Selling a policy in one tap creates an expectation the claim will feel the same. It usually doesn’t — and the gap is where embedded insurance will earn or lose its reputation.
Saudi insurers paid out SAR 53 billion in claims in 2025 — the Insurance Authority's number, up ten percent in a year. Behind the aggregate are millions of individual moments where insurance stopped being a checkout card and became the actual product: a crashed car, a hospital admission, a cancelled trip. Embedded distribution has spent its energy perfecting the moment of sale. Its next reputation will be built at the moment of claim.
The expectation embedded selling creates
A customer who bought cover in one tap, inside a brand they trust, reasonably assumes the claim will live in the same place. When it doesn't — when the marketplace that happily sold the policy answers a claim enquiry with a phone number and a shrug — the experience gap reads as betrayal, and it lands on the platform's brand, not the insurer's. This is the asymmetry platforms underestimate: the insurer carries the legal liability for claims, but the platform carries the reputational liability for the whole journey it chose to embed.
Who actually owns the claim
Legally, the underwriting insurer: assessment, approval, payment and the regulator's scrutiny of all three belong to the licence holder, and no platform should want otherwise. But "the insurer owns the claim" cannot mean "the platform exits the story." The workable division we've converged on with partners:
- The platform owns awareness and access: the customer can find "my policy" and "start a claim" where they bought it, without hunting for an insurer portal they never chose.
- The infrastructure layer owns the plumbing: routing first notice of loss to the right insurer with the policy context attached, tracking status, and surfacing it back to the platform's UI.
- The insurer owns the decision and the payment — visibly, so accountability stays where the licence is.
Design principles for the claims moment
Pre-fill works here too. The same transaction data that made the quote one screen makes first notice of loss one screen: the policy, the vehicle, the customer are already known. Asking a claimant to re-type what the platform already holds is the same sin at a worse moment.
Status beats speed you can't promise. Platforms cannot promise an insurer's settlement timeline. They can promise the customer never wonders what state their claim is in. Visible status — received, with the insurer, decision made, paid — converts an anxious wait into a process.
Never editorialise a decline. When an insurer declines, the platform's job is clean delivery of the insurer's reasons and the escalation route — not defending the decline, and not distancing itself with "that's the insurer's decision." Both destroy trust in opposite directions.
The honest limits
An embedded surface cannot fix a slow adjuster, and a platform should not accept service-level promises its insurer panel hasn't signed. The practical move is upstream: make claims performance a panel-selection criterion with the same weight as price. A cheaper premium from an insurer with a slow claims desk is margin borrowed against your own brand.
Why this is worth building now
Most embedded programmes are young; their books are only now aging into meaningful claims volume. The platforms that treat claims UX as a launch-day feature — not a year-three retrofit — will hit that maturation with loyalty instead of churn. The one-tap sale opened the relationship. The claim is where the customer decides if it was real.