The Future of Work Chatbots, Voicebots, and Virtual Assistants

LIMITATIONS REMAIN

There are limitations for what even the latest AI can do with any of these technologies, Tobey admitted. “AI needs to be trained like your best representative, which means that it knows and is empowered to say, ‘I don’t have an answer to that question’ when it doesn’t.”

Such training is designed to avoid “hallucinations”—incorrect, sometimes wildly incorrect, answers. However, some technology experts argue that hallucinations won’t be avoided entirely.

“Chatbots are in a good place right now,” Schneider said. “But they’re not fully living up to their promise when you interact with them.” Schneider explained that while the technology is available to have bots do much more, different companies vary in their acceptance and deployment of AI capabilities due to hallucination concerns, particularly in the regulation-heavy financial services and medical industries.

Depending on the underlying technology stack, a new chatbot can be added in as little as several minutes. For more complex uses, this can take up to a month.

“Every company is unique,” Tobey agreed. “Many that we are working with have some sort of other technology and data they want to integrate. That could mean that deployment might take days or weeks.”

Even with the more complex deployments with the lengthier time frames, a new chatbot deployment is still much faster than it had been only a short time ago. According to Tobey, this is because companies are using models already built with historical data, meaning that initial chatbot training is already completed. “So, you’re able to generate value from Day One,” she explained.

Schneider and other experts recommended deploying different chatbots and voicebots for different uses, rather than a chatbot designed to be a “jack of all trades.”

STAFFING IMPACT

With their capacity to work quickly and untiringly, these AI-powered devices will fulfill the work of many humans, though how many people will actually be displaced is a matter of much debate. In a CBS News interview, Klarna CEO Sebastian Siemiatkowski said AI could do the work of 700 humans. He and other technology experts are quick to add that many or even most of these people will move to more intellectually challenging jobs that require human intervention rather than responding to easily automated customer questions such as “What is my balance?”

“We’re dealing in some ways with worker shortages,” Tobey said, noting that while failing to add workers to handle all of the work going to AI-powered devices appears to be taking jobs away, instead, the jobs are changing. “Data scientists are becoming more like prompt engineers,” Tobey noted. “You no longer have to be able to write SQL queries. There are areas where some of the very routine and monotonous roles are evolving away. That does mean job elimination in some industries, but not all.”

For example, in customer service, there will be a melding of AI-powered bots and human agents, with each taking the lead for different types of interactions, according to Tobey.

“You can’t remove one and have the other take over everything. Humans are necessary to make the bots as good as they are. And bots are helping make humans more efficient and be able to do their jobs more effectively. There has to always be that synergy.”

2025 AND BEYOND

As the research reports above indicate, strong enterprise growth of all of these technologies is expected to continue at least for the next several years.

“I think one of the things you’re going to start seeing is a lot more buy in about breaking down silos between different departments and nodes of data to deal with customers in a way that’s going to revolutionize work,” Tobey said. “Many of the companies that we work with, particularly in the consumer-focused or entertainment industries, customer support operations, marketing and web departments, are getting together. They’ve actually never met before, and they’re working together.”

It's clear that the future of work will, in some areas, rely on chatbots, voicebots, and virtual assistants.