How AI predicts hurricanes and answers calls for help in their aftermath

Hurricane Ian wreaked havoc in southwestern Florida yesterday with 150-mile-per-hour winds, knocking out power for millions of residents. The near-category 5 storm is being called one of the most powerful storms to strike the United States in decades. But officials say artificial intelligence (AI) is a new weapon in the fight to keep the lines of communication open during a disaster and also better predict the intensity of hurricanes in the future.

With emergency hotlines, hospitals and utility call centers being inundated with calls, speaking with a voicebot during a time fraught with anxiety and fear can help, said Sourabh Gupta, CEO of augmented voice intelligence platform Skit. (He added that Skit is not being used as an aid with Hurricane Ian.)

“Voice AI is capable of handling context-specific, effective conversations with customers, with no need for human intervention,’’ Gupta told VentureBeat. “In a sensitive or dangerous situation, voice AI can be used to provide customers with crucial information in real-time, answer questions and redirect the more complex calls to a human agent.”

Voice AI is faster and more efficient than the more traditional interactive voice response (IVR) technology at providing time-sensitive information and redirecting inbound calls to the right party, because it can handle a human-sounding, two-way interaction with customers, Gupta said.

The digital voice agent also sounds natural and empathetic, Gupta said. “Because of the agility of the technology, new use cases and conversational flows can be easily added through the platform on an ongoing basis.”

Modernizing contact centers

Augmented voice intelligence platforms like Skit are designed to help businesses modernize their contact centers by automating and improving voice communications at scale. “By enabling preemptive, intelligent problem-solving and seamless live interactions, we help businesses across industries streamline their contact center operations, reduce costs and enhance customer experience,’’ Gupta said.

The platform can be easily integrated with other applications and offers detailed metrics and analytics, according to Gupta. It supports multilingual, natural-sounding conversations that can be hyper-personalized depending on the use case and customer, he added.

Skit’s digital voice agent “can handle an unlimited number of inbound calls simultaneously, eliminating wait times, which can be particularly frustrating at a time of need,’’ Gupta said. “By handling the most common customer queries, the digital voice agent boosts the contact center’s containment rate and reroutes the most complex calls to a human agent.”

This does not take away jobs, Gupta said. “The voice AI is designed to lighten the workload of customer service employees and augment their work, rather than take over their jobs entirely. We predict that AI will enhance existing products, help organizations and employees maximize productivity by improving accuracy, and simplify existing processes.”

Gupta claimed that voice AI will “advance customer service and customer experiences as a whole” and actually help create new types of jobs. 

“For instance, the Skit digital voice agent helps automate nonproductive calls, allowing the workforce to focus on more value-added tasks,’’ he said. This creates better opportunities for human agents to focus on more significant and diverse functions.

Using AI to stay ahead of the storm

Before there’s a weather emergency, AI is also used to develop forecasting models that can help more accurately predict when, where and how strongly hurricanes may strike. 

In 2021, researchers at the U.S. Department of Energy’s Pacific Northwest National Laboratory developed a model to better predict hurricane intensity in both the near future and under future climate scenarios. The team used deep learning techniques to develop the model, which can, on average, more accurately predict hurricane intensity relative to models used at the national level — running on a commercial laptop, according to the PNNL.

The researchers fed information to algorithms that can detect relationships between hurricane behavior and climate factors like heat stored within the ocean, wind speed and air temperature, the PNNL said. The algorithms then form predictions about which path a storm may take, how strong it could become and how quickly it could intensify. 

The National Oceanic and Atmospheric Administration predicts that the intensity of hurricanes will rise, on average, by 1% to 10% in a warmer future, bringing with them greater destruction, according to models that project two degrees Celsius of global warming. So being able to predict where a hurricane will hit and how strong it will be could make all the difference in evacuating the right people at the right time.