Sentinel: The Hackathon-Built Lifeline for Disasters (Full Transcript)

Eleven Labs hackathon winner Sentinel uses voice AI to triage calls, share real-time guidance offline-to-online, and deliver responder dashboards when networks fail.
Download Transcript (DOCX)
Speakers
add Add new speaker

[00:00:05] Speaker 1: On December 11th, Eleven Labs hosted a worldwide hackathon. 1,300 builders came together in 33 different cities to develop projects and ship them in just three hours. In that time, teams created everything from gaming companions to shopping assistants and training programs for employees. Out of all submissions, worldwide, one team was selected as the winner. Please welcome Jason Yee, representing Sentinel, the global winner of the Eleven Labs Worldwide Hackathon.

[00:00:46] Speaker 2: In 2019, Australia experienced one of its worst natural disasters in history, the Black Summer bushfires. It burned over 24 million hectares of land, an area equivalent to the size of the entire United Kingdom. Imagine driving from London to Edinburgh and seeing nothing but ashes. And worst of all, 33 people lost their lives, many waiting for help that never came. Meet Sam. Sam is a father of two, and he looks outside and sees a glowing red sky. He prepares to evacuate his family, but when he turns to his phone for information, the screen is blank. Internet infrastructure is down, there's no maps, no updates, no escape, and no escape routes as well. And Sam's not alone. Thousands of families are also calling simultaneously, panicking, asking for simple advice, like is the road open, overloading the ready fragile network. This exacerbates an 8,400% surge in call volume. And in just 24 hours alone, traffic explodes from 164 calls to over 14,000, completely burying human dispatchers. The situation became so critical that the state emergency services were forced to shut down their systems just because they couldn't handle the call surge. And when Sam desperately dials for help, this is what he hears.

[00:02:11] Speaker 3: The line is busy right now. The estimated wait time is 15 minutes.

[00:02:17] Speaker 2: Time wasted is lives lost. Every second counts. And that's why we built Sentinel, the lifeline that never hangs up. Sentinel is an offline-to-online bridge that connects victims by delivering live, real-time data through voice, even when the internet is down. And when a network spikes, it scales instantly to answer non-emergency call volumes with ultra-low latency. And with phone calls using a thousand times less bandwidth than the internet, it actively preserves the already fragile network infrastructure. And Sentinel doesn't just talk. It supports rescue teams by automating location and victim data and summarizing the call context all neatly in a summarized dashboard. By using the latest 11 Labs voice model, we developed Sentinel to listen, understand, and coordinate rescue. First, it intercepts. When disasters strike and calls spike, Sentinel is the first line of defense, intercepting mass-volume calls that human dispatchers cannot physically handle. It then listens, provides guidance, and feeds live data to victims. Filtering out background noise and panic to extract verified data packets, and most importantly, providing location and summarizing the data of the victims. Afterwards, it charges the call, conducting real-time risk analysis to distinguish between questions like, Is the road open? to I'm trapped in a burning house, please help me. Escalating the true emergencies. And finally, rescue. It presents a verified data packet for a clean dashboard for first responders to see to make sure every second counts. And now let's go to a live demo. My name is Jason and I'm currently at the Magazine London Greenwich. There's fire on the horizon and I'm worried that it's bushfire. What should I do? I'm not in danger right now, but there's smoke on the horizon and it's coming towards my direction. Should I start evacuating?

[00:04:50] Speaker 4: I'm here if anything changes.

[00:04:52] Speaker 2: And just in case, where's the nearest hospital? And with this information, the information is now summarized onto a dashboard provided for first responders to see with the live map data as well. And please meet the team that brought Sentinel alive. So we only met four months ago at a different hackathon. And we have Andrew, who's our CFO specializing in financial software. And also Nilin, our CIO, who's an AI specialist. And Von, our CTO, who's focusing on building for an impact. Thank you very much and join us today and ensure that families like Sam never have to wait for help again. Thank you.

[00:05:39] Speaker 1: Thank you.

ai AI Insights
Arow Summary
Eleven Labs’ worldwide hackathon brought 1,300 builders across 33 cities to ship projects in three hours. The global winning team, Sentinel (presented by Jason Yee), addressed disaster-response communication failures highlighted by Australia’s 2019 Black Summer bushfires, where infrastructure outages and an 8,400% call-volume surge overwhelmed dispatch systems. Sentinel is an offline-to-online voice bridge that scales to handle non-emergency calls with ultra-low latency, uses far less bandwidth than internet traffic, and helps preserve fragile networks. It listens and extracts verified information despite noise and panic, performs real-time risk analysis to triage and escalate true emergencies, and summarizes location and context into a dashboard for first responders. A brief live demo showed the system providing guidance and nearby hospital information. The team—Andrew (CFO), Nilin (CIO/AI), and Von (CTO)—aims to ensure families don’t have to wait for help during crises.
Arow Title
Sentinel Wins Eleven Labs Hackathon with Disaster Voice Lifeline
Arow Keywords
Eleven Labs Remove
worldwide hackathon Remove
Sentinel Remove
Black Summer bushfires Remove
disaster response Remove
emergency dispatch Remove
call volume surge Remove
offline-to-online Remove
voice AI Remove
triage Remove
risk analysis Remove
low bandwidth Remove
real-time data Remove
first responders Remove
dashboard Remove
location extraction Remove
Arow Key Takeaways
  • Emergency communication can fail catastrophically when internet infrastructure is down and call volumes spike.
  • Phone-based voice systems can preserve bandwidth and remain viable when data networks are fragile.
  • AI can intercept and handle mass non-emergency inquiries, freeing human dispatchers for critical cases.
  • Noise-robust listening, information extraction, and location capture are essential during panicked calls.
  • Real-time triage and escalation helps route true emergencies faster.
  • Summarized dashboards for responders can turn chaotic calls into actionable, verified data packets.
  • Rapid hackathon-style development can produce high-impact prototypes addressing real-world crises.
Arow Sentiments
Positive: The talk frames a tragic problem (bushfires and overwhelmed emergency lines) but maintains a hopeful, solutions-oriented tone, emphasizing faster help, scalability, and impact through the Sentinel system and its hackathon win.
Arow Enter your query
{{ secondsToHumanTime(time) }}
Back
Forward
{{ Math.round(speed * 100) / 100 }}x
{{ secondsToHumanTime(duration) }}
close
New speaker
Add speaker
close
Edit speaker
Save changes
close
Share Transcript