinsured motorist route exists for a gig driver hurt by someone else, or for a passenger, is a separate question again. Both are jurisdictional, and both are in the rules for your jurisdiction below.

And the record that decides which phase applied belongs to the platform. A driver arguing that the app had closed, or that the job had ended, is arguing against a log they cannot produce.

Two ordinary complications make the picture worse than the phases suggest. A vehicle used for more than one kind of app work passes through a different set of exclusions in each, so an endorsement bought for one activity can leave the other uncovered on the same afternoon. And a vehicle driven by more than one person — a household car worked in shifts, or a car rented to a driver by somebody else — puts a question about permitted drivers on top of a question about permitted use, and the two are answered in different parts of the same document. Neither complication is unusual, and neither is visible to anybody until a claim is made.

The injured third party is the reason this matters beyond the driver’s balance sheet. Where no policy responds, the claim does not disappear; it goes wherever local law sends unsatisfied motor claims, and it goes there slowly. A company whose work was being performed at the time is a visible and solvent place for that claim to arrive, whatever the contract says about who was engaged as what.

What a driver should confirm, and what a business contracting drivers should confirm

For the driver, the work is to name the activity to the insurer in the platform’s own vocabulary and get the answer in writing. Not whether cover exists for delivery, but whether it covers this platform’s food deliveries in this vehicle at these hours, including the waiting phase, including the drive to the collection point, and whether it covers damage to the vehicle or only liability to others. A policy that answers only for liability leaves the earning asset exposed, which for a driver whose car is the business is the material risk.

For a business that contracts gig drivers, the exposure is not the driver’s repair bill. It is a third party’s injury claim arriving at a company that assumed it had contracted out the risk. The instruments that hold are evidence of appropriate cover collected before the first job and again at each renewal, a written requirement that the cover permits the use, a liability programme of your own that contemplates vehicles you do not own being driven on your behalf, and an honest look at whether your scheduling makes the required cover affordable for the people you expect to hold it. A requirement nobody can afford is a document rather than a control.

What we cannot tell you is where the boundary sits in your case, because the two documents that decide it are private: the driver’s policy wording and the platform’s terms as they stand today. Both change, the second without notice and without anyone’s signature, and an answer confirmed a year ago is not an answer now.