IBM Teams Up With Slack to Build Smarter Data-Crunching Chatbots

By Bloomberg

(Bloomberg) — IBM is teaming up with Slack Technologies Inc. to make it easier for companies to build custom chatbots into the startup’s workplace-messaging systems, the latest move by Big Blue to add more diverse business cases for its Watson artificial-intelligence technology.

The two companies will release a developer toolkit that includes Watson technologies and can integrate easily into Slack, they said in a statement Wednesday. International Business Machines Corp. will also build a chatbot — a conversation-based application — that will help IT departments identify and resolve issues without having to leave the Slack platform. The startup’s own customer-service bot will incorporate the Watson Conversation system, which includes technologies such as speech-to-text conversion and natural-language processing, or the ability for a computer to understand what a person is saying.

“Right now, if you say something to Slackbot in Slack, it will kind of blindly take what you ask it and do a search,” Slack Chief Executive Officer Stewart Butterfield said. The bot tries to match the query to a preset document and doesn’t have any AI technologies helping it decode the meaning of the question. With Watson’s technology, “the system will better understand and learn about the way users ask about things and will collect better results,” he said.

Sending messages using applications such as Slack or Facebook Inc.’s WhatsApp or Messenger has become an attractive communication method for both consumers and companies, leading to a boom in chatbots. Businesses are eager to interact with their customers on these platforms using bots, which can provide direct contact without requiring them to hire a bunch of customer-service representatives or salespeople to do the leg work.

A virtual agent that can answer questions and synthesize large reams of data is useful because “it’s a faster way to get people the right answer versus going through files and finding it themselves,” said David Kenny, general manager at IBM Watson. “It enables people to work with machines the way they work with each other — it’s a more natural way to work.”

Bot Expertise

Armonk, New York-based IBM has come across a number of old and new clients that are creating these kinds of conversational bots, building them from various Watson technologies — putting them together like Lego blocks. To make it easier for companies to build chatbots, IBM has also combined the relevant blocks to create a new service, Watson Virtual Agent, according to a separate statement. Companies using this service can pick a specific focus for the bot’s expertise, such as retail, and can train it using internal proprietary data, Kenny said.

The push in chatbots is IBM’s latest move to add more artificial intelligence to software, an extension of its overall bet that companies will pay for AI products that can automate business processes. The suite of Watson’s AI services is key to IBM’s strategy to drive growth in software sales, especially as demand ebbs for older products.

About 30,000 IBM employees use Slack’s messaging system, making it one of the San Francisco startup’s biggest customers. The goal of the new partnership for both companies is to use AI to quickly provide insights from the huge data sets collected by the messaging system, yet require a lot of processing power to analyze. Each Slack customer, for example, has archived messages about everything from human-resources questions to sales data to network issues, all information that can be fed to an artificial intelligence machine that can be used to advise people on later inquiries and problems — via, say, a chatbot interface. Both companies agree that they’re at the beginning stages of building new technology that can achieve that vision.

“It’s hard to overestimate the degree of difficulty or depths we have to go to,” Butterfield said. “IBM has huge investments in this area and they have a great idea of the opportunities.”