18-day redesign sprint · AI surveillance

Making surveillance feel simple.

Spacture.ai helps teams make sense of cameras, incidents, people and analytics from their phone. The app worked — but it was a collection of screens, not a system. I redesigned it into one.

Role
UX/UI & product designer
Timeline
18-day sprint
Project
Mobile app redesign
Focus
System & hierarchy
Live camera feed screen Home dashboard screen AI search screen priority alerts first natural-language search
At a glance

An AI surveillance app that worked — but felt like a collection of screens: inconsistent visuals, weak hierarchy, fragmented navigation.

I treated it as a system problem, not a screen problem, and rebuilt navigation, hierarchy and components around how owners and security teams actually scan information.

Outcome · 20+ screens rebuilt on one navigation model, one component set and one visual language — in 18 days.
Redesign sprint on an existing product.
Role
UX/UI & product designer
Timeline
18-day sprint
Users
Store owners · security staff · analysts · IT teams
Delivered
UX audit · visual system · onboarding & camera pairing · dashboard, feeds, AI search, detections, analytics, watchlists
Add team and toolsWho you worked with (if anyone) and the tools used.
The challenge

Eighteen days to rethink an app that already worked.

Rather than redesign isolated screens, I treated the product as a system and asked five questions:

  1. How should people move through the app?
  2. What information deserves priority?
  3. Where are the inconsistencies?
  4. How can the interface become cohesive?
  5. How can a complex surveillance product feel easier to use?
The goalNot to make the app look newer — to make it clearer, more consistent and easier to navigate.
The product's world, simulated: hours of footage from every camera, distilled into a short list of moments. Tap a camera or a filter to focus it.
DAY 1DAY 18 PHASE 01Define · goals PHASE 02Audit · friction PHASE 03Explore · visual system PHASE 04Design · mobile specs PHASE 05Refine · polish

Five phases, run back to back.

Existing experience audit

The app worked. The experience needed work.

Issue 01

Inconsistent visual language

Screens didn't feel like parts of the same product.

Issue 02

Weak information hierarchy

Important information competed with secondary content.

Issue 03

Fragmented navigation

Hard to tell where you are or where to go next.

Issue 04

Inconsistent components

Controls, cards, buttons and surfaces had no shared system.

Issue 05

Dated visual treatment

Didn't meet expectations of a modern AI/security product.

SYSTEM BEFORE AFTER
Add the old Spacture screens3–5 “before” screens, ideally the same views as the redesign (home, feed, detections). They'll sit side by side with the new ones. assets/v2/spacture/before/

The problem wasn't one bad screen. It was the lack of a system connecting the screens.

This shifted the project from screen redesigns to a system-level redesign.

Target audience

Four roles, one shared need.

Overview

Retail store owners

Understand what's happening without inspecting every detail.

Action

Security personnel

Fast access to cameras, incidents, people and next actions.

Analytics

Business analysts

Structured data that reveals operational patterns.

System

IT & security teams

Visibility into system status, roles, config and health.

Design implicationComplex information must become easy to scan and act on.
Objective 01

Make navigation easier

Clearer relationships between the major product areas.

Objective 02

Establish a visual system

One language for type, colour, surfaces, controls and interaction.

Objective 03

Make complex info scannable

Show what matters without making people process everything.

Visual tone & typography

Dark, technical, premium — in service of clarity.

Typography as hierarchy
2,481Page title / primary value
Live camerasSection header / action tag
Cam 04 · Aisle 3 · 14:02:11Metadata & timestamps
LIVEALERTOFFLINEStatus states
Dark surfacesIndigo accentsHigh-contrast typeRounded cardsClear status states
Base surface
Card surface
Indigo accent
Live / scan
Alert
Primary text

Colour roles in the redesigned UI.

Swap in the real colour tokensThese swatches are approximations — paste the hex values from your Figma styles.
Add the component library sheetButtons, cards, status chips, nav — one Figma export showing the system. assets/v2/spacture/components.webp
Entry & hardware flows

From first launch to a paired camera.

Onboarding & identity

Splash
01Splash
Get started
02Get started
Sign up
03Sign up
Log in
04Log in
Create profile
05Create profile

Store setup & camera pairing

No store added
01Empty stateNo store yet
Add store
02Add store
Store added
03Store added
Add camera
04Add camera
Camera pairing
05Pairing
QR scan
06QR scan
Edit store
07Edit store
Product area breakdowns

One product. One visual language.

Pick an area to see how the same rules carry through.

01Camera content 02Context 03Status 04Action EVERY FEED FOLLOWS THE SAME READING ORDER
Major system changes

What actually changed.

NAVIGATION
Unpredictable section relationships.
Clear structure, predictable hierarchy.
HIERARCHY
Elements fought for attention.
Primary, secondary and metadata separated.
COMPONENTS
Inconsistent repeated patterns.
One reusable component system.
VISUAL LANGUAGE
Screens felt disconnected.
Colour, type, surfaces and spacing work together.
OVERALL
A functional collection of screens.
A cohesive mobile product.
Key decisions

Four trade-offs I made on purpose.

Prioritization › volume

Most important first, depth on demand.

Consistency › novelty

Familiar patterns mean less to learn.

Scan, then investigate

Quick overview, smooth path into detail.

System serves content

Dark surfaces and indigo exist for clarity and recognition.

What I learned — AI products aren't only about AI.

Systems thinking matters

One screen is easy. Making 20+ feel like one product is the real work.

Hierarchy over density

A product can hold a lot of data without showing it all at equal priority.

Consistency creates usability

A unified component system cuts the interpretation users have to do.

Constraints force prioritization

Eighteen days meant focusing on changes with systemic value.

The goal wasn't to make Spacture look newer. It was to make the complexity of surveillance feel simpler.

Next case study

KiranaPro →