Subscribe Us

Header Ads

Google, Microsoft And Startups Are Going To War On Chatbot Technology

Google CEO Sundar Pichai demonstrated Duplex, a cloud-based tool for building virtual assistants, in May 2018. (Photo by Justin Sullivan/Getty Images)

In the business of tech, battles are fought over research, but war breaks out when money is at stake. That’s finally happened for chatbots, or software that allows people to hold conversations with machines. 

Investors and the tech press overhyped bots a couple of years ago, but companies from insurers to health providers are increasingly using them to boost their bottom lines. Earlier this week Google planted a major flag in the war as it pushes further into the enterprise business dominated by Salesforce, SAP and Zendesk announcing bot technology for call centers alongside a raft of other machine-learning tools for its cloud customers.

Business can use Google’s tool, called Contact Centre AI, to attach a phone number to a virtual assistant that can take calls. It’s built on Dialogue Flow, which uses Google’s expertise in natural language processing. The technology is similar to Google Duplex, a cloud-based voice bot tool. Google CEO Sundar Pichai demoed it in May 2018 by playing a recording of Duplex calling a hair salon to book an appointment for its “client,” a person using Google Assistant. The synthetic voice sounded eerily human, even injecting ums and ahs into its speech. 

Google potential customers, or pretty much any company that has a call center or customer services department, can do something similar with Contact Center AI.

Research suggests chatbot technology is already improving bottom lines. Healthcare companies and banks who use chatbots to deal with customer queries can save around four minutes, or more than 50 cents per inquiry, according to a report last year by Juniper Research, a British market intelligence company. 

Helvetia, a Swiss insurance company that booked $1 billion in 2017 pretax profits, says it boosted its conversion rate for getting customers to buy a new policy when it deployed bots, with the help of Berlin, Germany-based startup Rasa, to chat to them via text message.

Ergo, a division of German insurance giant Munich Re, has been using Rasa’s chatbot technology to automate around 30% of its customer service inquiries and save money, says Rasa’s founder Alex Weidauer. 

The good news for these companies and others is that there’s already a vibrant market for selling virtual assistant technology. Among the big players: IBM sells access to Watson Assistant, Facebook offers free access to, with restrictions, Microsoft sells access to Luis via an API, and Amazon sells access to Lex as part of its larger AWS cloud service. 

Weidauer says his startup and the other smaller players in this race have an advantage in offering a more customized service. One of his customers, the smart-scheduling startup Meekan, switched to Rasa from Microsoft Luis to develop its robot assistant because the latter charges for each “call” to its API; the more bot users you have, the more you pay Microsoft.

Having switched, “we can scale up without worrying about surprise bills,” says Meekan’s CTO, Eyal Yavor.    

A jump in downloads to Rasa’s open source software, which usually sees a sliver of product managers buy a subscription, suggests its sales are growing. Weidauer says the strongest demand right now is coming from insurance, banking and, in the last six to 12 months, healthcare.

The British healthcare startup Babylon Health, which expectsrevenue in the tens of millions for 2018, bases its business largely on a chatbot it developed for giving advice on ailments. There are similarly automated symptom checkers from startups like Britain’s Ada Health and Israel’s K Health. The latter built its chatbot technology from a database of 2.5 million patient records spanning 20 years, obtained from Israeli insurers.

For both businesses, the chatbot aims to reassure patients with enough information to keep them away from relatively more costly (and often overworked) human doctors. Doctors have said that between 50% and 80% of their patients shouldn’t even see them, says Mark Tluszcz, a partner at venture firm Mangrove Capital who has invested in K Health. “Oursociety has become [a] hypochondriac,” and self-learning systems like chatbots can help address that collective need, Tluszcz adds.  

Gartner, the research firm,posted a survey of chief information officers earlier this year that showed 21% of enterprises plan to deploy some kind of conversational interface in the medium to long term. So far just 4% of comparison have done so.  

“Every company in the world will have some form of AI and machine learning embedded in their contact center in the next couple of years,” says Mikhail Naumov, the chief operating officer of startup DigitalGenius, which has built customer service bots for KLM, Unilever and Eurostar. “The contact center is ripe for AI disruption.”