Windows
Audio
Shared audio
I reframed shared audio from a minimum-scope question into a findability problem. Research moved the plan from one entry point to six, and the Quick Settings tile we recommended is how people start sharing in Windows 11.
My part
Designer: framing, flows, entry points; co-ran the study
Duration
Nov 2024 – Jun 2025
Crew
Bluetooth PM and engineering, UX research, design
Ships on
Windows 11
Context & role
Shared audio lets one Windows 11 PC play the same audio to two Bluetooth LE Audio headphones, earbuds, speakers or hearing aids at once. It serves people watching or playing together on one device, and people who use hearing aids. The Core OS Bluetooth team owned scope, platform capabilities and the plan of record, and its engineers built the feature. I was the design partner to the Bluetooth PM from November 2024 to June 2025: I owned the problem framing, the stakeholder flow-mapping session and the proposed flows and entry points, and co-ran the usability study with a UX researcher.
Stakes
Bluetooth audio on Windows was one listener at a time, so people used wired splitters or put up with audio that drifted out of sync. Phones already shared audio and supported hearing aids. The platform team was stretched by partner requests, and the business risk was spending the investment on a competitive feature people would never find.
Reframe
Asked: how little can we build? Reframed: where will people look?
The early concept answered the scope question cheaply with one entry point in Sound settings, which matched an engineering view of sharing as just another sound output. Prior Windows research said people use Settings to finish a goal, not to explore. I redefined the problem as "where will people look for this, and what will they call it," because a feature placed where no one looks fails however little it costs. I anchored scope on one scenario, watching a film on a plane with someone, and ruled broadcasting to a room of listeners out of scope.
Illustration: one entry buried in Sound settings, versus device-first doors that all open the same panel.
Evidence
The feature was nearly invisible
First I ran a session where every stakeholder sketched how sharing should work. Three competing flows came out, which showed the team itself disagreed about where the feature belonged, and gave the study a precise question. A researcher and I then ran a 10-person usability study on a semi-functional prototype. Most participants couldn't find the sharing settings on their own and several gave up, yet everyone who got past the entry point could share, adjust volume and stop. They split between Sound and Bluetooth settings, and they thought in devices, not outputs or "sessions".
01
The team disagreed with itself. Stakeholder sketches produced three competing flows, which gave the study a precise question to answer.
02
Once found, it worked. Everyone who got past the entry point could share, adjust volume and stop sharing. The problem was finding it.
03
People think in devices, not outputs. Nobody looked for a sound output or a "session"; they looked for their headphones.
Key decision
Put the entry where people look, including the "untouchable" panel
The options were to keep the single Sound-settings entry in the plan of record, build new surfaces wherever people looked, or add device-first entry points everywhere people looked, all opening the same panel. We recommended the third, including a Shared audio tile at the top of Quick Settings. Stakeholders had ruled the Bluetooth Quick Settings panel untouchable; participants got stuck there and failed, so we used that evidence to add a "Share audio" action to it. I chose several device-first entry points that reuse one panel over a single settings entry, because people look for devices first, accepting more surfaces to negotiate and a compromise on the Bluetooth settings card.
Second call
Treat "not supported" as a designed state, not a dead end
Auracast-capable accessories would take time to become common, so many people would open the feature with headphones that couldn't join. The cheap option was to hide incompatible devices. I chose to show them in the list, marked "Shared audio not supported", because hiding them would make people think the feature was broken, accepting a longer list and more states to build. An empty list says "No devices are available" with a "Can't find a device?" help link, and hovering a greyed device explains why it can't join.
Illustration: three paths, and why device-first entry points won.
Craft
Every step of sharing has a designed state
The panel asks people to "Select two output devices to connect" and enables Share only once two are chosen. Each device shows "Connecting..." while it joins, and while sharing, both rows read "Sharing audio" with "Stop sharing" beside them. Two questions stayed open at handoff, and I named them rather than guessing: people's expectations of the system volume slider conflicted, sometimes within one person, and they split on whether sharing should end manually or automatically. A later preview update added per-headphone volume, which answered part of the first.
Illustration: each step of sharing, including headphones that can't join.
System
Devices first, one panel behind every door
People manage sharing the way they manage devices, so the feature appears wherever they already see their headphones. Four rules carried it across surfaces: put the entry inside device lists; route every entry point to the same panel, so there is one thing to learn and one thing to build; use the plain name "Shared audio", not a technology brand; and show devices that can't share, with the reason, instead of hiding them. Routing six entry points to one panel is what made the research affordable for a cost-pressed team.
Illustration: four rules carried the model across every surface.
What shipped
What shipped
Shared audio entered public preview in Windows 11 in October 2025, first on select Copilot+ PCs, then rolled out beyond Insiders in the May 2026 update. The path people take is the one the research argued for: open the Shared audio tile in Quick Settings, pick two LE Audio headphones, choose Share, and end with Stop sharing. It carries the plain name from the research rather than a technology brand, and selection is device-first. A later preview update added a volume slider for each headphone and a taskbar icon while sharing, which answered part of the open volume question. Public notes don't yet show the Sound and Bluetooth settings entry points, so the full six-door model is still in progress.
Outcome
What changed
Shared audio entered public preview in October 2025 and rolled out beyond Insiders in the May 2026 Windows 11 update. People start it from the Quick Settings tile we recommended, pick two headphones and choose Share, with the plain name and device-first selection from the research. A later update added per-headphone volume, which answered part of an open question from the study. No usage data is available yet.
Reflection
Test where people look before the plan hardens
Because the finding arrived after the plan of record existed, every change had to be negotiated against engineering milestones. A first-click test on the early mockups would have shown the device-first model months sooner, and next time I'd put "untouchable" surfaces in the study from the start. The principle I took away: when a platform team frames a problem as cost, the designer's leverage is to reframe it around how people think and make that evidence cheap to act on.
