VergeSense Predicts Human Behavior in Space and Why You Can Trust It

What Is Predictive Planning?

Predictive Planning is VergeSense’s next-generation approach to workplace and real estate planning. Instead of relying on static ratios, historical averages, or consultant rules of thumb, Predictive Planning uses AI to model how people will actually use space, before changes are made.

At the core of Predictive Planning is the VergeSense Large Spatial Model (LSM): a behavior-first AI system trained on one of the world’s largest real-world workplace datasets.

The Core Idea

Given a space and the people who use it, we predict how those people will use that space.

This is not generic “AI.”It’s not forecasting based on guesswork. And it’s not just extrapolating the past.

It’s about modeling the interaction between people and space, such as how work happens, where it happens, and how those patterns change as offices fill up or change shape.

Why VergeSense Is Uniquely Positioned

VergeSense has spent 8 years collecting behavioral data across more than 200 million square feet of offices, capturing how real employees use real spaces, at extremely high granularity.

That dataset represents:

• Millions of observed workdays
• Thousands of office layouts
• A wide range of industries, regions, and workstyles
• Detailed, space-level usage patterns

This “human behavior dataset” is what allows VergeSense to move beyond assumptions and say—with confidence:“We deeply understand how people work, and how they actually use offices.”

What the Large Spatial Model Predicts

The Large Spatial Model predicts behavior, not just occupancy. Specifically, it predicts:

• What types of work people need to do (focus work, collaboration, meetings, etc.)
• Which spaces they’ll choose for that work (desks, focus rooms, conference rooms, phone booths)
• How those patterns change when you change:
  • the space (layout, space mix, floors)
  • the people (team mix, industry, workstyle)
  • the load (how many people show up)

This is not week-by-week forecasting. It’s answering a deeper question:“Given a certain demand for work and a certain office design, how will people behave?”

How Predictive Planning Works

You can think of Predictive Planning as four layers working together:

1. Inputs: Space Context + Behavior Context

We take in two types of inputs:

Space (Design) Context

• Floor plans
• Space types and counts
• Layout and adjacency
• Enclosed vs open areas

Behavior (People) Context

• Industry or region
• Team mix (e.g. engineering-heavy vs sales-heavy)
• Categories of work
• Historical usage (when available)

2. Usage Fingerprints: The Behavioral Core

Inside the LSM, those inputs are fused into a Usage Fingerprint, a behavioral profile of how an office works. Think of a Usage Fingerprint as an office’s behavioral DNA.

It predicts, at a space-by-space level:

• The probability a space is used,
• How different types of work distribute across the office,
• Typical meeting sizes vs intended capacity,
• How behavior shifts as the building fills.

As VergeSense CTO and Co-Founder Kelby Green puts it:

“When one person arrives, where do they go first? What fills up next? And what runs out first? That’s what the Usage Fingerprint shows you.”

The Four Components of a Usage Fingerprint

Each fingerprint is built from four measurable signals:

1. Person Distribution: How people spread across space types as occupancy rises.
2. Typical Gathering Size: How rooms are actually used (e.g., a 10-person room hosting 3 people).
3. Load-Scaling Factor: How behavior changes under pressure (e.g., conference-room camping drops at high load).
4. Space Popularity: Which rooms get claimed first, reflecting real preferences.

Together, these reveal the hidden bottlenecks and inefficiencies that static ratios never catch.

3. Monte Carlo Simulation: Turning Probabilities Into Reality

The LSM outputs probabilities, which are powerful but abstract. So we take the next step.

Using a Monte Carlo simulation approach, we run thousands of simulated “days” in your office using those probabilities. Each simulation is a plausible snapshot of reality:

• Which rooms are occupied,
• How many people are in each space,
• Where congestion forms,
• What fills up first.

This lets customers answer practical questions like:

• How often will we run out of conference rooms at 80% occupancy?
  
• Which floors break first—and why?
  
• Where are we overbuilt or under-built?
  

4. What-If Scenarios: Re-Running Reality

Behavior and space are tightly coupled. You can’t just “tweak” an output with simple math.

So Predictive Planning works by:

1. Adjusting the context (people or space),
2. Re-running the model,
3. Comparing outcomes

Examples:

• What if this floor becomes more engineering-heavy?
• What if we remove a floor?
• What if we convert desks into collaboration rooms?

Because the model is re-run end-to-end, every scenario reflects emergent, realistic behavior, not spreadsheet assumptions.

Precision You Can Trust

Usage Fingerprints support multiple precision tiers:

• Measured Precision: Generated directly from VergeSense sensor or WiFi data, highest fidelity.
• Modeled Precision: Generated from VergeSense’s 200M+ sq ft benchmark dataset, highly accurate even without local sensors.

As compared to: 

• Legacy Baseline (What Most Companies Use Today): Static ratios and consultant rules of thumb, opaque, outdated, and not behavior-backed.

No matter where a customer starts, Predictive Planning provides a more transparent, defensible, and realistic foundation than traditional methods.

As VP of Product Kanav Dhir explains:“With Usage Fingerprints, assumptions aren’t arbitrary. You can trace every number back to real behavior and defend your planning decisions with confidence.”

Why This Matters

Most planning tools assume:

• every seat is equal,
• every room is used as intended,
• behavior is static.

Reality is messier. Desks sit empty while rooms are full. Small meetings spill into large rooms.Phone booths become campsites.

Predictive Planning exposes these patterns before they become problems, so teams can right-size layouts, rebalance space types, and design offices people actually want to use.

Looking Ahead

Predictive Planning is the foundation of VergeSense’s broader spatial intelligence roadmap.

As the platform evolves, customers will be able to model not just supply, but demand, connecting attendance policies, behavior, and space design into a unified, adaptive system.

The workplace is no longer static. Your planning tools shouldn’t be either.