The starting position
What the programme looked like before — volumes, cycle time, where claims were going and what it was costing.
A worked example
A brokerage with no claims team.
In October we spent an afternoon with a mid-sized brokerage carrying a heavy transport and logistics book. No claims team, account managers running claims on top of the broking. This is what we found, and what we put to them.
What we foundWhat we proposed
No dedicated claims team
Account managers ran claims end to end on top of the advice, and the transport weighting meant more claims per adviser than a brokerage that size would normally carry.
The desk sits behind the account manager
They keep first contact, the client relationship and the placement decision. We take ingestion, lodgement, filing, chasing and the files that eat a day each.
The record was split in two
The broking system held the claim number. A separate document store held the file notes and the follow-ups. Everything in between was email and phone.
We sit across both and replace neither
Clara AI reads the mail, summarises it, drafts the lodgement and files the documents against the claim. If anyone has to key something twice, we have built it wrong — that is the test we asked them to hold us to.
No view across the book
Nowhere to see open claims, which insurer was slow, or where a file had stopped. Chasing ran on memory and diary notes.
Every claim carries the insurer’s own clock
Code of Practice timeframes run on every file, and overdue obligations escalate on a named ladder — drafted by Clara AI, sent only once a person has approved it.
Aggregate programmes tracked by hand
Their fleet clients carry a large aggregate deductible, and the erosion position was rebuilt by hand whenever somebody asked for it.
A live aggregate position
Paid, reserved and incurred against the deductible, moving as the claims move, with a forecast of where it lands at the current run rate — early enough in the year to do something about it.
- Pilot scope
- Three or four of their larger aggregate fleet clients — enough volume and complexity to prove it on files that matter, without touching the rest of the book
- Length
- Twelve weeks, with a defined exit, no minimum volume and no multi-year lock-in
- Measured on
- Broker touches per claim, how long acknowledgement takes, and whether the aggregate position holds up at the renewal meeting
What ours will contain
No case study without a number in it.
Plenty of claims case studies describe a process. Ours are written to answer the question a buyer actually has, which is what changed.
What we changed
The pathways, authority limits and integrations we set up, and what we deliberately left alone.
What moved
Cycle time, indemnity spend, days off road, recovery performance and complaints, measured against the starting position.
In the meantime
We will do it on your claims instead.
A case study about someone else's fleet is a weaker argument than a cohort from your own.
Send a month of claims
De-identified is fine. A single month is enough to see where the time and the money go.
We run it through the workflow
The same triage, the same pathways and the same reporting a live programme gets.
You get the findings
Cycle time, leakage points, what we would have flagged at intake and what the reporting would have shown.
Want to be a reference client?
Foundation clients get the terms, the attention and a say in how their story is told — if they want it told at all.