One automated path, two kinds of client

A broker with a personal lines book and a commercial book is advising two kinds of client about the same claims process, and the advice is not the same. Take the questions in turn and answer each of them twice.

What the client has to produce

For the personal client this is the whole ball game. Their photographs, taken on the day with a phone, become the basis of an estimate and therefore of the first offer, and nobody will inspect the vehicle unless something forces an inspection. The advice is concrete and unromantic: more images than the app asks for, each panel straight on and at an angle, a wide shot showing where the vehicle sat, the other vehicle and its plate, the surroundings, and anything that explains why the damage looks the way it does. Wet, dark and dirty are the conditions under which a model reads worst, so a second set in daylight is worth more than any argument made later.

For the fleet client the production problem is organisational rather than photographic. The driver has a phone, but the driver also has a schedule, a shift ending and an opinion about paperwork. What the fleet needs is a reporting routine that survives a bad afternoon: who is called first, what is captured before the vehicle moves, and where the images go so that they reach the insurer instead of a messaging thread on somebody’s phone. A fleet that already runs telematics has corroborating data the insurer may value, and a decision to make about handing it over.

Which claims the automation actually decides

Here the two diverge in a way that is easy to state backwards. Most personal motor claims fall inside the value band where an automated path can decide, so for that client automation is the normal case and human handling is the exception.

Most commercial motor claims sit above the band — except for the minor property damage that makes up the bulk of a fleet’s claim count. The fleet therefore meets the automation on precisely the losses it has most of and thinks about least: kerbed wheels, clipped mirrors, reversing damage, car-park contact. Those are the claims where an under-scoped estimate is least likely to be challenged, because nobody in the fleet has the time, and where the aggregate cost of not challenging is large by the end of a year.

What speed is worth

The personal client wants the claim resolved, and resolution speed is a real benefit that the industry could not offer before. Someone paying for a repair out of a current account is under pressure that grows daily, and a fast fair settlement is better for them than a slow larger one.

The fleet client is not buying settlement speed. It is buying vehicle availability, and those are different products. A fast payment on an under-scoped estimate produces a van that goes into a repairer, comes out incomplete, and returns for a supplement — which costs more days than a slower, correctly scoped authorisation would have. When a fleet client is choosing between insurers on the strength of claims speed, this is the distinction worth drawing for them.

What happens when the decision is wrong

The personal client has a complaints process, and after it, in many markets, an external route to a regulator or an ombudsman-type body. What that route is and what it can order is jurisdictional and must be read rather than assumed; the regulator for your jurisdiction is named in the rules below, and its own publications are where that route is described.

What the client needs from the broker before that stage is narrower and more useful: a written reason for the decision. A number with no reasoning behind it is difficult for an insurer to defend at any escalation, and asking for the reason in writing tends to produce either a better explanation or a better number. The dispute that works is about scope — an operation that should be on the estimate and is not — rather than about the total.

The fleet client has the complaints process too and will rarely use it, because it has something faster. It has an account, a broker with volume, a service conversation, and a renewal. Escalation for a commercial client is relational and it works, which is also why the fleet’s small claims go unchallenged: the leverage exists and nobody spends it on a mirror.

Data, and who else sees it

For the personal client the questions are what the insurer collects, how long it keeps it, whether it consults shared industry databases, and what a flag on this claim does to the next one. Most clients assume less sharing happens than does.

For the fleet client the same questions arrive with an employment dimension attached. Telematics that corroborate a claim also record a driver’s behaviour, and a fleet that hands over data to support a claim has handed over data about a person who works for it. That has consequences under local data-protection and employment rules that a broker should flag rather than advise on. What those rules require — consent, notice, something else, or nothing — is not stated here for any jurisdiction, and flagging it early is worth more than guessing it correctly.

Reading an insurer’s claims stack without a technical briefing

Nobody will show a broker their architecture, and it is not necessary. The evidence is in what the insurer sells and what it writes.

Look at the claim journey in the marketing: app-first notification, photograph capture, a settlement figure promised in a stated period, a network repairer, an instant payment method. Each of those implies an automated path behind it, and the more of them there are, the higher the threshold is likely to be. Look at the wording: does it require photographic evidence, does it reserve the choice of repairer to the insurer, does it say anything about how a valuation is determined. Look at how the insurer talks about claims performance — an insurer publishing touch-time and straight-through rates is telling you what it optimises. And look at how its declines read, if you have seen any: a decline letter that states specific claim facts comes from a different operation than one that states a conclusion.

The broker’s own position in this

There is a placement question underneath all of it that is easy to leave unasked. An insurer with an aggressive automated path is often the right recommendation for a personal client: the claims that client is likely to have are the claims that path handles well, and speed is what they want. The same insurer can be the wrong recommendation for a fleet running older vehicles on urban routes, where the losses are frequent, the damage is often structural behind a bumper, and an under-scoped first estimate costs vehicle days rather than money. Those are opposite conclusions from the same feature of the same insurer, and a broker who has a single view on claims automation across the book is not advising either client properly.

The record matters here too. A recommendation made partly on claims handling should say so in the file, in the terms it was actually made — the speed, the documentation demands, the escalation route — because a client who later has a badly handled automated claim will ask what they were told, and the answer should exist in writing before the question does.

None of that tells a broker where the threshold sits. That number is not published by anyone, and a client who wants to know how their claim will be decided cannot find out.

What a broker can do instead is decide what to put in writing before a claim exists — the inspection question, the written-reason question, the escalation function, the retention period, the telematics question — and place business partly on the answers. A client who has to discover how an automated claim is handled while having one is being advised too late, and the person in the chain who could have asked earlier is the broker.