From customer service bots to Amazon.com Inc.’s Alexa, machine learning algorithms spend a lot of time holding conversations with people, and developers need powerful tools to build apps that run on those systems.

Communications platform Twilio Inc. today announced one such tool: Autopilot, which it calls the first fully programmable conversational artificial intelligence platform for building custom bots for chat, interactive voice systems on telephone and home assistance apps.

When customers want to interact with a business, and they call a hotline or open a chat window on a website, sometimes what they need is simple. It’s something that a bot can quickly understand and deliver the correct information.

However, for those times when the problem is complex and cannot be automated by business logic, Autopilot can hand off the customer to a service agent after collecting the initial contact data needed to describe the problem so the agent can jump right in.

Autopilot also allows developers to code once and then deploy across most communication channels — including telephony, text messages, chat, Alexa, Slack and Google Assistant. With this capability, the company said, developers can worry less about where the system is deployed and focus on customizing its ability to bridge the gaps among customer, machine and support.

“Machine learning is the most transformative technology of our time,” said Nico Acosta, director of product and engineering at Twilio. “However, until now, the tools available for building machine learning-powered conversational experiences have been too complex and not optimized for developers, which has led to poor customer experiences. We built Autopilot to make companies successful in building bots that delight users, instead of frustrating them.”

According to Gartner Inc.’s Market Guide for Conversational Platforms, the market is set for growth, because only 4 percent of enterprises have deployed conversational interfaces but 38 percent plan to or are actively experimenting.

Much of the reason for that is that Gartner predicts that by 2021, 15 percent of all customer service interactions will be handled by AI systems. That would mark a 400 percent increase since 2017. The proliferation of home assistants, such as Google Assistant and Alexa, have put the expectation that some part of customer service is “always on” and artificial intelligence can do this for business.

“There has always been a question of how companies treat their customers who want to contact them,” Acosta told SiliconANGLE in an interview. The implementation of AI can also go a long way to reduce the workload on call centers while also making customer support more of a priority. “There’s a tradeoff between how long they’re willing to let customers wait in a hold queue verses how many customer representatives they can hire.”

With Autopilot, developers gain access to an array of tools designed to allow them to quickly build smart chat applications that connect to a powerful Natural Language Understanding and Machine Learning system — with the power of that system providing the brain, developers can focus on building business logic and improving customer experience.

With AI, developers can also use existing company data to train bots, using similar machine learning algorithms, to trigger actions or replies based on natural conversations.

As for making a better customer experience, the AI subsystem allows developers to separate business logic from presentation. Using what Acosta calls “conversational style sheets,” developers can customer tone of voice, language localization, error and success messages to fit the audience of the bot. In this fashion, businesses can get a bot that represents them in a way that other humans understand.

With today’s announcement, Autopilot is now generally available to developers and businesses, further details are available on Twilio’s website.

“There is no cookie cutter solution that will solve everyone’s problem,” Acosta said. “Twilio believes that the future is in the hands of developers, and they need the tools to build that future and that future is conversational.”

Photo: Twilio