Issue 02
Inside Substance
Short, irregular updates from inside the company. For the people backing us, advising us, and watching us.
June 2026

FELIX ran: violations found, limits exposed

78 /100
High Risk  ·  Immediate Intervention Recommended
Risk Rate
~10/sec
Peak Danger
100/100
Footage Analyzed
8 min

What 8 minutes of footage revealed

FELIX ran for hours across the facility. In just the first 8 minutes of analyzed footage, the model produced a sustained risk score of 78/100: High Risk, immediate intervention recommended, with peak danger hitting 100/100. Across that window, the model fired roughly 10 risk signals per second. Every PPE gap, proximity violation, and crowd-density spike, on every frame.

I reviewed FELIX's output against my time running Amazon warehouses. The categories the model is flagging are what an experienced operator would catch on the floor. FELIX just does it at machine scale.

A human inspector sees this once a year, for an hour, on a scheduled visit. FELIX sees it continuously, on every frame. That is the underwriting signal we are building.

The score is built from 9 weighted factors. Not a qualitative checklist, but a quantified, persistent signal running on every frame.

V1's inference focuses on dynamic worker safety: PPE, crowding, proximity to equipment. Static facility hazards like unsafe stacking, fire egress, and capacity violations come into scope with V2's expanded sensor suite.

Inference Live
Note

The risk score reflects our initial scoring methodology. We will continue refining the model and factor weights as we scan more facilities and gather more data.

10 risk signals per second, for 8 minutes straight.

V1 revealed what matters

V1 was never meant to ship as a finished product. It was a test of how AI-driven risk assessment holds up in a real warehouse. Two observations stood out, and both shape V2 directly.

Learning I
Context Matters as Much as Detection
Detection is only half the problem. A warehouse is a dynamic, three-dimensional environment where every aisle, rack, and storage area contributes to the risk profile. A score from one viewpoint is incomplete by definition. V2's multi-sensor coverage expands visibility across the full facility, moving from isolated observations toward a complete view of risk.
Learning II
Precision Unlocks Trust
V1 confirmed we can build a digital map of a warehouse. The lesson was how much spatial precision matters: the closer a detected risk ties to a specific location, the more usable the information. Precision improves clarity, strengthens auditability, and gives both operators and insurers something to act on. The future of warehouse risk assessment is not just detection; it is a navigable digital twin of every facility.
FELIX V2 · Target: June 2026
Add 01
Additional sensors for multi-angle coverage
Brings facility-wide context to every detection. Multiple angles replace isolated observations with a complete view of the facility, producing a more defensible score and a harder case to argue against at renewal.
Add 02
Thermal camera
Hazard detection beyond visible light: heat sources, electrical anomalies, and fire precursors. The risks that produce the worst losses, invisible to both standard cameras and human inspectors.
Fix 01
Improved sensor fusion: LIDAR and camera working together
Spatial precision down to specific locations. Each detected risk maps to where it happened in the facility, not just when it appeared in a frame. This turns inference output into something operators and underwriters can act on, and is the foundation of the navigable digital twin we are building.

The goal for June: V2 built, deployed, and scanning. That scan is the real test of what the full sensor suite produces when everything works together.

V1 revealed what matters. V2 is built around it.

Two tracks. Both moving

Substance is running two GTM motions in parallel: brokerage and MGA. Here is what moved on each this month.

01Brokerage
Signed
First Broker Partner
Our first distribution relationship is in place. Brokers already have the relationships with the facilities we want to underwrite. Signing this partner gives us a path to clients without building every facility relationship from scratch. A real unlock for GTM.
Active
Three Carrier Conversations
Active dialogue with three carriers on brokerage relationships. This is the capacity side of the equation, the infrastructure for placing risk once we are writing it. These are not introductory calls; they are substantive conversations about structure, appetite, and how a FELIX-monitored facility looks to a carrier's underwriting desk.
Filed
Licenses in IL, NJ & CA
Applications filed in Illinois, New Jersey, and California: three of the largest commercial warehouse markets in the country.
02MGA
Building
Warehouse Legal Liability Program
This is our first proprietary program, where Substance is the underwriter, not just the risk monitor. We have committed to Warehouse Legal Liability as the first product line and are actively building the program toward launch. The long-term business is continuous monitoring as the basis for how risk is priced, renewed, and managed. This is the first line of that business.
Robotics / Data
From V1 to V2
  • V1 inference model deployed Done
  • FELIX V2: multi-sensor and thermal June 2026
  • Sensor fusion: LIDAR and camera Q3 2026
  • Digital twin from every scan Q3 2026
  • Full autonomy in the field Q3 2026
Business / Insurance
The platform behind the policy
  • First broker partner signed Done
  • Licenses filed in IL, NJ, CA Done
  • Carrier capacity locked in Q3 2026
  • Warehouse Legal Liability program live Q3 2026
  • Substance MGA tech stack (Policy admin system, Aether CAT model, Broker portal, Underwriting platform) Q3 2026

Three specific asks

If you are reading this, you are already in our corner. Here is where we need the most help right now.

More from Substance