Pano AI speaks to Canada 2020

On October 23, 2025, Pano AI co-founder and CEO Sonia Kastner to the Canada 2020 audience in Ottawa.

Remarks to Canada 2020, October 2025

When we talk about wildfires, we use certain phrases. It was a force of nature. It moved like lightning. There was nothing we could have done. We use these phrases because that's how it feels when homes reduced to ash, when smoke fills the sky. When a town evacuates in minutes. It feels like something bigger than us, an unstoppable act of God.

Over time, that language starts to do something dangerous. It implies we're powerless that our only job is to keep responding to fires the same way we always have, even as they become bigger, faster, and more destructive. The truth is [00:01:00] we're not up against something unstoppable or the divine. Yes, our landscapes have become drier.

Our forests are overgrown, and climate change makes everything hotter, faster, and harder to predict. But those conditions are more catastrophic because of the way we're managing fire. We are still relying on outdated equipment, slow coordination and reactive thinking, and that's the part we can change.

When we label these events acts of God, we let ourselves off the hook. We miss the opportunity to adapt, invest, and innovate, and we simply can't afford to do that any longer. The good news is we don't have to. We now [00:02:00] have a multitude of new approaches for addressing every phase of the wildfire lifecycle from prevention and preparedness to response and recovery.

We just heard a little bit about that from bellwether just now today. I'm zooming in on one part of this trend. Fire tech, not because tech is a whole story. It's not. In fact, some of the new approaches are as low tech as a herd of goats that eat flammable leaves. But tech is the world I come from and it's an area with very exciting progress.

Before I continue, allow me to introduce myself. My name is Sonia Kassner. I'm the co-founder and CEO of Pano ai. Pano was founded in 2020 with my old friend Arvin Satya, and he and I had both spent decades in the tech sector, specifically smart home and Internet of Things. For example in 2014, I worked at Nest.

On the Nest cam, [00:03:00] one of the first AI enabled smart cameras, and at Cisco Arvin helped to wire up entire smart cities like Barcelona and Copenhagen. Then in 2020, we came across an executive order from the state of California that explicitly called upon the private sector. To develop innovative ideas for wildfire mitigation.

As we dug in more, we were shocked to learn that first responders were being sent into fast moving blazes with limited information. No real-time data, not even a shared map to us. This didn't look like an insurmountable problem. It looked familiar. A real world challenge that modern day technology could address.

Arvin and I looked at the state of tools available to fire professionals and thought we can do better than [00:04:00] this. And that's why we founded Pano. Now, the reason I was invited to speak here today along with Joshua is that, as you all know, wildfires are becoming an existential threat here in Canada and across the globe.

As you all know, the worst, the last few fire seasons have been some of the worst in history. In 2023, there were over 200,000 Canadians displaced from their homes, and millions impacted by smoke inhalation. And it's not just a western province's problem. In Quebec, 4.5 million hectares burned an $8 billion of property damage was incurred.

And underlying this destruction are some very worrying trends. For example, the vapor pressure deficit, which is a measure of how thirsty the air is, has gone up by 70% in some regions. This is a strong predictor of wildfire risk, and it's increasing as a [00:05:00] square of temperature rise. And another trend, urban Conflagrations, which we had all but eliminated in the 1930s, have returned, for example, the infamous Fort McMurray fire in 2016 and the Jasper Fire in 2024.

In short, the conditions are getting worse at an alarming rate, and the tools have not kept up. At Pano ai, we're working to address one very specific tools gap, providing actionable intelligence in the early moments of a fire. Often when it's just a couple of hectares, think of it like a nest cam for the forest.

We deploy. Deploy ultra high definition cameras on high vantage points like cell towers. We rotate the cameras every minute and upload the data to the cloud, as well as ingesting satellite data and [00:06:00] other useful feeds. We then run proprietary AI algorithms to detect new fire ignitions in minutes.

Each AI detection is confirmed by a human analyst using the camera footage. We then push out rich alerts to a wide array of fire professionals simultaneously. This allows them to coordinate an aggressive response and contain the fire while it's still small. Now you may be thinking that detecting wildfires is a simple and easy AI problem, but it's actually not.

Wildfires are relatively rare and they never look the same twice. They break out in different times of day and in many types of weather and terrain. Fortunately, Panos machine learning team has trained our models to handle this complexity. [00:07:00] It starts with our rich dataset. Over 1 billion images capturing countless fire types, terrains and conditions.

And we don't just use off the shelf models. We have built multiple custom models to detect smoke during the day and heat at night from both satellites and our own cameras. And now we have chosen to intentionally set our AI thresholds low to detect even the faintest hint of smoke. This means a lot of AI detections, but we don't alert on every single one.

Instead, we use a human in the loop review to filter out dust, clouds and other false positives. We have found that combining human intelligence with artificial intelligence yields the fastest and most accurate performance. So now let me give you an example of Panos early detection solution in action.[00:08:00]

On a 32 degree July day in Oregon's, the Yamhill County Pano detected the first wisps of what would become the Kch road fire. Our system sent an alert 14 minutes before the first nine one one call came in. Our alert included precise coordinates and high resolution imagery, allowing crews to respond swiftly.

The nine one one call on the other hand was vague and did not include the location or the severity of the fire. Thanks to our intel first responders reached the fire while it was still small and contained it to under one hectare, preventing it from becoming a major disaster. One fire chief told us if it had not been for Pano, we would still have been battling this fire days later.

Now, thanks to Suce success stories like this, I'm [00:09:00] proud to report that Pano is delivering its solution at scale. And this is thanks to early adopters like Rogers Telecom here in Canada, Xcel Energy in the United States and Southern cross forests in Australia. That's on the enterprise side. We also have customers who are government agencies like the Washington State Department of Natural Resources.

We currently have hundreds of camera systems deployed across one state in Canada, 13 states in the US and five states in Australia. We detected thousands of wildfires this summer. We have onboarded 600 fire agencies and thousands of fire professionals. We've signed over a hundred million dollars of customer contracts, and we're currently monitoring over 16 million hectares of land.

Pano has become a mature company with a proven commercial grade solution, and we are not alone. [00:10:00] I'm excited to announce that Pano is just one example of the wildfire tech innovation explosion, and you just heard that Bellwether is another example today as well. When we founded Pano, there were just 10 companies building fire tech.

Today there are 500. These solutions range from sensors that can detect a tree falling on a power line before it starts a fire to models that can predict where fuel reduction can have the greatest impact. There's even non-toxic retardant that can be used in no interesting ways. So you might be wondering if all this amazing tech exists, why haven't I heard about it?

Here's the truth. It's not the innovation that's lacking, it's the adoption. We can't just rely on innovative early adopters like Rogers Telecom and Washington State [00:11:00] Department of Natural Resources. We need wildfire tech to go mainstream, and that's why I'm speaking to all of you here today. We need federal and provincial agencies to embrace technology wholeheartedly.

Let me be clear. We will never stop every wildfire. There will always be wind, there will always be lightning. But in the world we're building, we don't wait until the sky turns orange to act. Fire doesn't catch us off guard. If we do this right, the next generation won't remember a time when we call these events acts of God.

Instead, they'll inherit better tools, better systems, and a clear plan for facing fire head on. We can prevent every spark, but we [00:12:00] can stop pretending we didn't see it coming. Thank you.