Intercom releases what could be the first civilized bot
The customer communications platform says its new Operator bot breaks new ground with rules for manners.
The civilized bot has arrived.
As bot makers begin to differentiate their wares, customer communications platform Intercom is out today with a bot — named Operator — which it says is the first one to care “what an end user’s experience is like.”
For its 17,000 or so client companies, Intercom provides a dashboard that receives live chat messages through a brand’s website chat window, emails, custom application or profile pages on Twitter and Facebook Messenger.
Although Intercom has provided some automated communications, such as messaging through those channels to collect email addresses, all other communications through its platform have been chats — either live or asynchronous, where the message remains visible and is answered later.
Operator is its first bot, and it is available for sites, emails or custom applications, although not yet for Twitter or Messenger.
“This bot cares more about user experience than it cares about getting all the info the team may need,” Product Manager Mark Ryan told me. The main goal, he added, is to augment live communications with the brand, not to automate them.
Operator has several key skills, each of which is tempered with rules so that it will avoid annoying the human user. Users can enter free form queries or responses, which Operator will process using Intercom’s internally built inference engine and machine learning.
Its initial skill set includes simple tasks like collecting contact info, setting response time expectations and answering questions by sharing relevant articles.
But it won’t keep asking for your email, it doesn’t interrupt when a customer or an agent is typing, and its messages are intended to be short, precise and adjusted for the person’s particular background or needs. It won’t keep sending messages if the user doesn’t engage in conversation.
Being helpful is its primary goal, not collecting lead info or pushing a new product.
It automatically tries to answer questions using a brand’s knowledge base, for instance, but it won’t suggest a related article if it sees that you were just in the help center, as you probably already searched. If a user immediately wants to talk to a live agent, Operator won’t try to block that effort by demanding — as some bots do — that the user talk to it first.
If you’re reading a document on a brand’s help center and click the “sad” emoji to indicate you’re not happy, Operator will show up to direct you to a live agent.
It’s designed to only suggest answers that are useful, Ryan said. Over time, its machine learning is expected to figure that out, according to users’ responding messages and to a satisfaction survey that Operator sends when the conversation is over.
“A lot of bots have rules,” Ryan said, “but they don’t respect the user and what they’re trying to achieve.”
He added that Intercom didn’t do an exhaustive survey of the thousands of bots out there to determine that Operator is the first to be well behaved. The claim is made because Operator’s principles of interaction make respect for the user and courtesy a central tenet, and Intercom is unaware of any other bot doing the same.
No data yet on whether manners help with a brand’s bottom line, like increasing customer satisfaction or brand recognition.
But, Ryan told me, “it seems like it’s the right thing to do.”