Overwhelmed emergency hospital hotlines. Shortages of doctors and nurses. No sector of the health care industry is being spared against the current pandemic, forcing it to quickly adapt in order to support the exponential emergencies.

• Baidu, a Chinese company developed a thermal sensor to detect, screen and single out individuals with high fever, with a margin of error within 0.05 degrees Celsius;
• Wuchang Smart Field Hospital in Wuhan has 14 robots as part of its medical staff who can disinfect, deliver medicine and track patients’ temperature changes;
• Providence St. Joseph Health System in Seattle is using a coronavirus chatbot tool to quickly assess potential symptoms instead of overwhelming the hotlines. So far, it’s had over 40,000 chat sessions since going live.

The power of AI is being seen before our very eyes. Unlike human intelligence, artificial intelligence can support near infinite queries, which is exactly what is needed during a pandemic with sudden and explosive surges in demand.

Financial services are certainly not being spared. Call centers are working around the clock to serve worried customers and business owners regarding loan defaults, mortgage rescheduling, and credit card fee payments, usually resulting in several hours of waiting time. Insurance companies are another industry struggling to deal with the pandemic and its resulting lockdowns: over-stretched customer service, staff shortages and insurance brokers unable to meet prospects face-to-face are just a few of the headaches to deal with.

So how can virtual chatbots help the insurance industry under unprecedented challenges?

Conversational AI is built with the purpose of offering customers optimal experience and enabling agents to work more effectively. High-quality chatbots are capable of:

• Reducing operations costs by 30%;
• Streamlining the pre-sale and the sale processes;
• Accelerating the underwriting and claim processes;
• Eliminating the waiting time completely.

However, most interactions with major chatbots in the insurance industry are robotic, outdated and sometimes even frustrating. As it turns out, chatbots are quite similar to humans, where they need to be trained and educated first, before becoming experts in a field. Only then can a virtual assistant accurately assess a customer’s needs.

The state-of-the-art technology supporting the bots is known as Natural Language Processing which has recently attained levels of human-like performance by applying deep learning (DL) and reinforcement learning (RL) algorithms in the NLP field. Olivo chatbot, powered and built by Koios Intelligence is an example.

We are now observing promising results, as large amounts of conversational data become available for training Deep Learning architectures. In technical terms, this means machine learning models effectively encode into words (latent semantic real vectors). The attention mechanisms in Deep Learning improved the contextual dimension, which enhance the classification of intent, as well as a prediction of the response, which results in significantly increasing the accuracy of a dialogue system.

The current crisis is now forcing entire industries, who have barely changed in decades to quickly adapt, and virtual chatbots seem to be a top contender in the game. However, despite their undeniable advantages, there should always be proper training beforehand and the right human/AI balance to ensure the most effective result.