According to Gartner’s Customer Service Chatbot Deployment Guide, while overall investment in chatbots is on the rise, over half of customer service support (CSS) leaders say they find low to moderate value in the chatbots they’ve implemented. And even though 70% of customers try self-service during their resolution journeys, only 9% of said journeys are wholly contained within those automated experiences. Regardless, over the next two years, 38% of organizations are planning to implement chatbots—a 40% increase in the adoption of chatbot technology.

Behold, the steaming cauldron of toil and futility at the center of the conversational AI marketplace. Roiled by the disruptive (and converging) technologies that are driving these changes (and their inevitable keystone position in business operations), this bubbling concoction feels new and unique. Turns out, it’s actually quite familiar. 

It’s hard to imagine businesses running without electricity now, but in the early 1900s industry struggled to adopt this distributed energy force in meaningful ways. When electric motors first became available, the factories that bought them just replaced a huge, central steam engine with a huge electric engine. Either engine type was put to work turning a central drive shaft that ran through the entire factory—sometimes even going outside and into other buildings. The factory’s tools were powered using belts running on the central shaft. To run one machine you had to run them all.

As Tim Harford explains in his BBC article, “Why didn’t electricity immediately change manufacturing,” factory owners had to think in new ways to take advantage of electricity. Electricity applied intelligently allowed factories to be cleaner, safer, and more efficient. “But you couldn’t get these results simply by ripping out the steam engine and replacing it with an electric motor … You needed to change everything: the architecture and the production process.”

Electricity can send power exactly where and when it’s needed, meaning factories could contain several smaller motors, each driving its own small drive shaft. The promise of conversational AI is similar. Successful implementation requires an open ecosystem where a shared library of information and code-free design tools make high-level automation and continual evolution an everyday thing. In such an ecosystem, intelligent bots are available to meet customers when and where they need help—sending power exactly where and when it’s needed. Those bots can use relevant data to create experiences that surpass what human agents alone are capable of providing. 

Hyperautomation is a far cry from one powerful engine turning all the wheels at once.

Most attempts at conversational AI fail because they begin with a gross misunderstanding of how to use it. There’s a ruinous and persistent idea that AI is meant to replace human workers. There’s no scenario where a call center can effectively replace its human agents with some sort of AI engine. Even point solutions that can tackle certain use cases are a stifled expression of conversational AI—like giant electric engines drawing too much power and delivering too little focused output.

By contrast, a company like Lemonade, which launched in 2016 as a tech-first rental insurance provider, exhibits the right approach to conversational AI. Their entire business is built around conversational AI. Customers love interacting with Maya, an efficient and engaging bot that can sign customers up to a policy in minutes. Employees have a friend in Cooper, an internal bot that helps developers manage work flows. By the end of 2018, Lemonade was insuring 425,000 customers across 24 U.S. states and expecting sales of over $57 million. Unburdened by the outdated corporate structure of insurance giants, Lemonade has added homeowners, auto, pet, and life insurance to the mix and expanded into the European market.

As Harford points out, once electricity was being used in ways that met its potential to revolutionize business, average wages jumped and hiring became centered on finding quality candidates over hiring in quantity. Electricity became an ally for trained workers, unburdening them and providing more autonomy. As more factory owners realigned their thinking about electric motors, new manufacturing innovations spread.

“You would think that kind of leap forward must be explained by a new technology,” Harford writes. “But no. The economic historian Paul David gives much of the credit to the fact that manufacturers had finally figured out how to use technology that was nearly 50 years old.”

Conversational AI isn’t half a century old, but too much time has already been spent swinging it around in the dark. Recognizing that conversational AI is meant to be an ally to workers is a start. It’s also critical to begin the hard work of connecting the people, systems, and things inside our organizations. In order to leverage electricity, factories needed extensive wiring that could distribute this new form of power. In order to leverage conversational AI, organizations need to create their own intelligent communication fabric that allows for meaningful conversations between humans and machines. 

With intelligent communication fabric in place, the lights begin to come on. It becomes far easier to conceive of new ways to use conversational AI. Designing, testing, deploying, and iterating on automations becomes a daily activity. Conversational AI becomes an ally that frees employees from mundane tasks and gives them more time to solve problems creatively. The air inside becomes positively electric.