Claims and Underwriting – Operational Drag Leaving Margins Behind

Published on: 03/06/2026 Published in: Blog

Operational drag is where margin goes to hide in claims and underwriting

Most insurers do not lose margin in one dramatic moment.

They lose it quietly.

In the extra touch that should not have been needed.
In the rekeyed field that adds no value.
In the avoidable referral.
In the expert reviewing work that should have been routed earlier.
In the claims file that moves too slowly between teams.
In the underwriter chasing information instead of judging risk.

That is operational drag.

And in claims and underwriting, it becomes trapped margin.

This matters because the economics are tightening. EY’s latest UK motor analysis forecasts a 111% net combined ratio for 2026, meaning insurers would pay out £1.11 in claims and expenses for every £1 earned in premium. At the same time, insurers are still being asked to speed up claims, sharpen underwriting, improve service, and make AI useful. KPMG’s latest insurance CEO outlook says carriers are accelerating AI adoption to improve underwriting efficiency, speed claims processing, and strengthen trust.

The problem is that many insurers still look for margin improvement in the obvious places only:
pricing,
headcount,
fraud controls,
vendor management,
expense reduction.

All important.

But a lot of margin is still hiding in the workflow.

A straightforward submission takes too long to quote.
A simple claim picks up complex friction.
Claims and underwriting do not learn from each other quickly enough.
Data arrives too late in the process.
Teams stay busy, but the process still feels noisy.

That is not just inefficiency.

It is a commercial issue.

Because in many insurers, the operating model is still making good people work too hard to get ordinary things done.

The insurers that stand out will not just “automate more”. They will redesign claims and underwriting around a sharper question:

What should move instantly, what should be guided, and where should expertise really sit?

That changes the game.

Simple work should not feel complex.
Guided work should not consume expert time unnecessarily.
Expert judgement should be reserved for the cases that genuinely deserve it.

This is where better triage starts to matter.

Not everything needs the same workflow.
Not everything needs the same level of touch.
Not everything needs the same person involved.

Insurance trends point in exactly this direction, with underwriting evolving toward more intelligent, AI-supported decisioning and “ease of doing business” becoming a competitive differentiator. EIOPA also says nearly two-thirds of European insurers are already using generative AI, though most are still at proof-of-concept stage. That tells us the opportunity is real, but production value still depends on workflow design, governance, and operating readiness.

The other big missed opportunity is the gap between claims and underwriting.

Claims should not just be an operational output. It should improve underwriting.
Underwriting should not just set risk appetite and move on. It should keep learning from claims patterns, leakage signals, and portfolio reality.

The insurers that build that feedback loop faster will protect more margin than the ones that keep treating claims and underwriting as adjacent, not connected.

Data matters too, but only when it appears at the point of work.

Most insurers already have plenty of data. The real issue is timing and usability. If policy context, claims history, fraud indicators, portfolio signals, and next-best-action prompts do not show up inside the workflow, people fall back on rework, chasing, and manual judgement to fill the gaps.

That is why better analytics spending is still happening in the market. Reuters reported strong demand for Verisk’s analytics in underwriting, claims processing, and fraud detection, which is a good signal that insurers are still investing where better decision support can reduce friction and protect economics.

And this is where AI needs a more disciplined role.

The question is not “Where can we add AI?”

It is:

where can AI reduce friction without reducing trust?
where can it support decisions without obscuring accountability?
where can it improve cycle time, leakage, or service quality in a measurable way?

KPMG says trust remains central, especially where pricing, risk profiling, and claims payouts need to be fair and explainable. That is why the best AI use cases in insurance are usually not the most theatrical ones. They are the ones where workflow, ownership, and outcome are already clear.

The insurers that pull ahead will not just have better average turnaround.

They will feel different:

underwriters spending more time on judgement
claims teams spending less time gathering and more time deciding
brokers experiencing less friction
customers repeating less information
leaders seeing a clearer link between workflow quality and margin quality

JD Power’s 2026 property claims study found claims were resolved 3.4 days faster than the year before, with faster repair and payment cycle times improving satisfaction. That is the wider point. When drag comes out of the workflow, the benefit is not only internal efficiency. The market feels it.

So the next margin story in insurance is not just about pricing sophistication or bigger transformation language.

It is about finding where the process is still too manual, too noisy, too late, or too dependent on expert effort for work that should already be flowing cleanly.

That is where trapped margin sits.

And the insurers that stand out will be the ones that know how to remove it:

route earlier
decide faster
use data sooner
reserve expertise for the work that deserves it
connect claims and underwriting more tightly
apply AI where trust can still be defended

That is not just efficiency.

That is competitive advantage.

References:

You may also like