The architectural document the room is built to.
Nine specifications, each measurable. Twelve months of continuous field measurement across six installs. 2,840 employees in the consented sample. Published so the category cannot be diluted by imitation, and so a CHRO can defend the line item without a sales conversation. Below the threshold, the room is something else.
Foreword.
This document exists because the category drifted. For fifteen years Indian companies bought wellbeing in three forms, an app, a webinar series, an EAP contract. The budget went up. The numbers a CHRO could defend in front of a CFO did not.
For our first six years inside Indian workplaces, the question was "what programme can you bring to our floor?" We ran festivals, retreats, eight-city tours, and learned what a tired body actually needs from a room before it needs anything from a vendor. By year five the question changed: CHROs stopped asking what we would bring, and started asking what it would take to put the work inside the building.
We built the first Zen Garden inside Flipkart Bengaluru in 2024 and measured it for twelve months. The room paid for itself in retention math inside year one. This document is what we wrote down so the second, third and sixth installs would not be guesses. A room that meets the nine specifications is a Zen Garden. A room that does not is something else.
Recovery is an architecture problem.
For thirty years, workplace recovery has been treated as a behaviour problem. Apps. Webinars. Awareness weeks. The premise: the worker is failing to choose recovery, and we will help her choose better.
The argument of this Standard is that recovery is an architecture problem, not a behaviour problem. A floor either gives a body somewhere to recover, or it does not. There is no third state. This is why eight of the nine specifications are physical, not procedural, square feet, acoustic isolation, lighting temperature, material choice. The Standard does not ask the worker to choose recovery. It installs recovery into the floor.
Recovery cannot be added as behaviour. It can only be installed as infrastructure.
Nine specifications. Below the threshold, the room is something else.
Each specification has a rule, the measurable, non-negotiable threshold; a body, the rationale from twelve months of field data; a benchmark, what the founding install actually delivers; and an audit method a qualified architect uses to verify it at handover.
Minimum 350 square feet, dedicated.
Below 350 sq ft the three Zones cannot coexist without acoustic and visual interference. Dedicated means no overlap as a meeting room, phone booth, breakout, library or kitchenette. The footprint is calibrated to 200–600 people on the floor; above 600 commissions a second room.
STC 45 minimum on every shared wall.
STC 35 is normal office. STC 45 is the threshold at which the body stops listening for interruptions and begins to release; the cortisol curve does not move below it. Resilient-channel construction, mineral-wool batt, automatic door bottoms, ceiling clouds tuned to the 500–2000 Hz band.
Circadian-tuned, 2700K to 3500K.
Overhead office fluorescent disqualifies a space. A room that looks the same at 11am and 5pm has failed. Each install ships with a DALI fixture set on a three-state curve: morning grounding, midday alignment, evening release. Only the resident expert can override the schedule.
Wood, fabric, brass. No glass walls.
Built from materials that absorb: FSC-certified Indian hardwoods, linen and wool, natural stone, brass, terracotta, lime plaster. Glass walls are disqualifying, a recovery space the floor can see into is a fishbowl, and utilisation collapses inside two months. A brass plate at the threshold carries the install number, date and version.
All three Zones present, none collapsed.
Three Zones, each a distinct stage of the recovery arc: Balance (grounding, the reset), Blossom (awareness, the opening), Bliss (alignment, the release). Separated not by walls but by floor material, ceiling height, lighting and modality. A single open room is a wellness room, not a Zen Garden.
One named, certified expert. Five days a week.
Every Zen Garden has a dedicated, named, on-site resident expert. Not a rotating contractor, not a once-a-week visitor, not a chatbot. In the room during business hours, available without booking, named on the brass plate and in the monthly report. By month four, the most-loved member of the floor across every install.
Two hundred modalities, minimum.
At least 200 distinct modalities: breathwork, sound, movement, art-based methods, somatic release, listening, journaling, micro-meditation, gratitude. Below 200 the room reads as a yoga studio. The breadth produces the 60–70% utilisation we measure, people return because Monday's need differs from Thursday's. 14,200+ sessions logged.
Two anchors per day. Otherwise open.
A morning anchor and an afternoon anchor, held in the Balance Zone by the resident expert, free twenty-minute drop-in practices. Outside the anchors the room is open and the expert is on demand. Programming is the rhythm; open access is the volume. Utilisation peaks on Tuesday at every install in the cohort.
A gratitude book. A visit log. A monthly report.
Three artefacts kept by the resident expert: a leather gratitude book in the Bliss Zone where employees write anonymously; an anonymised digital visit log at the entrance; and a monthly summary to People & Culture covering utilisation, dominant modalities and one expert observation. The documentation is the answer to the CFO question.
Twelve months. Six rooms. One dataset.
The first Zen Garden was installed inside Flipkart Bengaluru in January 2024. Five more permanent installs followed. Twelve months of continuous instrumentation produced the dataset every specification rests on. If a spec disagrees with the data, the spec is wrong, and we rewrite it.
Instrumented four ways: anonymised threshold visit logs; resident-expert modality logs; a six-question Maslach short-form survey across three waves (n = 2,840); and an opt-in bio-marker panel measuring cortisol, HRV and afternoon focus. A senior cohort (n = 480) is tracked by access-card grade. The Tuesday Index, utilisation peaking on Tuesday, replicates across all six sites.
The senior cohort, calendars they cannot escape and decisions they cannot offload, turn out to be the people most starved of legitimate, unobserved recovery time.Finding 04 · Senior cohort behaviour
How we measured.
Instrumentation. Each room is instrumented at handover. Visit events are logged via an anonymised entry counter; daily users are de-duplicated by a passive proximity reading. No badge data, no name, no facial recognition. The room counts bodies, not identities.
The consented panel. A subset of employees opts in to a bio-marker panel, with consent renewed each quarter. The Flipkart panel covered 2,840 employees, held against itself across weeks 1, 13, 26 and 52. The 25% is the within-panel delta between Q3 baseline and Q4 measurement, the composite of cortisol, HRV and self-reported focus, all moving in the same direction.
What we did not measure. Not attrition, too many confounders inside a single year. Not "engagement," survey scores are not bio-data. We measured the only thing that holds up under a CFO's question: the body's signal at the close of the working day.
A strict reading.
This is one site, one year. The 25% reduction is the Flipkart install, measured against Flipkart's panel. It is not a population statistic for Indian offices, it is the strongest finding we have. The next five installs show the same direction, but are not yet a year old. The Standard says what it can say, and no more.
STC 45 is the floor, not the ceiling. We have measured up to STC 50 in a Mumbai install and seen incremental gains. The resident-expert specification is the most expensive, and the data is clear that the rotating-contractor variant did not produce the same engagement curve. The economics are real, and the spec is non-negotiable for a reason.
Audit checklist. Fifteen items.
Fifteen items a qualified architect verifies before handover. If any one fails, the install is not certified to Volume I.
The Council that revises the Standard.
The Zen Council is a private membership of CHROs, founders and modality experts whose work has informed this volume. It exists because the Standard has to evolve, and cannot evolve at Zariyaa's desk alone. Once a year, in November, the Council meets in Bengaluru to debate the next volume. Membership is by invitation only and is not for sale; every certified workplace becomes eligible at its twelve-month anniversary.
Volume II · in field testing.
Status: active field testing. Publication target May 2027. Volume II expands the specifications with a chapter on Stay-Lock, the structural design choices that have, in the diagnostic data so far, correlated with a measurable reduction in senior-bench turnover. Eighteen new specifications, drawn from the five installs after Flipkart, with panels totalling 8,400+ employees. Also publishing: the November 2026 Council revision notes, and three new case studies. An institution publishes on cadence.
Citation.
Bohara, K. & Bohara, R. The Recovery Standard, Volume I. Bengaluru: Zariyaa, December 2024. Nine architectural specifications for workplace recovery, with twelve months of measured field data from the founding cohort. n = 2,840.
The install starts with three days.
21-day install · 24h reply · Cohort 2026 · Bengaluru.