Reading Shelf
This page provides everything from light and easy to deeply technical reading. Below is the under the hood; the why and how of Inari's technology and mission.
Tier 1 - The Premise
This is the orientation tier. These reads describe the problem Inari Watch exists to solve — that detecting wildfires is largely a solved problem, while turning a detection into coordinated action is not — and the commitments that govern everything built on top of it: bounded authority, a human on every consequential decision, and honesty about scope. Read this first; the rest of the shelf assumes it.
Frames the US Government Accountability Office's (GAO) findings and academic-review consensus in plain language. The core argument: detection technology is mature; decision and response workflows remain manual; the gap between seeing a fire and acting on it has become the binding constraint. This paper ends with one paragraph on what closing that gap looks like operationally. This is the lead piece — the one a procurement lead or journalist forwards to a colleague.
2 Page Brief - 3 min Read
1 Page Statement - 2 min Read
Our mission statement expanded into a one-pager that articulates what Inari Watch builds, who it builds for, and the non-negotiables — particularly the bounded-autonomy / human-authority framing. This is the document that answers "what kind of company is this?" for a thoughtful evaluator and preempts the inevitable "wait, are your drones autonomous?" question.
Tier 2 - The Architecture
The how. These reads describe the system of systems — three physical layers unified by a bounded agentic coordination fabric — explain why that structure makes safety a property of the design rather than a matter of trust, and lay out what it actually takes to stand the system up in a single protected region. This is the engineering and the deployment model, not the pitch.
The architectural argument distilled from our white paper §4 — the three physical layers (sensing mesh, aerial fleet, ground control) and the agentic coordination fabric that unifies them. Includes the SoS diagram, the end-to-end flow walkthrough, and the integration points with existing agency systems (CAD, ICS, ADS-B, NOTAMs). This is the piece a utility wildfire program manager reads before scheduling a discovery call.
6 Page Brief - 8 min Read
5 Page Brief - 7 min Read
The agentic coordination explained for a reader who has read about AI safety incidents and wants to know whether Inari Watch is the careful kind or the reckless kind. Articulates the six agent classes, the human authority boundary, the policy registry, and the append-only decision log. This is the piece that converts skeptics into evaluators — if you can convince someone you've thought about safety more rigorously than they have, the rest of the conversation is downhill.
Tier 3 - The Evidence
The tier written to be checked. These reads confront the published data head-on — including the findings that complicate the convenient story — to locate where the real leverage in wildfire response actually lies, and it grounds the one piece of future hardware in spray physics and canopy science rather than ambition. Bring a pen.
8 Page Analysis - 12 min Read
Quantitative analysis of where time gets spent between sensor detection and operator-authorized response in current agency workflows. Cites published agency timelines, the GAO Spotlight findings, and the cost-and-loss frameworks that translate saved minutes into avoided losses. This is the piece that grant reviewers and insurance partners read with a pen in hand.
The AKA introduced honestly: what it is, what it isn't, where it sits in the roadmap (Phase II, not Phase I), and the science underneath (Sauter Mean Diameter targets, canopy interception literature, kinetic atomization principles). This read dives into our long-term physical actuator within the larger integrated system. This is the piece for technical readers who saw "drones" and want to understand the engineering.
5 Page Paper - 7 min Read
Tier 4 - A Point of View
The stance tier. These pieces take clear, sometimes inconvenient positions: that the category does not need another detection platform, that "agentic" becomes a dangerous word when it travels without a boundary, and that an early-stage company should publish its misses and not just its wins. (Coming Soon...)