Fusun Uzun
Let’s talk
Fusun Uzun
Loading{{ pct }}%
Fusun Uzun hello@fusunuzun.comcopied to clipboard

Product designer based in Toronto. Over the past decade, I’ve led design work from early exploration through delivery, across consumer products and emerging technologies. My work has ranged from replacing the manual processes behind a home-services marketplace with AI to turning seven days of inquest testimony into a 360° documentary. Both came down to the same challenge: finding a form for complexity without passing it on to the person experiencing it.

01{{ peekTitle }} Work ↓

Selected Work

hello@fusunuzun.comcopied to clipboard LinkedIn
© 2026 Fusun Uzun Back to top
Fusun Uzun hello@fusunuzun.comcopied to clipboard

Home Project Experience Redesign

Designing for the way home projects really unfold

Company
Quartermaster
TIMELINE
2025–2026 
Design complete · Partially implemented
Role
Lead Product Designer
Project workspace on desktop: HVAC installation with a payment due, visits, chat and quote

Overview

Quartermaster connects Toronto homeowners with contractors (Pros) recommended by their neighbours. By 2025, AI-assisted intake and automated matching had changed how projects began, but everything after that was still built around a one-time transaction: quote, book, pay and review, in that order, in a workspace originally designed for desktop and later adapted to mobile. There was no way to save a draft or pause a project, and recurring services weren’t supported, even though they were central to Quartermaster’s ongoing relationship with homeowners.

Real projects didn’t move in a straight line. Cleanings repeat, renovations are paid in stages, quotes get revised and visits get rescheduled. The model had to handle all of that while showing homeowners where their project stood and what to do next.

The technical rebuild gave us the opportunity to change the experience. I led the research and end-to-end design of the homeowner experience across web and mobile, connecting Pro selection, booking, payments and feedback within a more flexible project workspace. The experience was designed to accommodate how projects evolve and how relationships with Pros continue beyond a single job.

Context

From a fixed sequence to a flexible project structure

I structured quotes, visits, payments and chat as independent activities within a shared project. What homeowners saw depended on the relevance of each activity, rather than where it fell in a sequence:

  • Needs attention: Actions and changes requiring the homeowner’s response.
  • Supports a decision: Quotes, conversations and details needed to evaluate options.
  • Available for reference: Completed activities, records and history.

This allowed multiple activities to happen in parallel while keeping the most relevant information in focus.

How do I decide?

Bringing contractor selection into one context

Before booking, homeowners compare contractors, discuss requirements, review profiles and evaluate quotes. Previously, these activities were spread across separate screens, requiring homeowners to move between them to assess each option.

I brought each Pro’s profile, quote and conversation into a dedicated panel, keeping the broader selection of Pros accessible within the project. Testing revealed different navigation preferences across platforms. On web, participants tended to review one Pro at a time, while some mobile participants preferred switching directly between Pros. We designed different navigation patterns for each platform rather than forcing the same interaction across both.

Messages belonged with the contractor

We first collected pre-booking messages in a separate mini-inbox on the dashboard. Having every chat in one place seemed tidy, but homeowners struggled to tell which message came from which Pro. The conversation was cut off from the profile and quote it related to. Moving each thread into that Pro’s card, with the latest message surfaced, tested better: homeowners could see who said what alongside their profile and quote. That was worth the extra height it added to every card.
What needs me now?

One project, multiple timelines

A recurring cleaning can have twenty past visits, one coming up, and a payment due. Visits and payments follow different schedules, so combining them into a single timeline would have buried whatever homeowners needed to act on.

I organized visits and payments around what’s due, what’s coming up and what’s already happened. Each card brings the most relevant information forward while keeping upcoming and past activities accessible. A cancelled visit moves to previous visits without affecting the next scheduled visit. For payments, homeowners can see what’s due, what’s been paid and what remains without having to piece together the full payment history. The structure had to hold at both extremes: a biweekly cleaning on its twentieth visit and a renovation on its third payment.


How did it go?

Feedback at any point after booking

Local referrals are at the heart of Quartermaster. Homeowners can leave feedback at any stage of a project, rather than waiting until the job is complete. To give each review some context, AI generates a short project summary based on the details the homeowner shared, which they can edit or omit.

Reviews also connect to Home Team, a feature developed by another team that lets homeowners save the Pros they trust. A positive experience is a natural moment to add a Pro to Home Team, making it easier to pick up where they left off when the next job comes around.

Project activity on mobile with the referral prompt: would you refer this Pro to your friends and neighbours?
Your review: AI-generated project summary, referral question, star rating and written review
Add as your Go-to Pro: your Team is the place to add your Go-to Pros
Why start over?

The next job starts with context

The project brief stays with the homeowner beyond the initial request. It provides the information Pros need, carries into the project workspace, and can be reused for future work.

When a homeowner returns, the AI assistant uses the previous brief as a starting point, asking what has changed rather than collecting the same information again. This applies to rebooking a Pro, extending an existing project, or requesting a new service.

09:41
Rebooking
WHAT CHANGED

What we parked

Drafts turned into a question about AI

When we tested a draft state for projects that needed more planning, participants expected AI to be built in: help choosing materials, sequencing the work and deciding what to ask contractors. That changed the problem from storing unfinished requests to helping homeowners define the project. Doing that well was beyond the release scope, so we deferred drafts until we could build them with AI assistance.

Four levels of depth for every block

Mobile ordering settled which block came first, but not how much of each to show. I defined every block at four levels, from a one-line minimized state to a full maximized view, so the dashboard could expand what mattered at each stage and collapse the rest. We deprioritized this logic because of its engineering cost and chose to test the fundamental components first.

Fusun Uzun hello@fusunuzun.comcopied to clipboard

AI-Assisted Intake & Matching

Finding the right Pro starts with understanding the project

Company
Quartermaster
TIMELINE
AI experience shipped in 2024
Matching experience shipped in 2025
Role
Lead Product Designer
09:41
AI intake: project summary with next step
09:41
Pro selection: matched Pros to request quotes from
09:41
Pro selection: quote submitted and awaiting responses

Overview

Quartermaster’s marketplace initially relied on the Operations team to interpret requests, identify the right Pros and guide homeowners through booking. That hands-on approach helped establish the service, but to grow, those decisions had to move into the product.

I designed the homeowner-facing experience for that transition. AI-assisted intake turned often-vague requests into structured projects, with the Operations team stepping in when needed. Automated matching then recommended suitable Pros based on neighbourhood referrals and eligibility criteria.

Requiring an AI conversation for every submission raised concerns about drop-off, but submission rates held. About 40% of requests arrived outside staffed hours and could now reach Pros without waiting for the Operations team.

Automated matching expanded the number of requests the marketplace could handle by roughly a third without growing the Pro network. Each project reached about 40% more eligible Pros. That increased reach, but not competitive bidding, which remained an important constraint on booking.

In 2025, projects with competitive bids booked at 62%, against 26% without.

What’s actually wrong?

Building a better project brief

Homeowners rarely arrive with a diagnosis. Some know they need an electrician; others just hear a noise or see a leak. The old form asked for a category and an optional description, and a thin description made the job easy for Pros to skip. The intake moves quickly when a request is clear and asks focused questions when it isn’t. Neither the homeowner nor the AI has to diagnose the problem, only describe it well enough for a Pro to act on. Each conversation ends in a structured brief covering scope, category, timing, location and constraints.

The Operations team defined the essential questions for each service category, and the AI decided which to ask based on each homeowner’s answers. There was concern that asking too much up front would cause drop-off. In testing, though, homeowners welcomed questions that helped them describe the job, including things they hadn’t considered. What slowed the exchange was repetition and prompts that took too long to get to the point. I initially thought three questions at once would feel overwhelming; in these relatively low-stakes conversations, people handled that format better than I expected. We kept the questions focused and limited follow-ups when an answer was skipped.

01Gather information progressively
02Structure the project requirements
03Generate a project summary
AI intake·
Desktop chat: homeowner connects with a Quartermaster support team member, who picks up from the AI conversation
Support handoff

We also looked beyond the first request. In the redesigned experience, project history carried forward: when homeowners rebooked, the AI could start with their previous project summary rather than ask them to describe the same work again. A simple stepper showing progress through the conversation also tested well and was planned for a later phase.

Who should I trust?

Matching built on neighbour referrals

The matching logic combined eligibility rules with a ranking system. Eligibility determined which Pros could take on a project, while ranking considered proximity, referrals and past performance. Pros recommended by homeowners in the same neighbourhood got priority, so local recommendations stayed central to how homeowners found Pros.

Research showed that most homeowners wanted to manage around three competitive quotes, so we surfaced five eligible Pros at the start. The team discussed a “Select all” option to speed things up. But every extra Pro meant more conversations, follow-ups and decisions for the homeowner to manage. I wanted selection to stay deliberate rather than encourage contacting everyone at once. Homeowners still needed room to widen their options when supply was thin or a Pro declined, so they could add Pros at any point without starting over.

As matching moved from the Operations team into the product, homeowners also needed to understand why particular Pros were recommended. I added a short explanation of how matching worked, alongside information about Pro vetting and the Good Neighbour Guarantee. Making the selection process more transparent gave homeowners the context to assess the recommendations rather than simply accepting them.

Request quotes·
Why these Pros
No Pros available
Fusun Uzun hello@fusunuzun.comcopied to clipboard

Chat Redesign

Keeping homeowners and Pros connected beyond a single project

Company
Quartermaster
TIMELINE
Shipped in Q1 2024
ROLE
Lead Product Designer
Redesigned homeowner chat

Overview

At Quartermaster, chat was organized around individual projects. Each new project opened another conversation between the same homeowner and Pro, plus a new support channel, and important messages got buried. Pros, who were receiving so many messages from Quartermaster, stopped engaging in chat and called homeowners instead, which turned those conversations into a black box for us. I led the redesign of chat around ongoing relationships and brought pre-booking communication into the platform.

Why did Pros stop responding?

More nudges made more noise

When Pros stopped responding, the product sent more reminders, adding to the noise that had pushed them away. Support messages and automated updates already competed with direct conversation. And because so much of the discussion before booking happened by phone, the Operations team couldn’t see it. A Pro might already have arranged a quote visit while the system still treated the request as unanswered, and kept sending reminders.

Legacy desktop chat with a separate thread per project
Who am I talking to?

One conversation per relationship

We replaced per-project conversations with one continuous thread between a homeowner and a Pro. Messages from current and past projects stayed together, so both had their history when they worked together again. The trade-off was navigation: a single thread cut duplication but could get hard to use as the relationship built up projects. I added a project menu inside the conversation so either person could move between projects without going back to separate inbox threads.

Redesigned Pro chat
Pro chat on mobile with the Active projects sheet: plumbing installation and plumbing repair
Fusun Uzun hello@fusunuzun.comcopied to clipboard

CTV News Redesign

One system, fifteen local news sites

COMPANY
Bell Media
TIMELINE
Shipped in 2020
Role
Lead UI Designer - News
CTV News redesign components

Overview

CTV News is one of Canada’s most-watched news organizations, both locally and nationally, and has a network of fifteen local news operations. Each site shared a content system but had its own custom-built front end, so every improvement had to be designed, built and launched fifteen times. Bell Media set out to move them all onto a single codebase and shared design system. I designed the modular homepage system and article template it was built on.

The first phase went live in 2020 and drew 5.8M unique visitors, up 29% from the year before. The system also proved itself in the local rollout. Producers could build each new local site themselves, without custom design or development. We started with the larger newsrooms (Montreal, Vancouver and Calgary) and then launched a new local site every two or three days.

ONE SYSTEM, FIFTEEN NEWSROOMS

Evolution of a single teaser

A news homepage has no fixed density: a quiet morning and an election night need different structures. I designed a modular system that grows from a single teaser into story clusters, and made the clusters themselves responsive, so producers could resize them manually and instantly instead of being bound to the grid.

Scanning vs reading

Scanning vs reading

This modular approach offers a clear hierarchy that helps readers scan the content and guides them to what is happening, what the most important story of the day is and what matters to them throughout the page. Horizontal scroll components with improved signalling break up the story clusters and help readers quickly choose what to engage with.

CTV News homepage story clusters
Instant vs slow discovery

Instant vs slow discovery

Each module is designed for a different pace of reading. Happening Now offers a quick way to catch up: a chronological list of headlines with red dots and timestamps.

Special series serve readers who want to explore a subject in more depth. Each gets a larger, image-led package with a lead story, supporting pieces, explainers and a link to the full series, giving readers a sense of what’s available before they click.

CTV News series package on tablet
09:41
CTV News series package on phone
09:41
Happening Now and Editor's Picks on phone
Fusun Uzun hello@fusunuzun.comcopied to clipboard

Shadows of the North

Shadows of the North, an interactive experience commissioned by Amnesty International Canada, is a reflection of experiences with immigration detention in Canada through the mechanics of generative art.

Year:
2021–2022
Agency:
Supernova Art Collective
Project Type:
Generative and interactive experience
Role:
Co-Creator - Designer

About the project

Shadows of the North is a reflection of experiences with immigration detention in Canada. Commissioned by Amnesty International Canada (English Section), Supernova Art Collective curated and directed the artistic content in consultation with researchers from Amnesty International and Human Rights Watch, who acted in an advisory capacity drawing on the research and interviews they conducted for their joint report, “I Didn’t Feel Like a Human in There: Immigration Detention in Canada and Its Impact on Mental Health.”

Most of the testimonies, quotations and informational text in Shadows of the North are taken directly from the report published on June 17, 2021. In those instances, voice actors substituted the voices of immigration detainees, but we aimed to honour their regions of origin. We were privileged to include the words of “Amina” and Abdelrahman in their own voices.

Credits

By Supernova Art Collective

Producer: Fusun Uzun

Creators: James Essex & Fusun Uzun

Web & GL Development: James Essex

Visual Design: James Essex & Fusun Uzun

Sound Design & Mixing: Sound Atlas & Duygu Asena Unlu

Voice Consultant: Lopa Sircar

Voice-Over Artists: Joseph: Omari Newton · Michelle: Jane Dut · Omar: Noaman Rahin · Jonathan: John Echano · Charles: Kym Dominique-Ferguson · John: Bob Youssef · Jason: Vincent Smith · Ken: Macdonald Nkemda · Usman: Sam Youssef · Karim: Bob Youssef · Tiffany: Ya Ju Weng · Amina: Own voice · Abdelrahman Elmady: Own voice

Produced with the co-operation of the Union of B.C. Performers (UBCP/ACTRA)

Fusun Uzun hello@fusunuzun.comcopied to clipboard

I started as a copywriter at agencies including TBWA and Young & Rubicam in Istanbul, before moving to Canada to study design. Along the way, I co-founded BFK, a small creative practice working across documentaries, websites and visual identities. After completing my MFA in Digital Futures at OCAD University, where I received a Governor General’s Academic Medal, I shifted my focus to product design. Over the past decade, I’ve designed consumer-facing products across public media, venture studios and startups.

Copywriting, documentary and product design are different disciplines, but making something feels much the same in each. Every project needs structure, but not always the same one. The first attempts rarely resemble what the work becomes, thankfully. I love researching the problem and the people in it, but it’s in the making that I find the gaps and traps in my thinking. The satisfaction of a launch is brief. Once the work is out in the world, it has to stand on its own.

Earlier work in branding and graphic design, alongside experiments, collaborations and documentaries.

{{ t.title }}
Fusun Uzun hello@fusunuzun.comcopied to clipboard