The ninety-two days
nobody can see.
Until now.
Ninety-two days is your own average to fully settle a house claim. Most of those days are honest work — assessors, schedules, the Natural Hazards portion — but your customer sees none of it. This is a concept for making the wait visible, valuable, and shorter-feeling: the claim prepared in the open, and every decision still made by a person.
“An average of 92 days to fully settle a house claim.”
Consumer NZ insurance research · claims settled Sept 2022 – Aug 2023The morning after the storm
Water came through the ceiling overnight. Lodging the claim in My Tower takes minutes — your own site says online claims save an average of five minutes, and photos help validate twice as fast. Then the part nobody designed: the silence between lodgement and a decision, measured in weeks, while the customer refreshes an app that has nothing new to say.
The 2023 floods and Cyclone Gabrielle brought Tower 8,900 claims — your chief claims officer called it five years of large-loss claims in about two weeks. Half-year profit to March 2026 more than halved on $18.5m of large-event costs, while 15,000 new customers arrived in a year. The next event is a when. The claims experience during it is a choice.
What happens now
The claim is real work from the first hour — triage, photos, the assessor diary, the Natural Hazards portion. All of it invisible, so every day of it reads as neglect.
What it costs
In the 2023 surge, 35% of enquiries reaching the insurance ombudsman were about delays, against a 9% norm. The work was being done. Nobody could see it.
What the concept does
Shows the claim being prepared as it happens, pays the wait back as excess credit, asks the one question that changes triage — and stops. A person decides.
Every claim moment, one pattern.
Each of these is a moment your customer already spends with you. Each can carry what the phone above shows: watch the work, earn the wait, answer one thing.
Be relevant — right thing, right moment · Be respectful — short, skippable, quiet by design · Be measurable — real impact, on the record · Be Aotearoa — built here, for how Kiwis actually wait
Every waiting moment in a house claim
Six gaps between the storm and the last tradie leaving, and what a household could see in each.
The evidence gets taken at 2am, while the damage is still exactly as it happened.
The claim shows its working, so a week of real effort stops reading as neglect.
The urgent ones stop waiting in the same queue as everything else.
They stop taking days off work for a visit that has not been scheduled.
A choice with two clear consequences, instead of a form with two boxes.
The longest part of the claim finally has a schedule the household can see.
Four steps, every time.
- The signalSomething in your own systems says a person is now waiting, and roughly how long for.
- The predictionHow long this one will really take — not the average, this one.
- The valueSomething worth their attention goes into that window: an answer, a credit, a thing prepared while they wait.
- The exchangeThey get value. You get a customer who stayed, and one optional answer you have always wanted.
This is the reward layer, running for real on this page. Drag to scratch it. Nothing is collected, nothing is sent — it is here so the idea is something you do, not something you read.
US retailers lose an estimated $37.7 billion a year to customers who abandon because of long queues, split roughly $15.8b to competitors and $21.9b abandoned outright (Adyen with 451 Research). The New York State Department of Labor reported a 48% reduction in hang-ups after introducing virtual queuing. Theme parks moving to virtual queues reported a 36% lift in per-capita spending, because guests spent the wait somewhere that sold something. A Seattle-Tacoma Airport study found 63% of passengers said they shopped or ate more because they saved time in the security queue.
Every figure on this page carries its source. Overseas figures are shown as overseas figures — we have not found a published New Zealand equivalent, and we will not invent one.
Ninety-two days, out loud.
Most of those days are real work — assessors, schedules, the Natural Hazards portion. The customer sees none of it, so it all reads as silence. Press it and watch the same claim run with the work shown.
Every figure here is illustrative except the 92, which is Consumer NZ insurance research, claims settled Sept 2022 – Aug 2023. This is a concept timeline, not a measurement and not a promise.
-
Day 1Lodgedwhat they get todayA reference number and a promise to be in touch.with the wait designedPhotographed room by room while the damage is still exactly as it happened. The evidence an assessor will ask for in three weeks is already attached.
-
Day 2–9Triagewhat they get todaySilence. The urgent claims and the simple ones wait in the same queue.with the wait designedRead against the flood overlay and the policy wording, and sorted by what it actually needs — make-safe today, assessor next week, or nothing at all.
-
Day 10–34Make-safe and assessmentwhat they get todayA booking window and a wait with nothing in it.with the wait designedA time from a real diary, what the assessor will look at, and what the household can do this week that counts.
-
Day 35–61The cover questionwhat they get todayThe longest silence, and the one that generates the most calls.with the wait designedThe cover position drafted the day the assessment lands, with the wording it rests on — held for a person to approve, never sent by a machine.
-
Day 62–84Repair or cashwhat they get todayA number arrives with no working.with the wait designedBoth options side by side with what each means for this house, and where the excess comes off.
-
Day 85–92Settledwhat they get todaySettled, and nobody remembers the middle.with the wait designedA record of every decision and what it rested on, which is also the audit trail.
The claim, rendered
A live panel showing this property against Tower’s own published flood-risk view, the assessment progressing, and which portion is Natural Hazards Commission rather than Tower. The customer stops guessing which queue they are in.
A labelled line, not a chatbot
It answers what stage the claim is at and what happens next, from the claim record rather than a script. It says when it does not know and hands to a person. Every answer is marked as prepared by a machine.
The wait pays for itself
While the assessment completes, one funded slot: temporary accommodation, drying and make-safe, or replacement essentials. It is relevant because the claim says the house is wet — not because someone bid highest.
One question, answered once
Inside the wait: is the house safe to sleep in tonight? An answer to that changes the triage immediately, and it is the one question a survey three weeks later never gets.
The funded slot above shows a category, never a company. assembl has no agreement with any third party named or implied on this page, and would not put a partner's name on a concept before that partner had agreed to it.
One storm. Every customer’s claim.
The same weather event, prepared for four different households. Every draft stops before it sends — and the ones that need a person say so.
A template sends the same thing to everyone and changes the name. What matters here is the case that should get something different — or nothing at all.
The rules beside each draft are placeholders. A pilot replaces them with your actual policies. We have not seen those, and this concept does not guess at them.
The check that runs before a person ever reads it
A storm claim is where a warm template does the most damage. These are the checks — and the ones that block are the point.
Drafts are held against the Fair Insurance Code 2020, CoFI fair-conduct duties, the Natural Hazards Insurance Act 2023, the Fair Trading Act 1986, and the Privacy Act 2020.
Press run to see what it catches — and what it refuses to produce at all.
One room. Your people, your agents, every action on the record.
Every concept on this page would run inside a private room like this: your team and the agents in the same space, drafts appearing with receipts, a named person saying yes. It runs on infrastructure you approve — and what is said in the room stays in it.
receipt sources: your rate card · rules: yours · held for approval
Client information stays in your room. Six months later, “where did this come from?” is one line, not a search.
Yours
Runs where you approve — your own room, your own record. Nothing in it leaves it.
On the record
Every agent action lands as a line a person can read — not a log only a vendor can open.
People decide
Drafts wait for a named person. The room shows who said yes, and when.
This panel is a concept picture, not a live room — a pilot stands up the real one, on infrastructure you approve.
The storm board, before anything sends
Illustrative and fictional. No production access is requested by this concept.
| Cohort | Claims | Prepared | Waiting on | Status |
|---|---|---|---|---|
| Water damage, home liveable | — | Update drafted · assessor booked | Claims manager sign-off | Awaiting approval |
| Natural Hazards portion applies | — | NHC explainer + update drafted | Claims manager sign-off | Awaiting approval |
| Answered “we need help” | — | Accommodation call queued first | Named owner per file | Priority — held from bulk |
| Possible total loss | — | Nothing sent | Claims lead — senior call first | Held — no template |
| Photos missing or unreadable | — | One-tap photo request drafted | Customer | Parked |
Cohort sizes are deliberately blank. This concept has no Tower data and does not estimate how many customers anything affects.
Every draft carries its own working.
Not a log somebody has to go and find. The provenance travels with the work — what it read, which rules it held, who must approve it, and what it refused to do.
This matters most on the day someone asks why it said what it said.
- Artefact
- Claim status update — water damage, home liveable
- Read
- Policy wording (placeholder) · claim status fields (placeholder) · address flood-risk rating band · the customer’s one answered question
- Rules held
- Fair Insurance Code 2020 · CoFI fair-conduct duties · Natural Hazards Insurance Act 2023 · Fair Trading Act 1986 · Privacy Act 2020
- Refused
- No coverage decision was stated by any draft. No repair or settlement figure was calculated or implied. Total-loss and displaced-customer files were held for a person.
- Approver
- A named person. Unsent until then.
- Prepared
- —
- Reference
- —
Why this concept, for Tower, now.
Nothing on this page is modelled, projected or borrowed from a slide. These are your own published facts, and the concept above is built on them.
The same evidence, on one page you can print or send: the one-pager →
What this will never do
It does not send. It prepares, and a named person sends. It does not publish to any channel, commit spend, move money, or make a decision that belongs to a person. It does not pretend to be a person — every draft says it was prepared by a machine and approved by a human.
No production access is requested by this concept.
Scope
One weather event’s house-claim cohort, lodgement to decision-drafted. Six weeks. Not the whole claims book, not assessment itself, not settlement.
Access
Policy wording, your claim-status field names, your update templates, and the address risk-rating band. No customer records. No claims history. No settlement figures.
Scorecard
Updates accepted without rewrite. Held files a person agreed should have been held. Status calls the contact centre no longer received. One claims manager’s honest answer to “would you send this?”
Fail any line of the scorecard and we change the design or stop.
What a pilot actually asks of you.
Written out in full, because the honest version of this is short and most of it is your time, not your money.
One named person
The approver. Every draft stops with them and nothing reaches a customer until they say so. Expect about two hours a week — reading drafts, not managing a project.
One situation, not the business
One weather event’s house-claim cohort, lodgement to decision-drafted. Six weeks. Not the whole claims book, not assessment itself, not settlement.
Read-only access, nothing that writes
Policy wording, your claim-status field names, your update templates, and the address risk-rating band. No customer records. No claims history. No settlement figures. No production credentials, no write permissions, no access to anything the pilot does not need.
Your actual rules
The policies, limits and words the drafts must be held against. Every rule on this page is a placeholder standing in for yours. An afternoon with whoever owns them is usually enough.
Six weeks, then a real decision
Scored on: Updates accepted without rewrite. Held files a person agreed should have been held. Status calls the contact centre no longer received. One claims manager’s honest answer to “would you send this?” Fail any line of that and we change the design or stop. You keep everything drafted either way.
A fixed fee, agreed in writing first
No number is published on this page, because it depends on which situation you pick. Whatever it is, it is agreed before any work starts and it does not move.
Run the pilot.
Say yes to the six weeks above, and add anything you would change first. Nothing here is fixed — most pilots move a line or two before they start, and the notes below are how that happens.
Opens your mail app to assembl@assembl.co.nz with your notes in it. Nothing is collected by this page.
- Six weeks, one situation, scored against a written line
- Read-only access, nothing production
- A named approver — nothing sends without them
- A fixed fee, agreed in writing before anything starts
- Stop any time. You keep what was drafted
Pick a verb.
Not “book a demo”. Any of these is a real next step, and the last one is a perfectly good answer.
What is the one constraint we have got wrong?
Every concept is built from the outside. There is always something about how Tower actually runs that we could not see. One line is enough.
Opens your mail app to assembl@assembl.co.nz. Nothing is collected by this page.