Design the Wrong Answer Challenge

What You'll Be Doing

Tallgrass makes dispatch software for residential HVAC and plumbing contractors. Their mobile app, Tallgrass Field, is used by about 38,000 technicians.

Four months ago they shipped Snap Quote. A technician photographs a piece of equipment; a model identifies it and drafts a repair quote. Here is the screen the technician sees today, described exactly:

Header: “Snap Quote” Below the header: the photo the technician just took, at about a third of the screen height. Below the photo, in large bold type: Carrier 58MVC — 80% AFUE gas furnace Installed approx. 2007 · Est. remaining life 2–4 yrs Below that, a form, pre-filled and editable: Diagnosis: Failed inducer motor Parts: Inducer motor assembly — $412 Labor: 2.5 hrs — $340 Total: $752 At the bottom of the screen: one full-width blue button, Send Quote to Customer. There is no other control on the screen except a back arrow.

Two facts from the Tallgrass product team:

On an internal set of 900 photos, the model identifies the correct make and model 71% of the time. When the photo is taken in poor light or at an angle — which is most of them — it drops to about 58%.
Since launch, 1 in 9 quotes sent through Snap Quote has been followed by a correction call from the contractor to the homeowner. Before Snap Quote, quotes were typed by hand and the correction rate was 1 in 40.

The technician is standing in a basement. They have gloves on. They have one bar of signal. They are being paid per completed job.

Redesign this one screen so the technician catches the model when it is wrong.

Constraints to consider

You may not add a second screen or a confirmation step that requires a separate tap-through. Everything you change happens on this screen. Techs are paid per job and any added step gets skipped or gamed.
You cannot change the model or its 71% accuracy, and you cannot require better photos — techs will not retake them.
The equipment database is read-only to your team and it does not store a confidence score per field. If your design shows confidence, you have to say where that number comes from or design something that does not need one.
One-handed, gloved operation, outdoors or in a dark basement. Any control smaller than a thumb or any interaction requiring two hands does not exist.
Fifteen minutes. One screen, annotated. Not a polished file.

AI Usage Guidance

We expect you to use AI tools. We evaluate how you use them — not whether you use them. Evidence of iteration, redirection, and critical evaluation scores higher than a polished output with no process documentation.

The single highest-signal indicator: your video answer to the mandatory AI question. If you cannot name a specific moment where you redirected AI output, evaluators will assume you did not.

Mandatory AI question for your video: Walk me through one moment where you disagreed with, pushed back on, or redirected what the AI gave you — and what you did instead. Name the specific moment. Explain what the AI produced that didn’t meet the bar, what you did differently, and why.

Communication clarity matters for this role. We assess structure, stakeholder awareness, and ability to simplify complex ideas — not presentation style or accent.

A note on tools and fidelity. We are not scoring visual polish and we are not scoring your tool. A photograph of a pen-and-paper sketch, taken on your phone, scores exactly the same as a Figma frame. What we score is whether the annotations tell us what you decided and why. A rough sketch with eight annotations beats a beautiful screen with none. Do not spend your fifteen minutes on finish.

Submission: Upload each deliverable as a separate file directly on the Provn platform.

What You'll Accomplish

Identify where an interface silently invites a user to trust output that is often wrong

Design a correction affordance that fits a real physical and economic constraint

Express model uncertainty in an interface without access to a confidence score

Annotate a design so a reviewer can follow the decision without a live explanation

Explain a design decision to a cross-functional partner in under four minutes

How Your Work Will Be Scored

Interface Design Execution: Redesigns the screen within the stated physical and structural constraints, with annotations that make each decision legible Designing for Uncertainty: Changes what the interface implies about how much the technician should trust the output, without requiring a confidence score Video Walkthrough & Design Communication: Walks the artifact in the user’s order and explains why each decision was made AI Fluency: Shows directed, iterative, critically evaluated AI use

What to Submit

File 1 — Annotated Screen

Any File · one image, PNG or JPG, or a single-page PDFRequired

Format: no restrictions

Your redesigned Snap Quote screen with at least six numbered annotations. Each annotation says what you changed and why in one line. A photographed paper sketch is fully acceptable.

File 2 — README Document

DocumentRequired

Format: .pdf, .doc, .docx, .rtf, .txt, .md

Three required sections:

Section A — Written analysis: name the single riskiest moment on the original screen — the exact place where a technician is most likely to send a wrong quote without noticing — and say what your design does about it. Say also what you deliberately did not change. 150–250 words.
Section B — The one thing you would test: describe one thing you would put in front of five technicians to find out whether your design works, what you would watch them do, and what result would tell you that you were wrong. Two to four sentences is enough.
Section C — AI Usage Log (Mandatory): This is not a trick. We want to see how you work with AI — not whether you used it. In a short section of your README, document your AI collaboration process. For each significant interaction with an AI tool, briefly note: what you asked the AI to help with / what it gave you / what you kept, changed, or rejected — and why. Three interactions documented is sufficient. The log does not need to be exhaustive.

File 3 — Video Walkthrough

Video · 3–4 minutesRequired

Format: .mp4, .mov, .webm

Record as MP4 or MOV and upload directly on the Provn platform as a separate file. No ZIP files, no repo links, no third-party video platforms. Hold your sketch up to the camera or share your screen — either is fine.

Cover: (1) the riskiest moment on the original screen and what you did about it, ~60 sec; (2) walk your annotations in the order a technician would encounter them, ~60 sec; (3) the mandatory AI question, ~60 sec; (4) what you’d design next with more time, ~30 sec. Speak naturally — we’re assessing your thinking, not verbal polish.