Hayley Bance Operational Products
← Contents
03

Operational Products

When technology fits naturally into existing routines, better data becomes a byproduct of better work, not an additional responsibility.

I design software for people working in fast-paced operational environments, where attention is limited and every extra step competes with the work itself.

I believe operational technology only creates value when it disappears into the job already being done—reducing cognitive load instead of adding to it.

Data Collection With a Single Tap

StreamLine Smart Scale

Using AI and computer vision to remove work in the kitchen.

Rather than asking operators to manually record production data, StreamLine used AI and computer vision to automatically identify and label food as it was placed on the scale.

Instead of creating another task, data collection became a lightweight extension of the work already happening.

User testing shaped every aspect of the interface, helping us strike the right balance between a sleek, modern aesthetic and subtle moments of delight. The resulting design language felt innovative enough to build trust in the AI while using rewarding feedback to encourage consistent use throughout the day.

The StreamLine smart scale in use in a commercial kitchen
Fig. 1The StreamLine scale in use.
Diagram showing where the StreamLine scale sits in back-of-house and on the serving line, and the Carryover, Waste, and Prepared modes it captures
Fig. 2Where the scale sits: back-of-house and on the serving line, capturing carryover, waste, and prepared weights.
An operator weighing food with the StreamLine scale during service
Fig. 3The interaction, in service: weigh, confirm, done.
The StreamLine touchscreen mid-weigh, showing a live weight reading and Prepared, Carryover, and Waste mode buttons
Fig. 4The touchscreen itself: a live weight and a single tap to sort it.

The quality of operational data is directly tied to the quality of the frontline experience.

By making data capture nearly invisible, we created a system operators trusted and consistently used, laying the foundation for meaningful operational visibility and future AI-powered decision support.

2M+ lb Food recorded
10× Faster data capture
50% Reduction in food waste across participating kitchens
Flow diagram of the happy-path waste capture process, from weight detection through session confirmation
Fig. 5The happy-path waste flow, mapped.

From Data to Decision

StreamLine Insights Reporting

Helping chefs put production data to use.

Collecting operational data is only half the challenge. Once that data exists, people need experiences that help them quickly understand what matters without interrupting the work they’re already doing.

We used a system of anticipatory touchpoints that delivered information through the right channel at the right moment.

01

During Service

Embedded operational feedback provided immediate visibility while work was happening, helping teams stay on pace without interrupting service.

An embedded feedback badge reading 'Salad Bar ยท 21 Observations today'
02

Before Service

Daily email summaries gave chefs a quick understanding of yesterday’s performance before today’s shift began, making it easy to identify what required attention.

A daily email summary reading 'Good Morning, Chef! Here's what happened yesterday'
03

Weekly Reflection

Weekly Slice n’ Dice reports created space to step back from day-to-day operations, helping chefs identify trends, celebrate wins, and uncover new opportunities for improvement.

A weekly waste report showing total waste this week and a waste monitor trend
04

Deep Exploration in Dashboard

Rather than another dashboard, an AI-native workspace: chefs explore performance visually, or ask an assistant that’s read thousands of production observations and can answer in plain language.

05

Agentic Support

Beyond surfacing insights, we built focused workflows and agentic tasks that let chefs act on what they saw directly. Updating a forecast (previously a manual process spread across separate systems) became a single tap, with the agent handling the update everywhere it needed to happen.

The AI assistant confirming an update was applied successfully
The StreamLine Insights workspace, showing forecast suggestions, location performance, and highest-cost-of-waste items
Fig. 6The AI-native workspace is designed so that specific AI workflows populate “teaser” cards. The user taps to engage deeper with the chat experience.

What deserves my attention right now?

70% Reduction in food waste per menu item
5% Average decrease in cost per meal
$15K Average annual labor savings per kitchen

Workflows That Fit the Work

Sweetgreen Hot Prep Tool

Helping kitchen teams plan and track production throughout the day.

At Sweetgreen, I redesigned a prep planning system that had struggled with adoption by transforming it into a lightweight Kanban-style workflow that matched how kitchen teams already worked.

The Kanban-style cook board, showing a tray card being dragged from the to-do list into the in-progress column
Fig. 7The Kanban-style prep planning workflow, matched to how kitchen teams already worked.

The experience combined production planning with required food safety logging, eliminating duplicate work while helping operators understand what to cook, when to cook it, and whether they were keeping pace with demand.

Flowchart mapping the steps to cook a single ingredient, comparing the path with connected ovens against the additional manual steps required without them
Fig. 8User flow for recording a cooked item.
500% Increase in adoption of a previously underutilized operational tool
$700K Annual waste savings

Workflow design isn’t simply about efficiency.

It’s about creating experiences that respect attention.