Underwriting Is Consolidating Around Agentic AI. Total Loss Is Still Manual.

This month the agentic underwriting tools stopped being separate launches and started becoming a single stack. Total loss valuation for AMCs is still a manual call and a spreadsheet. Here is what actually shipped, why the gap is now about the record and not only the speed, and the one move to make.

Earlier this month we made two arguments on this blog. First, that underwriting was racing ahead while total loss stayed manual. Then, a fortnight later, that agentic underwriting had gone from argument to live product. Both described a widening gap.

The story has moved again, and the shift is worth naming precisely. The agentic underwriting tools are no longer a scattered set of launches. They are consolidating into a single stack. That is a different kind of signal, and it changes what total loss teams should do about it.

What shipped, and why it counts as consolidation

Three things landed in the first half of July, and read together they point one way.

Duck Creek acquired Send, an AI-native underwriting orchestration engine, and described the result as an agentic underwriting-to-core platform. Send carries the submission and its data. Duck Creek runs the policy core. Put together, one submission can travel from broker email to bound policy inside a single governed system, with agents doing the orchestration in between.

Alongside it, Hyperexponential introduced hyperoperator, an agent built to take a broker submission from first email to a priced, quoted risk inside a carrier's own appetite and authority controls. It is entering controlled customer rollout across the rest of 2026.

And Cytora, on the back of its Zurich deployment, added the production proof point. Zurich reported manual triage time falling by around 80 percent, and Cytora's chief executive said agentic AI is now "beyond pilots." That is not a demo. That is live.

The detail that matters is not any single product. It is the direction. An acquisition is a bet that the pieces belong together. When a core vendor buys an orchestration engine to build one underwriting-to-core stack, the market is telling you that agentic underwriting is no longer an experiment bolted on the side. It is becoming the default architecture.

A launch is a tool you can choose to adopt. Consolidation is the shape the whole market is settling into. Tools come and go. Architecture sticks.

Why underwriting went first

None of this is an accident of who had the better engineers. Underwriting went first because it is the part of the chain best suited to agents.

A submission is structured. A broker email, a schedule, a set of risk factors, a pricing model with defined appetite and delegated authority. The rules are written down. The volume is high and repetitive. That is precisely the profile where an agent earns its keep. Lots of similar decisions, a clear rulebook, and a measurable output at the end. Feed it the submission, let it triage and price against the model, put a human on the exceptions.

Total loss is not shaped like that, and that is exactly why it has been left until last.

Why total loss is still the manual link

A total loss settlement is a judgement wrapped in a negotiation. What is the vehicle actually worth. Which valuation guides apply and how do you weight them. Is there a pre-accident condition adjustment. Has the customer got a point about the comparable listings. What will this insurer actually accept, and where is the room. None of that reduces cleanly to a rulebook, because half of it is contested and the other half is customer-specific.

So while underwriting has been handing structured decisions to agents, total loss has stayed where it always was. A skilled person, a set of valuation guides, a spreadsheet, and a phone call with the insurer. The judgement in that call is often excellent. The record of it is thin.

The gap is now about the record, not just the speed

The earlier pieces framed the gap as speed. Underwriting fast, total loss slow. That is still true, but consolidation adds a sharper edge. The gap is now also about the quality of the record each side produces.

On the underwriting side, the file is machine-built. Structured data, every step logged, the whole decision reconstructable on demand. On the total loss side, the file is still hand-built, and the record of the judgement is thin.

That mismatch does not matter until the two files meet. In a valuation dispute, in a Consumer Duty file review, in an FCA thematic look at claims handling, the side with the clean machine record looks defensible and the side with the phone notes looks like it is catching up. The judgement can be sound and still lose the argument on the paperwork.

What the gap costs AMCs

Three costs, and they compound.

Expectation drift. When an insurer can move a submission from email to bound policy in minutes, a total loss settlement that takes days of manual work looks slower than it did, even though nothing about it actually got slower. The same insurer's claims team grows less patient with the cycle time.

Credibility on the record. A valuation challenge backed by a structured, evidenced file carries more weight than the same challenge backed by a handler's notes. As the underwriting side sets the standard for what a clean file looks like, the total loss file that cannot match it starts every negotiation a step behind.

Handler time in the wrong place. When the routine parts of a total loss file are still assembled by hand, the skilled person spends their hours gathering and formatting rather than valuing and negotiating. The scarce judgement gets less of the week, not more.

The one move to make now

Do not try to automate the valuation. That is the wrong lesson to draw from underwriting. The judgement and the negotiation are the job, and they should stay with your people. It is also the reason total loss is defensible. You cannot hand it to an agent and walk away.

The move is narrower and more useful. Bring your total loss file up to the same grade of record the underwriting stack now produces by default. Structured evidence in one place. Every comparator, adjustment and decision captured as you go, not reconstructed afterwards. A timestamped trail that stands up in a dispute or a review without anyone rebuilding it from memory.

That is a single, concrete change, and it is the highest-return one available in total loss right now. It does not chase underwriting's speed or copy its model. It closes the part of the gap that consolidation just made visible, which is the record. Speed follows a clean file. So does credibility.

Where TotalSettle fits

TotalSettle was built for this exact gap. It does not replace the settlement judgement, because that judgement is the job. What it does is take the manual weight off the file around it, so the evidence is assembled in one place, the valuation challenge is structured and grounded, and every file leaves a clean, timestamped record without anyone rebuilding it after the fact.

As underwriting standardises around a machine-grade trail, that is what lets total loss meet it on the record rather than trail it. Underwriting could hand the structured decision to an agent. Total loss cannot, and should not. But it can stop being hand-built from scratch every time, and that is the difference between feeling the pressure from automated underwriting and being ready for it.

TotalSettle is the total loss settlement portal built for AMCs. Faster settlements, cleaner audit trails, fewer FCA headaches. Book a demo at totalsettle.com.