Read almost any workplace design paper published this year and you will find a version of the same thesis. The technology office is no longer an office. It is part laboratory, part production floor, part demonstration centre. Engineering neighbourhoods sit beside open labs. Collaboration hubs sit beside focus work. The separation between functions, the argument goes, is exactly what slows innovation down, so the best workplaces remove it.
The argument is a good one. Proximity does accelerate problem solving.
What these papers rarely mention is sound. You can read whole reports on the anatomy of the modern tech workplace and never encounter the words acoustics, privacy, or distraction. The design conditions get described in detail. The consequence goes unnamed.
That gap matters more this year than last, as companies pull people back to the office in the name of collaboration, and building the space around it.
The Density Is Returning Faster Than the Investment
According to CBRE's 2026 Office Occupier Sentiment Survey, 89% of office-using organisations now require at least three in-office days per week, up from 78% a year earlier. Nearly two-thirds of technology companies say they plan to grow their footprint, a larger share than any other industry in the survey, and tech firms led US office leasing in the first half of the year with a 21% share.
The same survey found that only 14% of companies are making major improvements to their space.
More people, in more hours, in floorplates designed around openness, with no corresponding investment in how those floorplates handle sound. CBRE surveyed 97 companies, so treat this as sentiment rather than a market census, but the direction is consistent with what the leasing data shows.
The Research Is Unusually Clear About What Fails First
This is where the evidence gets specific, and where most workplace conversations get it wrong.
The intuitive assumption is that noisy offices are too loud. The measurement evidence says otherwise.
A 2025 study published in the Journal of the Acoustical Society of America, Yadav, Kim, Hongisto, Cabrera and de Dear, measured room acoustics and occupied sound environments across 28 offices alongside surveys of 349 occupants. Modelling acoustic dissatisfaction, the researchers found that lack of privacy contributed roughly 25% more than noise disturbance in predicting how dissatisfied people were.
Earlier longitudinal work points the same direction. A relocation study published in Ergonomics by Hongisto and colleagues followed workers moving from private offices into open plan. Speech privacy fell significantly. The noise level averaged across the workday did not change.
The problem is not the volume of the room. It is that you can understand every word of a conversation you are not part of, from four desks away. Human speech is the one sound the brain refuses to treat as background, and intelligible speech is the specific mechanism that degrades concentration.
Occupant data has been saying this for a decade. Leesman, which has surveyed workplace experience at scale for years, reports that noise levels matter to around 70% of employees while only about 35% are satisfied with them. Even inside their certified high-performing workplaces, satisfaction with noise sits near 45%. Their longitudinal view is more damning than the snapshot: satisfaction with noise has moved only marginally since 2013, while nearly everything else about workplace design has improved.
Leesman has also noted that dissatisfaction with noise levels carries the strongest correlation with employees saying their workplace does not support their productivity.
What an Interruption Costs an Engineering Organisation
For technology companies, the productivity argument is not abstract, and it is not new.
In research presented at CHI, Parnin and DeLine analysed thousands of recorded programming sessions across dozens of developers, supported by a survey of several hundred more. Their finding, widely cited since: after an interruption, a developer typically needs on the order of ten to fifteen minutes before resuming code editing, and only rarely resumes in under a minute when interrupted mid-edit.
Apply that to a floorplate where the design intent is deliberate adjacency between people who need to talk and people who need to think. The interruptions are not incidental. They are structural.
There is a compounding factor worth naming. Gensler's Global Workplace Survey 2026, a panel survey of 16,459 full-time office workers across 16 countries, found that employees most embedded in AI workflows spend less of their week working alone and more of it learning and socialising. As AI absorbs solitary execution work, what expands is conversation. More talking, in the same room, beside the people still doing deep work.
There Is Already a Standard Written for This Exact Room
Here is the part that surprises most workplace teams. The condition those write-ups and surveys describe without naming has had an international design standard since 2021.
ISO 22955, Acoustics, Acoustic quality of open office spaces, is the workplace design guidance layer built for exactly this problem. It takes the open-plan measurement framework and sorts it by what people are actually doing in the space. It starts from the core tension: open plan is asked to support two acoustically contradictory activities, oral communication and focused individual work. Rather than treating the floor as one uniform condition, it defines six space types by activity. The sixth is spaces that combine several activities at once.
That sixth type is the technology workplace as currently designed. A floorplate where engineering, collaboration, testing, demonstration and focus work share the same volume is not an edge case in ISO 22955. It is an anticipated condition with defined acoustic indicators.
Very few technology workplace briefs reference it. That is a specification gap, not a knowledge gap.
Adaptability Is the Stated Goal. Acoustics Is Where It Usually Breaks.
Every serious workplace paper published this year lands on the same success criterion: the space must absorb new technologies, new team structures, new equipment and new ways of working without a major redesign every few years.
Applied to acoustics, that criterion is unforgiving.
Built acoustic solutions are permanent decisions. Partitions, screen heights, ceiling absorption and room geometry are fixed at construction. They are correct for the layout they were designed around, and progressively less correct with every reorganisation, every team that doubles, every lab that turns into a demo suite. Re-solving them means construction.
Adaptive sound masking behaves differently. It raises the background sound level in a controlled, engineered way so that speech from a distance stops being intelligible, without raising it to the point of being noticeable itself. Because it is a networked, tunable system rather than a built assembly, the acoustic condition of a zone can be adjusted when the use of that zone changes. A floorplate that is re-zoned is re-tuned, not rebuilt.
This is also why masking appears in performance standards rather than only in product catalogues. The WELL Building Standard's Sound concept treats sound masking as one of several complementary levers alongside absorption, barriers, and background noise limits, with defined commissioning levels and spatial uniformity requirements.
One important caveat, and it is the reason masking has a mixed reputation among people who have experienced it done badly. A masking system that is installed but not properly calibrated to the room is worse than none at all. Spatial uniformity and spectral shaping are what separate a system people never consciously notice from one they complain about within a week. This is engineering work, not a volume knob.
The Practical Shift: Specify It as Infrastructure
The most useful idea in current workplace design thinking is that planning now starts with technical workflow before it becomes spatial design. Equipment first, environmental conditions second, layout third.
Acoustics belongs in that first pass, alongside power, network and HVAC, for three reasons:
- Zoning is a design decision, not a commissioning decision. A floorplate with a lab at one ambient level, engineering benches at another, and focus space at a third needs distinct acoustic zones defined before the ceiling is closed. Retrofitting zone boundaries is expensive.
- Dedicated privacy spaces are now standard in the floor plan. Product demonstration centres, cybersecurity operations rooms, AI research areas and pre-launch prototype spaces all carry speech privacy requirements. These are the same requirements law firms and financial institutions have specified for years, arriving in a new sector under different names.
- Exposed ceilings change the acoustic math. The exposed ceilings and open plenums that define the current tech aesthetic remove the acoustic ceiling tiles that used to absorb most of the noise overhead. That is a solvable condition, but only if it is known early.
- Portfolio consistency requires a written standard. Organisations running centres across multiple countries, including the rapidly maturing global capability centres documented in the nasscom-Zinnov India GCC Landscape Report 2026, need one acoustic performance target and one verification protocol that travels, delivered locally and commissioned to the same measured result everywhere.
The gap, plainly
Technology companies are expanding their offices, densifying them, and designing them around deliberate adjacency between activities that are acoustically incompatible. The research says privacy fails before volume does. A standard already defines what good looks like. And most briefs still treat sound as something to address after occupancy, when the complaints start.
The workplace papers are right that the office has become something more than an office. They are just missing the part where every barrier you remove is a specification you now owe.
Bringing acoustic performance into the specification
Soft dB, the most advanced sound masking system available, is engineered for exactly this condition: floorplates where activities with different acoustic needs share the same volume, and where those zones will change. Calibration is performed and permanently owned by Soft dB, so the system delivers a measured, verifiable result at handover and stays correct as the space evolves.
If your team is planning a technology workplace, a global capability centre, or a portfolio standard, we can help translate the acoustic requirement into a specification before the ceiling closes.
Request an acoustic consultation
FAQ
Is sound masking the same as white noise?
No. White noise has equal energy across all frequencies and is generally perceived as harsh. Engineered sound masking uses a spectrum shaped specifically to cover the frequency range of human speech, then calibrated to the room so it remains below conscious notice while reducing speech intelligibility at distance.
Does sound masking make an office louder?
No. When done properly, sound masking actually makes an office feel calmer and quieter. It raises the background sound level in a controlled way, to a level occupants do not consciously register. The perceived effect is the opposite of louder, because distant conversations stop resolving into words and stop drawing attention. Noisy distractions get effectively muffled out under the smooth background sound layer generated by the masking system.
Does an open ceiling rule out sound masking?
No. With the right emitters, an open ceiling plays to masking's strengths rather than against them. The real design challenge is that there is no ceiling plane to mount into and no plenum to conceal the system, so emitter choice, placement and calibration all have to account for the exposed deck from the outset. Our approach turns that deck into an asset. Up-firing emitters, the SMS-STR and SMS-STX, project sound upward at the structure, which reflects and scatters it back down across the floor. Instead of aiming sound at people, you are using the deck itself as a diffuser, so the masking arrives evenly rather than pooling under each speaker, and it stays hard to localize. Even, source-less background sound is exactly what makes masking effective. The ceiling condition that is supposed to make masking harder becomes the thing that distributes it well.
Is there a standard for open plan office acoustics?
Yes, though "standard" splits into a few things. The core one is ISO 3382-3, built specifically to measure open plan offices: it defines metrics like distraction distance and the spatial decay rate of speech across the floor. Then you have ISO 22955 adding design targets on top of ISO 3382-3. ISO 22955 provides technical guidance on the acoustic quality of open office spaces, defining six space types by activity with corresponding acoustic indicators, including spaces that combine multiple activities. In North America, the equivalent framework is the ASTM E-series. ASTM E1374 is the umbrella guide for designing and evaluating both open-plan and closed offices, and it pulls in two companion standards to handle the specifics. ASTM E1130 defines how to objectively measure speech privacy using the Articulation Index, quantifying how intelligible a voice is at a listener's location relative to background noise. ASTM E1573 covers sound masking, setting out how to measure and report performance in A-weighted and one-third octave band levels, which is how you verify an installed system delivers the right spatial uniformity, temporal stability, and spectral shape across a space.
Building certification systems like WELL, LEED, and BREEAM don't measure acoustics themselves but set targets that reference these standards, with WELL being the most demanding on speech privacy and the only one that explicitly calls for sound masking technology.
