Chatbots looked upon to ease burden of Healthcare Systems

The ongoing COVID-19 pandemic has put tremendous pressure on health care systems. Healthcare systems are experiencing pressure not only for critical care but from anxious public looking for answers.

In a research work carried out at the Kelley School of Business, Indiana University finds that chatbots working for reputed organizations can ease the burden of medical providers. And, can offer reliable guidance to those with symptoms. Chatbots are software applications that carry out online chats via text-to-speech or text.

To validate the efficacy of chatbots, researchers carried out an online experiment with 371 participants. The participants viewed a COVID-19 screening chat between a hotline agent – human or chatbot – and a user with mild or serious symptoms.

Online Screening Experiment validates Efficacy of Chatbots

During the session, the researchers examined if chatbots were persuasive, provided satisfying information that is likely to be followed. The results showed some negative bias against the ability of chatbots, perhaps because of recent press reports. However, when the perceived ability is same, chatbots were reported to be more positive than human agents. This is a good news for healthcare organizations that are struggling to satisfy user demand for screening services.

Meanwhile, the ability of the agent is the primary factor that drives user response to screen hotlines – chatbot or human. When ability of chatbot or human agent is same, users view chatbots same, or rather more positive than human agents.

Earlier, before the pandemic, chatbots were recognized as a technology that could increase the speed how people interact with researchers and obtain medical information online.

“Because chatbots can be scaled up, they can serve an unexpected surge in demand when is paucity of qualified human agents,” wrote the authors. Also, chatbots provide round-the-clock service at a low function cost, they added.

Chatbots lack Natural Conversation; Startup Vies to deploy Humanlike ones

For most, conversations with chatbots for customer service, resolving queries is not pleasant. First, chatbots are not humanlike, second, chatbots do not have conversational memory to hold more natural conversation. Citing this, comedian Bill Burr refuses to call automated customer service lines for queries. He fears, years later, on his death bed, all he would have in his mind is time wasted with chatbots.

So much so, the frustration of completing the most straightforward task using an automated customer service makes anyone question the purpose of life.

To address this, the startup Posh has entered the fray. The endeavor of the company is to make conversations with chatbots more natural and less maddening. The company is vying to accomplish this using an artificial-intelligence powered system. This system uses conversational memory that eases out for users to complete tasks.

Retaining Conversation Memory Feature of New Chatbots

“Previously, bots subsumed what the user said at face value, and would not connect with what the user said in the conversation,” said Posh co-founder and CEO. In places such as banks with tellers, or for customer service, what the user says in the past is very important. For such needs, Posh focused on to develop bots more humanlike. They did this by giving them the ability to remember historical information in a conversation.

For this, currently, Posh’s chatbots are used over more than dozen credit unions across text-based and voice channels. With the help of a well-defined help customer base, the company has allowed to train its system only on the most relevant data, and to improve performance.

As for the future, in a course of time, the founders plan to partner with other companies in a bid to gather industry-specific data. This would also to expand the use of Posh’s system without compromising on performance.