Ask the Deciding Question Before the Claim Is Filed

A homeowner opens a warranty request and types six words: "Brown stain on the hall ceiling."

Anyone who has worked a warranty desk knows what happens next. The claim lands in the queue. Someone opens it, reads it, and immediately knows the one thing they need and do not have: what is above that spot. A bathroom means one trade and one provision. A roofline means another. Nothing above it but attic means a third. Until somebody knows which, the claim cannot be assigned, cannot be scheduled, and cannot be ruled on.

So the office writes an email. The homeowner answers on Thursday, or next week, or after a second reminder. The claim sits. When the answer finally arrives, it is one sentence long, and it was one sentence the homeowner could have written in four seconds on the day they filed.

That gap, between the moment a claim is filed and the moment somebody asks the obvious question, is one of the quietest costs in warranty administration. It does not show up as a defect or a complaint. It shows up as cycle time, and as a homeowner who is convinced nothing is happening.

The homeowner will never be closer to the answer

At the moment somebody files a warranty request, they are standing in the house. They can see the stain. They can walk upstairs. They can put a hand on the wall and tell you whether it is damp right now.

Three days later, at a desk, replying to an email, they are guessing from memory about a room they are not in. The answer you eventually get is worse than the one you could have had, and it cost a week to obtain.

The question did not get harder. It got asked at the worst possible moment.

This is not advice about how you should run intake. It is what IQ now does inside Cascade Connect, and the rest of this post is the specifics.

What IQ actually does when a homeowner types

The homeowner describes an issue in the request form and presses "done describing this issue." IQ reads what they wrote and works out the handful of physical causes that would explain it, then asks the one or two questions whose answers separate those causes from each other. On a ceiling stain that is:

What is directly above that part of the ceiling? Is the stain damp to the touch right now, or does it only darken after rain?

Those questions are not on a list somewhere. Nobody wrote a ceiling-stain rule. They fall out of the possible causes for what this homeowner actually described, which is why the same machinery handles a sloping floor, a sticking window and a sewage smell without anyone maintaining a script for each.

Three things about how it behaves are worth stating plainly, because they are where the design decisions live.

It asks what the homeowner can answer, not what a technician could. A question qualifies only if somebody standing in their own house can settle it by looking, measuring or remembering, in under a minute. If the leading causes can only be separated by opening a wall, IQ asks nothing and says so, rather than inventing a question to look thorough.

It keeps their words. The original description is stored verbatim and is never overwritten. What the homeowner reported and what they answered stay separable in the file forever, which matters on the one day somebody reads that file adversarially.

Skipping is a real option. Leave the boxes blank and send. The claim files exactly as it would have before. A homeowner pressured into guessing would put a wrong fact into their own claim, which is worse than no fact at all.

The best outcome is no claim at all

Sometimes the deciding question is not diagnostic. It is something the homeowner can go and try.

Several outlets go dead in a bathroom and a garage. The homeowner reports a wiring fault. IQ's first question is whether the GFCI has been reset, because in a meaningful share of those cases the answer is that it has not, and pressing one button ends the matter.

Price that out the ordinary way. A filed claim, a queue slot your rep pays attention to, an email, a week of the homeowner assuming they have been forgotten, a dispatch, a truck roll, and a subcontractor invoice for a service call that found nothing wrong. All of it to arrive at a button that was always within arm's reach.

So IQ marks that kind of question differently. When the answer requires going and doing something rather than just looking, the form offers to hold everything they have written so they can go and check. If it turns out to be sorted, there is nothing else for them to do and no claim to process. That outcome is better for the homeowner than any repair you could have scheduled for them.

It is also the opposite of how warranty software is usually pitched. Processing claims efficiently is the wrong ambition. The claim you never have to open costs nothing to assign, nothing to schedule, and nothing to defend.

The thing IQ refuses to do

There is an obvious shortcut here, and it is worth explaining why we do not take it, because from the outside the shortcut looks like a better product.

Reading "brown stain on the hall ceiling," it would be easy to hand the homeowner a tidier version: "Brown stain on the hall ceiling, below the upstairs bathroom, consistent with a supply line leak." It reads beautifully. It demos even better. And a homeowner shown that text will very often click accept, because people accept pre-filled text that reads well.

It is also a fabrication. Nobody said there was a bathroom above that ceiling. It is simply the most common cause, and a model that has been asked to be helpful will supply it. The moment it is accepted, that guess is a fact on the record in the homeowner's name, a determination gets made on it, and if the claim is ever disputed, your file shows the homeowner asserting something they never said.

So IQ does exactly two things with a description. It restructures what was actually written, changing no facts. And it asks for what is missing, as a question with a blank in it. Every sentence in the final text traces to something a human typed. The homeowner's answer is evidence. A guess would be a liability you generated yourself.

That is a constraint we chose, and it is worth asking of anything you evaluate in this category: does it ever write a fact the homeowner did not state?

What this changes on the desk

The claim arrives with the answer already attached. The person who opens it is not composing an email, they are making a decision. The homeowner is not waiting on a question they could have answered instantly. Your trades are not being dispatched to discover something a sentence would have told you.

And the file contains what the homeowner actually said and what they actually answered, with nothing invented in between.

The fastest warranty claim is the one that never needed to be filed, because somebody was asked the right question while they were still standing in the room.

Where this fits

Intake is the front door, and it compounds. A claim that arrives complete gets assigned faster, ruled on once, and closed without a second conversation. Across a book of builders that stops being about individual claims and starts being a staffing number.

Asking the question is one piece of what IQ does. It also rules on coverage against that home's own warranty document and no one else's, drafts the homeowner response, routes the work to the right trade, and learns from the calls your team actually makes. People decide at every step. Patent pending.

If you administer warranty for several builders at once, the platform built for that operator is Cascade Connect for warranty-management companies. If you are a builder running your own program, start with the platform and pricing, see it side by side with a legacy portal, or read how the IQ claims engine works.

Or skip all of that and request a demo. Bring the vaguest, most useless claim description sitting in your queue right now, the one that has been waiting a week on an email. That is the one I want to show you.