← Back to work
Shipped Design Lead — Direction Enhancement

AI Color Match

An enhancement to Lowe's Paint Color Wall: upload a photo of a color and get an instant match from Lowe's paint line — designed and directed with the team, not by me hands-on.

Context

A design direction role, not a hands-on one

Paint Color Wall was already live. As Design Lead, I directed design for its newest enhancement — an AI-powered color match — reviewing and shaping the work of Sr. Designers Lauren Solomon (initial release) and Oliver French (recent work), while the Visualizer team built the underlying match technology in-house.

The Problem

Matching a color is still mostly guesswork

Customers trying to match an existing paint color relied on browsing the color wall manually, bringing in a physical swatch, or using a generic color-matching app — all slower and less reliable than they should be for a task this common. Solving it well meant fewer mismatched purchases, higher conversion on paint, and less friction on a routine task.

Research

Built on existing usage, validated for the new case

This built on existing Paint Color Wall research and usage patterns, layered with validation specific to the new AI match capability. One early, deliberate finding shaped the whole feature: we chose not to brand it as "AI" at all. Customers don't care about the technology — they care about finding their color. The feature is framed entirely around that outcome.

Design Decisions

Extend the existing experience, don't replace it

01 A parallel path, not a replacement

Visual Color Match lives as its own tab alongside Interior, Exterior, Cabinet/Furniture, My Rooms, and Color Wall — customers who just want to browse or search by name lose nothing; the AI path is additive.

Visual Color Match sits as one tab among several — never the only way in.
02 One required input, one optional

Step 1 (an inspiration photo) is required; Step 2 (a photo of the customer's own room) is optional. Uploads are restricted to interior images, with error states for low-quality or unusable photos — keeping the required step simple while still offering the payoff for those who want it.

Step 2 is explicitly labeled optional — visualization is a bonus, not a blocker.
03 One confident match, visualized immediately

The result is a single named color match, shown applied to a real room instantly — with Similar Colors offered as clearly secondary alternatives, and a one-tap save to "My Color Palette" so the match isn't lost.

The match becomes real the moment it's returned — shown in a room, not just as a swatch.
Outcome

Impact

⚠ AI match is too new for its own metrics yet — numbers below reflect overall Color Wall performance, for business context.
224K
Monthly Users
17%
Visualizer → Configuration
New
AI match — early days

Visualizer-to-Configuration tracks how often a customer who interacts with the visualizer — finding a color they like — moves forward into product configuration or adds to cart. At 17%, it's a meaningful signal that visualization drives real purchase intent, not just browsing.

Reflection

Directing, not just designing

Collaboration fuels ideas — but leading design direction sometimes means reeling a designer back in. There were moments reviewing early concepts where I had to steer the work toward something simpler, given scope and time. As a designer, my instinct is to think big; as a lead, part of the job is knowing when to trade that instinct for something concise and shippable.