--- title: "AXA Climate Part I: Building a Climate Risk SaaS in the Eye of the Storm" excerpt: "How we incubated a risky SaaS to handle global natural disasters, while hunting for product-market fit." subtitle: "" slug: "building-climate-risk-saas-in-the-eye-of-the-storm-axa-climate-part-1" date: "2025-04-11" category: "Product Case Study" tags: ["Incubation", "Geospatial", "SaaS", "Climate"] author: "Sébastien" image: "/blog/climate/part1.jpg" imageAlt: "iss056e162175 (Sept. 12, 2018) --- Cameras outside the International Space Station captured a stark and sobering view of Hurricane Florence the morning of Sept. 12 as it churned across the Atlantic in a west-northwesterly direction with winds of 130 miles an hour. The National Hurricane Center forecasts additional strengthening for Florence before it reaches the coastline of North Carolina and South Carolina early Friday, Sept. 14. https://en.wikipedia.org/wiki/File:Dramatic_Views_of_Hurricane_Florence_from_the_International_Space_Station_From_9_12_(42828603210)_(cropped).jpg#filelinks" bgFormat: "dark:bg-gray-950 " articleFormat: "" imageFormat: "contrast-150 grayscale " cardFormat: "" hideByline: false status: "published" readingTime: "10 min" --- # Building a Climate Risk SaaS in the Eye of the Storm (Part I)

iss056e162175 (Sept. 12, 2018) --- Cameras outside the International Space Station captured a stark and sobering view of Hurricane Florence the morning of Sept. 12 as it churned across the Atlantic in a west-northwesterly direction with winds of 130 miles an hour. The National Hurricane Center forecasts additional strengthening for Florence before it reaches the coastline of North Carolina and South Carolina early Friday, Sept. 14. https://en.wikipedia.org/wiki/File:Dramatic_Views_of_Hurricane_Florence_from_the_International_Space_Station_From_9_12_(42828603210)_(cropped).jpg#filelinks

## It always starts with a phone call… It's 3:04 AM. My phone rings. I'm on duty tonight. > Hello, this is Predict services. >We're issuing an alert for a tropical cyclone over the eastern Pacific, tracking toward Miami. Currently a Category 3: sustained winds at 180 km/h. Landfall expected in ~12 hours, based on NOAA and ECMWF forecasts. > You have three exposed assets. High risk of disruption, likely structural damage. No immediate threat to personnel. The system may strengthen to Category 4. Next update in 6 hours. I hang up, and open Argos to assess our clients' exposure.

This was a Tuesday at AXA Climate. But how does a PM end up on cyclone duty? ## Disasters by the numbers: why it's everyone's business Floods, fires, hurricanes, droughts — disasters adapt to every season and continent. And they're hitting harder, faster, and more often than ever: **Worldwide, 2023 alone saw $360 B in global losses**, with only half of that insured ([Munich Re](https://www.munichre.com/en/company/media-relations/media-information-and-corporate-news/media-information/2025/natural-disaster-figures-2024.html)). The impact for companies? The bigger your business, the bigger your risk exposure. Every new factory, logistics hub, or supply route becomes another point of exposure. And even when you're insured, traditional policies take months to payout, when restarting operations can cost millions per week. Worse: in many regions, even in France, insurance is no longer even an option. Premiums are skyrocketing, or coverage is denied altogether. "The risk is simply too high." That's why AXA launched something new. ## From Parametric Insurance to Climate Incubator AXA Parametrics began as a spinoff of AXA, focused on parametric insurance, a data-driven product designed for the age of climate extremes. The premise was simple but powerful: instead of waiting months for claims to be assessed, payouts could be triggered within days — using live geospatial indicators like wind speed, rainfall, or a cyclone's path across concentric impact zones. That solved part of the problem. But it didn't go far enough. Many businesses only seek coverage after they've already been hit. But when the risk gets too high, even parametric insurance stops being an option. Insurance alone wasn't enough. Businesses needed to **anticipate, prepare, and adapt to climate risk**. That's why AXA Parametrics evolved into AXA Climate, turning into a climate-focused incubator. The mission: launch and nurture new climate ventures across sectors, from crop insurance and carbon markets to flood risk, ESG compliance, and resilience consulting. Each initiative - internally dubbed *speedboats* - followed the same incubation process: > Pitch → 3, 6, or 12 months of incubation → Pivot, Scale, or Kill. > ## My mission? A bet on the future.

AXA

Climate

Cymo

Argos

One of those speedboats was called **Cymo** — and that's pretty much how it was pitched to me: > We're launching a global alerting service for natural disasters. We already have a team testing markets and hypothesis with white label products. And we need someone to build the real thing in parallel — a future-proof scalable SaaS platform and the team that goes with it. We gave it the codename Argos. So while we chase the market with shelf-ready products with Cymo, you build the foundation for the future, for Argos. > I was hooked, for sure, but my PM internal alert system went off. I raised the obvious concern: *"Isn't it risky to hunt for product-market fit with one team… while the other builds a futuristic platform?"* The answer was basically: *Shut up and take my money.* > "We're making a bet because we believe in this product, and we think like insurers. > > > Risk is just financial exposure. We've got the funds, so we have high risk tolerance." > I guess that was my first lesson in risk management. I signed off on that. I became the first PM of Argos, leading a team that would grow to nearly a dozen: three frontend, three backend, three data engineers, plus data scientists, a designer, and then another PM. A dream team, and a total candy shop for a curious, problem-solving mind like mine. I got to collaborate with Data Scientists, Risk Managers, PhDs, storm modelers, wildfire specialists… Walking on clouds? Sure, but we needed something more solid to build on: a market. ## Navigating the pivot storm And pivot we did. Over two years, Cymo searched for signs of market readiness — while we quietly built a flexible, under-the-radar SaaS platform designed to adapt to whatever stuck. We tested 10+ go-to-market strategies across industries, continents, buyer types, ICPs, and use cases. Some showed promise. Others hit hard blockers: low maturity, lack of awareness, or no real incentive to act. We learned. We adapted. Then we pivoted again, one canvas after the other.

## Our survival playbook: optionality & mitigation Now, how do you build and scale a complex, geospatial data product — handling the sensitive topic of global natural disasters — with no defined market, no users, and moving targets? ### Optionality by design While Cymo chased signals, we build the infrastructure to absorb them. But we couldn't build something new for every pivot. Nor could we ignore them entirely. So we adopted a strategy of loose coupling: stay close enough to learn, far enough to avoid disaster. We borrowed from the very problem we were solving: **detect risk, adapt fast, stay resilient.**

#### **Mission** Build and scale a future-proof geospatial SaaS platform for climate risk.

#### **Anti-mission** Survive pivots, and potential market failure.

#### **Disaster scenario** Building a product nobody needs, in a market that might not even exist.

### **Mitigation strategy in 6 principles** 1. **Proto-users = proxy stability.** We worked with AXA's internal risk managers as proto-users — stand-ins for real clients, to filter noisy market signals. They gave us continuity in discovery, with far less volatility than our shifting clients and prospects. 2. **Build the primitives.** No market = no MVP. So we started with Geospatial Product 101: → auth, geo‑search, asset mapping, import, exports… Core features any incubated speedboat might need in the future. 3. **Benchmark, baseline, plug-in** Catch me if you can → Competitors were years ahead, so we defined our baseline, built the essentials, and made it plug-and-play. 4. **Ingest and display anything** Insurance portfolios. Industrial sites. Weather and risk data. We built flexible pipelines to ingest, structure, and map any kind of spatial data, at any scale. 5. **Be modular > be monolithic** Every major feature - map, alerts, pipelines, APIs - was built to be swappable. Feature flags let us test & swap safely in production. GraphQL gave us flexibility across frontends and use cases. 6. **Stay agnostic** From design system to backend to data models, we avoided hardcoding any vertical. We had to be ready for whatever came next. --- ## What Survived All the Pivots? Read Part II. You've seen why AXA needed Argos—and how we set sail into uncharted waters Curious about what survived all those pivots?. In Part II, I crack open the black box:

AXA Climate Part II

Break apart in case of disaster

Peek into the app and stack that powered real-time climate alerts, thousands of events, and millions of assets, at 3 AM.