The VergeSense Large Spatial Model (LSM) is a proprietary AI foundational model purpose-built to predict how humans use physical space. Trained on 8 years of real-world occupancy behavior across 250M+ square feet of workplaces, spanning industries, geographies, and space types, it's the only spatial AI model of its kind. The LSM powers VergeSense Predictive Planning for enterprise customers and is available via API for technology partners.

A Usage Fingerprint is a compact behavioral profile that predicts how employees will interact with a particular office. Think of it as an office's behavioral DNA. It captures four signals: how people spread across space types, how rooms are actually used relative to capacity, how behavior changes under pressure, and which spaces get claimed first. Unlike static planning ratios, Usage Fingerprints reflect real behavioral patterns that shift as occupancy changes.

The LSM takes two inputs, space context (your floor plan, space types, layout) and behavior context (industry, team mix, work patterns,) and predicts how those people will interact with that space. It generates a Usage Fingerprint, then runs Monte Carlo simulations (1,000+ runs) to translate behavioral probabilities into a range of realistic, floor-level outcomes. This makes it possible to quantify questions like: How often will we run out of meeting rooms? Which space types bottleneck first? At what occupancy does user experience degrade?

Monte Carlo simulation is a method that generates many plausible scenarios by sampling from predicted probability distributions. VergeSense uses it because workplace outcomes aren't deterministic, meaning the same headcount can produce different experiences depending on meeting sizes, team behavior, and space preferences. By running 1,000+ simulations, the LSM shows the full range of realistic outcomes, not just a single "breakpoint" number. Think of it as replaying the same day 1,000 times to see what would realistically happen.

No. The LSM can generate forecasts from a floor plan alone, using behavioral benchmarks from its 250M+ sq ft training dataset calibrated to your industry and building type. This is called "modeled precision." When VergeSense sensor data or WiFi data is available, the model shifts to "measured precision," generating Usage Fingerprints directly from your workplace's actual behavior for higher fidelity.

Large language models are trained on text from the internet. They can summarize a report but cannot predict how consolidating a floor will affect employee experience, because they've never learned how humans behave in physical space. The LSM is trained on spatial behavioral data — real occupancy patterns from millions of square feet of offices. It predicts the interaction between people and space, which requires understanding probability, physical constraints, and behavioral dynamics that text models can't capture.

Partners can embed spatial intelligence directly into the platforms their customers already use. IWMS and planning platforms can add scenario modeling and breakpoint forecasting. Workplace experience apps can offer predictive busyness signals. Smart building systems can optimize HVAC and cleaning based on predicted demand. Design tools can score layout performance before construction. The API provides behavioral distributions, simulation snapshots, or derived planning metrics depending on integration needs.

Predictive Planning is VergeSense's customer-facing product built on the LSM, available via the Meridian Platform. It allows enterprise real estate and workplace teams to forecast space demand, run what-if scenarios, identify breakpoints, and quantify ROI, in hours instead of the months a consultant study would take. Every forecast, scenario comparison, and risk assessment in Predictive Planning is generated by the LSM's behavioral modeling and Monte Carlo simulation engine.

The LSM was trained on 8 years of anonymized workplace occupancy data collected by VergeSense sensors across 250M+ square feet of offices in 50+ countries. The training set includes 5,000+ diverse floorplans across industries and regions. This covers how people actually use individual spaces (desks, meeting rooms, focus areas, collaboration zones), how behavior changes across days, seasons, and occupancy levels, and how usage varies by industry, region, and space design. No personally identifiable information is used.

Yes. The LSM API is available now for technology partners. Through a simple integration, partners can bring spatial intelligence into the platforms their customers already trust. Contact VergeSense to discuss your use case and get API access. For full technical details, visit vergesense.ai.