Foyr | Client - Smart Trade AI
2026
From vague complaint to qualified job
Homeowners usually know something's wrong before they know what's exactly wrong. Smart Trade's AI chatbot turns a vague complaint into a structured, prioritized job a service business can act on.
Context
Home repairs start with a guess, not a diagnosis
Most homeowners can describe a symptom to a problem like a strange noise, a leak, no hot water, but not the actual cause. Traditional service intake assumes the opposite: a caller who can name the problem, answer a checklist of technical questions and know which kind of technician they need. That mismatch is where most of the friction in home services actually starts, before a technician is ever booked.
Smart Trade, built by Foyr, is an AI-powered home-service platform that sits between the homeowner and the service business. A homeowner describes their problem in plain language to an AI chatbot, which turns that conversation into a structured, qualified job a manager can act on.
Challenge
Ways a normal service call goes wrong
Traditional intake breaks down in a few specific, repeatable ways.
Homeowners describe symptoms, so a lot of context gets lost in translation as it passes from customer to employee to technician.
Manual intake is repeating the same questions on every call, and missed or unanswered calls during busy or after-hours periods are a well-documented pain point across the home services industry, one of the direct reasons Smart Trade wanted an always-available AI front door instead of a phone line.
Not every conversation is a real lead, some are just information-seeking and treating them the same wastes a manager's attention.
Solution
Two interfaces, one connected workflow
The AI Chatbot
Repair vs. Replace comparison with estimate - The AI identifies the exact unit from a photo, lays out repair against replacement side by side with a plain-language recommendation and by choosing one, the AI leads straight into the full estimate, cost range, what's included and deposit terms. All the technical work happens before this point without the need for the homeowner to self diagnose the issue. The homeowner instead just has to make a decision.
The Backend Dashboard
All Entries dashboard - Every incoming job with an AI-generated confidence score, a suggested action and a status, scannable without reading a single full conversation transcript.
Assign Technician - Technicians ranked by expertise match, availability and distance, with a Recommendation of what's the best expertise match, available in the preferred slot, lightest load in the zone. The manager can still Hold, Reschedule or Override. The AI makes a suggestion but it doesn't make the final call alone.
Schedule and dispatch board - A week/day/list view of every visit, unassigned jobs visibly flagged in a different color so nothing slips through. The operational counterpart to the chatbot's conversational simplicity, dense on purpose, built for a manager who wants to scan information fast.
Key Decisions
Design decisions made inside a structure defined by the Product Manager
Scoping note: The Product Manager defined what SmartTrade is and how the three modules connect. Design's job was how each interface works and feels within that. The decisions below are written to reflect that boundary, not to claim structural or strategic choices.
Two entirely different interfaces for two audiences, not one shared - The homeowner experience needed to be conversational and simple. The manager experience needed to be dense and scannable.
The estimate was designed to explain reasoning, not just show a number. Product scoped the estimate to exclude payment and accounting. Within that scope, the design decision was to still show what's included and why the cost is required.
The AI recommends, the manager decides - On technician assignment, the AI ranks every technician and explains its top pick directly. Hold, Reschedule and Override sit next to the recommendation. The cost: showing the reasoning takes more screen space and requires the AI's logic to be explainable in plain language, which is more work than a simple auto-assign button would have been.
Reflection
Designing for clarity, confidence, and everyday financial workflows
One of the biggest things I learned on this project was how to carry forward context someone else had already built, instead of starting research from scratch myself. The company had already put together the spec sheet and done the contextual research before I joined, so my focus was fully on the design decisions themselves.
This was also one of the first times I got to work hands-on with a properly built, established design system. The company already had other dashboard and chatbot projects before this one, so the system were already in place and I got to design within that instead of building one from zero.