Twilio Hones in on Contact Center, AI, Payments
While Twilio made headlines on Monday when it announced its $2 billion acquisition of email API platform SendGrid, the company wasn't done for the week. Today at its annual Signal user conference, Twilio cranked out contact center, artificial intelligence (AI), and payment-related announcements.
Flex Goes GA
To start, today marks the general availability of Twilio's Flex contact center solution, revealed at Enterprise Connect 2018. Built on Twilio's API stack, Flex delivers a fully customizable contact center solution that natively includes workforce optimization through technology gained in the Ytica acquisition.
Demand for Flex has far surpassed Twilio's expectations, Devang Sachdev, director of product marketing at Twilio, told me in a pre-briefing. With early Flex customers including Shopify, U-Haul, Scorpion, and Lyft, more than 3,000 agents are live on Flex today, he said. Additionally, Twilio has been working to expand the Flex ecosystem, and now has more than 145 consulting and technology partners supporting the solution
Flex has taken the contact center market "by storm," with potential interest from the base of two million developers and 50,000 active customer accounts Twilio has built up over the past 10 years, said Sheila McGee-Smith, contact center analyst and regular No Jitter contributor, in a prepared statement. Flex offers a "perfect cloud contact center solution to replace aging inflexible premises-based systems," she said.
Developers can get started with Flex for free with 5,000 active user hours available for building and testing prior to having to commit to a contract. But once customers decide to buy in on Flex, they have two pricing models to choose from, Sachdev said. They can either pay $1 per active contact center agent hour or pay $150 per user per month. And they can add on speech analytics capabilities for an additional 25 cents per active user hour or $40 per user per month.
Channeling AI with Autopilot
In other news, Twilio has announced Autopilot, the evolution of the Understand API it previewed last year at Signal. AI will be the next thing to drive the communications revolution, Twilio CEO Jeff Lawson told the media at a press lunch yesterday. However, too many AI products on the market require a master's in machine learning to use well, he added. Autopilot is Twilio's solution for helping enterprises build "bots that don't suck," Lawson said.
Autopilot is a fully programmable conversational AI platform for building custom bots, IVRs, and home assistant apps. Developers only have to write applications once and can deploy to IVRs, SMS, chat, Slack, Alexa, and Google Assistant without additional code, Nico Acosta, director of product and engineering at Twilio, told me in a pre-briefing.
Autopilot is aimed at automating self-service and other initial information-gathering conversations with customers. Depending on the interaction, Autopilot can hand off a conversation, along with all the context, to an agent so the customer doesn't need to repeat any information already shared with the bot.
"Machine learning is the most transformative technology of our time," Acosta said. With Autopilot, Twilio's goal has been to optimize and simplify how developers create machine learning-powered conversational experiences, he added.
Autopilot is available today for developers in public beta, accessible via the Twilio Console, a widget in Twilio Studio (Twilio's drag and drop tool for non-developers), and integrated into Flex.
Partnering on Payments
Twilio also announced
"Historically, taking payments over the phone has been either insecure or painfully difficult and complex for businesses to support," said Kris Gutta, senior product manager of Programmable Voice at Twilio. "