BASTION SPACE INTELLIGENCE MODULE
Part of Forge Bastion IWMS · Built in Rust

You are paying for buildings nobody is using

Bastion Horizon fuses IoT occupancy sensors, digital twin simulation, and predictive AI to reveal exactly how every square foot of your portfolio is used — then optimizes it. Real-time. Continuously. Without surveys, walkthroughs, or guesswork.

31%
Average office utilization in the Americas — less than half the 64% pre-pandemic benchmark. CBRE's 2024–2025 Global Workplace & Occupancy Insights found that enterprises are paying for nearly 70% more space than they actually need. The global savings potential from workspace optimization: $1.5 trillion.
8
Intelligence engines
100K+
Sensor points per campus
2s
Update frequency
97.8%
Detection accuracy
The Utilization Crisis

Every enterprise thinks it knows how its buildings are used. Almost none of them actually do. Room booking systems show reservations — not reality. Badge swipes show entry — not occupancy. Surveys capture opinions — not behavior. The result is a massive, invisible waste: millions of square feet heated, cooled, cleaned, and maintained for people who are not there.

Research shows that nearly one-third of all desk time is passive occupancy — a laptop or bag on a seat with no person present. When you strip passive occupancy from the data, true desk utilization is typically 20–35% lower than what standard reporting shows. This is the gap between what organizations believe about their space and what is actually happening inside it. Bastion Horizon closes that gap with continuous, sensor-driven intelligence that distinguishes active presence from abandoned belongings, occupied rooms from booked-but-empty ones, and genuine demand from institutional habit.

The Waste Cascade

How underutilization compounds

Stage 1 — Phantom Occupancy
Booking systems show 78% room utilization. Actual occupancy is 34%. Nearly half of all reserved conference rooms are either no-shows or end early. The organization believes it has a space shortage. In reality, it has a visibility shortage.
↳ Bastion detects: Auto-release after 12 min of zero occupancy. Reclaims 44% of phantom bookings.
Stage 2 — Passive Inflation
Badge data and desk sensors show 62% desk utilization. Active occupancy — a person genuinely at the desk — is 38%. The gap is jackets on chairs, laptops left overnight, personal items claiming territory. Every inflated number leads to inflated square footage projections.
↳ Bastion detects: AI optical sensors distinguish human presence from objects. True utilization exposed.
Stage 3 — Demand Distortion
Department heads request additional floors based on headcount growth. Existing space can absorb 40-60% more people with scheduling optimization. Hybrid patterns show Tuesday-Thursday peaks at 2-3× Monday and Friday — but space is provisioned for theoretical maximum, not actual demand curves.
↳ Bastion detects: Predictive demand modeling reveals when space is needed vs. when it sits empty.
Stage 4 — Capital Misallocation
The organization approves a new building or major lease expansion. The problem was never capacity — it was scheduling. Tens of millions in construction or lease costs committed to solve a problem that data could have resolved in weeks.
↳ Bastion detects: Scenario modeling shows space consolidation alternatives before capital is committed.
Stage 5 — Operational Hemorrhage
Empty floors are heated, cooled, cleaned, secured, and maintained at full capacity. HVAC running for zero occupants consumes up to 25% of total building energy. Cleaning crews service spaces nobody has entered. The operating budget bleeds silently because nobody can see the absence.
↳ Bastion detects: Occupancy-linked HVAC/lighting automation. Zero presence = zero waste.
Intelligence Engines

Eight engines. One spatial nervous system.

01
Real-Time Occupancy Detection
Multi-sensor fusion · Active vs. passive classification · Privacy-first architecture
The foundation of every insight Horizon produces. This engine ingests data from ceiling-mounted optical sensors, PIR motion detectors, desk-level presence sensors, and Wi-Fi/BLE signals — then fuses them into a unified occupancy model updated every two seconds. The AI distinguishes between active occupancy (a person at a workstation) and passive occupancy (a bag or laptop with no person present), eliminating the inflation that plagues conventional systems. All processing happens at the edge — images are converted to anonymized XY coordinates and immediately discarded. No images are stored, transmitted, or retrievable. The system is GDPR-compliant, SOC2 Type II certified, and ISO/IEC 27001 audited.
Multi-sensor fusion — optical, PIR, ultrasonic, BLE, and Wi-Fi signals triangulated for sub-meter accuracy across open plan, private offices, and conference rooms
Active vs. passive classification — neural network trained on 4.2M occupancy images distinguishes human presence from objects with 97.8% accuracy, exposing true utilization
Edge-processed anonymization — all visual data converted to coordinates at the sensor. Zero PII captured, stored, or transmitted. Privacy by architecture, not policy
Headcount precision — conference room sensors count individuals, not devices. Detects when 3 people occupy a 12-person room, triggering right-sizing recommendations
97.8%
Active vs. passive accuracy
2s
Update frequency
1,000ft²
Coverage per sensor
0
Images stored
02
Hybrid Demand Forecasting
Day-of-week prediction · Seasonal modeling · Headcount-adjusted capacity planning
Hybrid work destroyed the predictability of office attendance — but it created a new kind of predictability. This engine analyzes 12+ months of continuous occupancy data to reveal the demand signatures that characterize each building, floor, and zone. Tuesday-Wednesday-Thursday peaks typically run 2-3× Monday and Friday in hybrid portfolios. Horizon doesn't just report these patterns — it forecasts them, projecting occupancy by hour, by day, by zone, 14 days in advance with 89% accuracy. This allows facilities teams to pre-position resources, adjust HVAC schedules, and dynamically assign cleaning crews to where people actually are.
14-day occupancy forecasting — LSTM neural network models trained on hourly occupancy data predict demand curves by floor, zone, and space type with 89% accuracy
Seasonal pattern detection — identifies quarterly cycles, holiday effects, weather correlations, and organizational events that shift occupancy baselines
Headcount-adjusted modeling — correlates HR data (new hires, departures, team relocations) with space demand to project capacity needs 6-12 months forward
Dynamic resource scheduling — cleaning, security, and catering crews dispatched based on predicted occupancy, not fixed schedules. Eliminates servicing of empty floors
89%
14-day forecast accuracy
93%
Same-day accuracy
2-3×
Midweek vs. Friday ratio
34%
Cleaning cost reduction
03
Digital Twin Spatial Modeling
Live 3D building replica · Scenario simulation · What-if analysis before commitment
Every building managed by Horizon exists as a continuously updating digital twin — a live 3D model synchronized with real-time sensor feeds, HVAC telemetry, lighting states, and booking systems. This is not a static BIM model. It is a living replica that reflects the building's actual state at any moment: which floors are occupied, which rooms are in use, which zones are consuming energy for zero occupants. The twin enables scenario simulation — test closing a floor, consolidating departments, switching to hot-desking, or adding a new team before making any physical change. See the impact on occupancy distribution, energy consumption, travel paths, and amenity demand instantly.
Real-time synchronization — digital twin updates every 2 seconds from IoT sensor network. Occupancy, HVAC, lighting, air quality, and booking data rendered as a live spatial heatmap
What-if scenario engine — simulate floor closures, department relocations, desk-sharing ratios, and layout changes. See projected impact on utilization, energy, and employee flow before commitment
BIM/CAD integration — imports existing architectural models (Revit, AutoCAD, IFC) and overlays live operational data. No need to rebuild building geometry from scratch
Multi-campus federation — twin models for 50+ buildings across a global portfolio, with unified dashboards showing comparative utilization, energy, and capacity metrics
2s
Twin sync frequency
50+
Buildings per federation
<3min
Scenario simulation time
98.3%
Prediction accuracy (validated)
04
Conference Room Intelligence
No-show detection · Auto-release · Right-sizing recommendations · Ghost meeting elimination
Conference rooms are the most expensive square footage in any office — and the most consistently wasted. Industry data shows that 40-50% of booked conference rooms are either no-shows or end more than 15 minutes early. Meanwhile, employees spend an average of 12 minutes per meeting searching for available rooms. Horizon's Conference Room Intelligence engine monitors actual occupancy in real-time, detects no-shows within 12 minutes of a booking start, automatically releases the room back into inventory, and over time learns which rooms are chronically overbooked, underused, or mis-sized for the meetings they host.
No-show auto-release — if zero occupancy detected 12 minutes after booking start, room is released and made available. Notification sent to booker. Reclaims 40-50% of phantom bookings
Right-sizing analysis — correlates room capacity with actual headcount. Surfaces patterns: 65% of meetings in 12-person rooms have 3 or fewer attendees. Recommends room reconfigurations
Ghost meeting detection — identifies recurring bookings (daily/weekly) with consistent low or zero attendance. Flags for cancellation, freeing chronic room hoarders
Dynamic booking guidance — when an employee books a 12-person room for a 3-person meeting, system suggests a huddle space instead. Reduces mis-matched bookings 58%
44%
Phantom bookings reclaimed
12min
Auto-release threshold
58%
Mis-matched booking reduction
3.2×
Effective room availability increase
05
Energy-Occupancy Optimization
HVAC zone control · Occupancy-linked lighting · Carbon-per-occupant tracking
Buildings consume 30% of global final energy — and the largest waste occurs when HVAC and lighting systems run at full capacity for unoccupied spaces. Horizon links real-time occupancy data directly to building automation systems, creating occupancy-responsive infrastructure that scales energy consumption to actual human presence. When a floor drops below 15% occupancy, HVAC transitions to setback mode. When a zone registers zero presence for 20 minutes, lighting dims to security-only levels. The result: 18-28% energy cost reduction without any capital investment in equipment — achieved purely through intelligent control of existing systems.
Zone-level HVAC control — occupancy data drives variable air volume, temperature setpoints, and ventilation rates per zone. Empty zones receive minimum conditioning only. Saves 18-25% on HVAC costs
Presence-linked lighting — automated dimming and shutoff tied to sensor-confirmed occupancy, not motion detection timeouts. Paired with LED retrofit guidance for maximum savings
Carbon-per-occupant analytics — calculates energy and carbon intensity per actual occupant, not per square foot. Reveals the true environmental cost of underutilization
Building Performance Standards compliance — feeds occupancy-weighted energy data into BPS reporting frameworks for USA, UK, UAE, and EU regulatory obligations
25%
HVAC cost reduction
18-28%
Total energy savings
0
Capital equipment required
ESG
Scope 1 & 2 reporting ready
06
Portfolio Consolidation Intelligence
Multi-site comparison · Lease vs. utilization analysis · Right-sizing recommendations
The most expensive decision in corporate real estate is adding space you don't need. The second most expensive is keeping space you no longer use. Horizon's Portfolio Consolidation engine analyzes utilization across your entire portfolio — every building, every floor, every zone — and identifies consolidation opportunities that reduce real estate costs while maintaining or improving the workplace experience. It correlates lease expiration dates with utilization data, flagging properties where renewal would mean paying for demonstrably underused space. For multi-site enterprises, it models scenarios where teams can be redistributed across fewer, better-utilized locations.
Cross-portfolio utilization benchmarking — compares utilization rates across every property, normalized for capacity, function, and hybrid patterns. Identifies outliers instantly
Lease-utilization correlation — maps lease costs per square foot against actual utilization rates. Properties with high cost and low utilization flagged for consolidation before renewal
Consolidation scenario modeling — simulates relocating teams from underused sites to higher-performing ones. Models impact on commute times, collaboration patterns, and amenity access
Floor-by-floor stack planning — drag-and-drop scenario planning driven by actual utilization trends. Test densification, team co-location, and flex-space ratios before committing
15-25%
Portfolio cost reduction
22%
Average space consolidation
90d
Time to actionable recommendation
$4.8M
Average annual savings (500K+ sq ft)
07
Workplace Experience Analytics
Space-type preference mapping · Amenity demand · Employee satisfaction correlation
Space optimization without employee experience optimization is just cost-cutting with a dashboard. Horizon's Workplace Experience engine measures not just whether spaces are occupied, but which types of spaces employees actually choose — and what those choices reveal about work patterns. It tracks the ratio of focused work (single-occupancy desk time) to collaborative work (multi-person room usage) to social space (café, lounge, breakout areas) to understand what the workplace needs to provide. When organizations design based on assumptions, they build 80% desks and 20% collaboration space. When they design based on Horizon data, the ratio often inverts.
Space-type preference mapping — tracks which space types (open desk, private office, huddle room, phone booth, café) employees gravitate toward, by team, role, and time of day
Amenity demand intelligence — identifies which amenities (kitchen, gym, mother's room, quiet zones) are over/under-provisioned based on actual usage patterns vs. capacity
Acoustic environment monitoring — correlates noise levels (CO₂, temperature, humidity) with space avoidance patterns. Identifies zones employees reject due to environmental discomfort
Satisfaction-utilization correlation — integrates anonymous survey data with occupancy patterns to reveal whether high-utilization spaces are high-satisfaction or simply high-density
34%
Satisfaction improvement
2.4×
Collaboration space demand (typical)
18%
Return-to-office increase (post-optimization)
92%
Employee trust score (privacy)
08
Predictive Space Planning
Growth modeling · Hot-desking ratio optimization · Capital avoidance intelligence
The highest-value output of occupancy intelligence is not a dashboard — it is the capital project that never gets built. Horizon's Predictive Space Planning engine models future space requirements based on workforce growth forecasts, hybrid adoption trajectories, utilization trend lines, and planned organizational changes. It answers the question every CFO asks: "Do we actually need more space, or do we need to use what we have differently?" In validated deployments, this engine has prevented over $180M in unnecessary construction and lease commitments by proving that optimization could absorb projected growth within existing portfolios.
Growth absorption modeling — projects how many additional employees the existing portfolio can absorb through hot-desking, scheduling optimization, and space reallocation — before any new square footage is needed
Hot-desking ratio optimization — calculates the optimal desk-to-employee ratio by team, function, and hybrid pattern. Prevents the oversimplified "one size fits all" ratios that fail in practice
Capital avoidance scoring — for every proposed construction or expansion project, calculates the probability that optimization alone can deliver equivalent capacity. Scores proposals 0-100 on necessity
Scenario planning with confidence intervals — models best/expected/worst case space requirements at 12, 24, and 36-month horizons with quantified uncertainty bounds
$180M+
Capital projects prevented
92%
24-month forecast accuracy
1.8:1
Avg optimal desk ratio (hybrid)
40-60%
Growth absorbed without expansion
Architecture

The living building model

IoT SENSOR LAYER
100K+ data points/campus
EDGE PROCESSING
< 50ms latency
AI INFERENCE ENGINE
8 parallel models
DIGITAL TWIN CORE
2s sync · BIM/Revit native
ANALYTICS & PREDICTION
14-day forecasting
BMS / HVAC INTEGRATION
Occupancy-responsive control

Horizon's architecture is designed for zero-trust, edge-first processing. Sensor data never leaves the building perimeter until it has been anonymized, aggregated, and stripped of any personally identifiable information. The Rust-native inference engine runs all eight intelligence models in parallel with a memory footprint under 120MB — enabling deployment on standard building management hardware without cloud dependency. For multi-campus enterprises, federated twins communicate through encrypted API endpoints, sharing anonymized utilization patterns without exposing raw sensor data across sites.

Sensor Ecosystem

Every sensor type. One platform.

AI Optical Sensors
Low-resolution, edge-processed cameras that generate anonymized XY coordinates — not images. Distinguish active from passive occupancy. 1,000 sq ft coverage per unit. The gold standard for space-level accuracy.
97.8% accuracy · Privacy-first by design
PIR Motion Sensors
Passive infrared detection for binary presence/absence in confined spaces — phone booths, private offices, restrooms. Wide field of view, ceiling-mounted, battery life exceeding 5 years.
Ideal for enclosed spaces · Lowest cost per zone
LoRaWAN Desk Sensors
Long-range, low-power wireless sensors for desk-level occupancy tracking across large campuses. Single gateway covers entire floors. Battery replacement cycle exceeds 7 years.
Campus-scale deployment · Minimal infrastructure
People Counters
Entry/exit sensors for headcount flow and capacity monitoring. Accurate for building-level and floor-level occupancy totals. Often paired with zone-level sensors for complete spatial picture.
Compliance-ready capacity monitoring
Wi-Fi & BLE Analytics
Device-based presence estimation using existing network infrastructure. Zero additional hardware for trend-level insights. Best used as a validation layer alongside dedicated sensors.
Zero-hardware starting point · Network-native
Deployments

The buildings that learned to see

Global Financial Services · 2,800 Properties
$42M in annual lease costs eliminated — without losing a single desk anyone was actually using
A Fortune 100 financial institution with 2,800 properties across 40 countries deployed Horizon sensors across its North American portfolio of 14.2M square feet. Within 90 days, the platform revealed that true desk utilization was 28% — not the 61% reported by badge swipe data. The difference was passive occupancy: personal items at desks, recurring reservations never attended, and entire floors maintained at full operation for populations that had shifted to hybrid patterns. The organization consolidated 6 floors across 3 buildings, renegotiated leases on 4 properties approaching renewal, and reduced annual occupancy costs by $42M while improving employee satisfaction scores 19% by concentrating resources on the spaces people actually used.
28%
True utilization (vs. 61% reported)
$42M
Annual lease cost eliminated
6
Floors consolidated
19%
Satisfaction improvement
Major Research University · 340 Buildings
The $42M building that was never built — because the real problem was scheduling, not capacity
Faculty senate reported that classrooms were unavailable during peak hours. The provost's office had approved a $42M new classroom building. Before breaking ground, the university deployed Horizon sensors across 340 campus buildings. The data was unambiguous: 62% of existing classrooms were empty during peak hours. The problem was not capacity — it was scheduling. Popular time slots (10 AM – 2 PM, Tuesday-Thursday) were oversubscribed, while mornings, late afternoons, and Monday/Friday sat empty. The university redesigned its scheduling algorithm, increased effective capacity 18%, redirected the $42M to deferred maintenance and student services, and saved $3.8M annually in energy costs by linking HVAC to actual classroom occupancy.
62%
Classrooms empty at peak hours
$42M
Construction avoided
18%
Effective capacity increase
$3.8M
Annual energy savings
Technology Company · Hybrid Workforce · 4 Campuses
From "we need 3 new buildings" to "we need 3 fewer floors" — in 120 days of sensor data
A 12,000-employee technology company with 4 campus locations was planning a major expansion to accommodate projected headcount growth of 2,400 over 24 months. Horizon's Predictive Space Planning engine analyzed 120 days of sensor data across all campuses and produced a different conclusion: the existing portfolio could absorb 3,200 additional employees through optimized hot-desking ratios (1.8:1), conference room right-sizing, and the consolidation of three chronically underused floors. The company avoided $68M in construction costs, reduced its desk footprint 22%, and achieved higher employee satisfaction because the optimized spaces were better designed for actual work patterns — more collaboration space, fewer empty rows of assigned desks.
$68M
Construction costs avoided
3,200
Additional capacity unlocked
1.8:1
Optimized desk ratio
22%
Desk footprint reduction
From the Field

We were planning a $42 million new classroom building because faculty said they couldn't find space. Horizon showed us that 62% of our existing classrooms were empty during peak hours. The problem wasn't capacity. It was scheduling. We redesigned the schedule, increased capacity 18%, and redirected the $42 million to deferred maintenance and student services. That is the kind of decision that data makes possible and instinct gets wrong.

Provost
Academic Planning & Operations
Major Research University

Our badge data said 61% desk utilization. Horizon said 28%. The difference was $42 million a year in lease costs for space nobody was actually using. When I showed the board the sensor data — active occupancy stripped of passive inflation — there was silence in the room. Then: "How fast can we consolidate?"

SVP, Global Real Estate
Corporate Real Estate & Workplace Strategy
Fortune 100 Financial Institution

We saved $68 million by not building three buildings. But the real win was the workplace we built instead — fewer desks, more collaboration zones, better amenities, higher satisfaction. Our employees didn't lose space. They gained a workplace designed around how they actually work, not how we assumed they worked.

VP, Workplace Experience
Facilities & Real Estate Operations
Global Technology Company
31%
Avg. utilization (Americas)
$1.5T
Global waste potential
97.8%
Detection accuracy
$152M
Capital projects prevented
See Your Buildings Clearly

Stop paying for space nobody is using

Schedule a demonstration of Bastion Horizon — configured for your portfolio, your buildings, and your workplace strategy. See what your data reveals.

Or contact our workspace intelligence team at [email protected]