Digital Assistant Microsoft Bots

Here’s how to bring on the bots – and make consumers forget they aren’t human


By Lisa Lacy, n/a

May 23, 2017 | 7 min read

Even though research and advisory firm Gartner says 30% of web browsing will be screenless by 2020, Purna Virji, senior Bing Ads PPC training manager, told marketers at the Inbounder on Monday (May 22) not to panic.

Consumers don't mind AI if brands can make them forget they’re not human.

Consumers don't mind AI if brands can make them forget they’re not human. / Pixabay:

Pointing to a Cortana video in which women plan a 30th birthday party in Las Vegas without any landing pages and instead rely on digital assistants to do the heavy lifting, Virji noted this is where consumers are headed, but she also reassured the Inbounder audience while there will be a shift away from screens, voice won’t kill screens – just like mobile didn’t kill desktop.

According to Virji, we are training artificial intelligence (AI) to be human-like so that it looks at something, understands it, recognizes it and responds back – like Microsoft’s, or the #HowOldRobot, which is learning to understand faces and tell users how old they look, and Xiaoice in China, a social assistant on networks like Weibo and Touchpal – and who Virji said is designed to be users’ friend. Like, say, if you tell Xiaoice you broke up with your boyfriend, she’ll put you on a 30-day breakup plan, Virji said.

As a result, Xiaoice has 40 million active monthly users – and 25% have told her they love her. That’s 10 million people professing love for an AI.

“Xiaoice is a sophisticated conversationalist with a distinct personality. She can chime into a conversation with context-specific facts about things like celebrities, sports or finance but she also has empathy and a sense of humor,” Bing said in a blog post. “Using sentiment analysis, she can adapt her phrasing and responses based on positive or negative cues from her human counterparts. She can tell jokes, recite poetry, share ghost stories, relay song lyrics, pronounce winning lottery numbers and much more. Like a friend, she can carry on extended conversations that can reach hundreds of exchanges in length.”

Xiaoice has even given weather reports in China and Virji said she is more human in her descriptions of upcoming weather and its impact on residents than actual human meteorologists. This, in turn, makes an interesting point: Consumers are willing to engage with AI if we can make them forget they’re not human, Virji said.

Chinese ecommerce site even implemented a Xiaoice marketing channel to provide product advice and help customers make purchasing decisions and Virji said it is one of’s most highly profitable channels.

This is in part because bots remove a lot of pain points. Like, say, when shopping for furniture, if a question arises about a given product or category, consumers have to leave the site and do a separate search for answers before returning – sometimes leading them to other brand websites. A bot, however, can secure the answer for that consumer – and it’s still using search, it’s just much easier for the consumer – and in the same branded interface.

Virji pointed to on-demand ride services like Uber and Lyft and said the reason consumers like them so much is because they don’t have to take out their credit cards at the end of the trip. And this seamless experience is what chatbots replicate.

Look at the Skyscanner bot, Virji said. A user can easily find a number of flight options – and even add his or her spouse to a chat when they are planning a trip rather than trying to coordinate schedules otherwise.

Bots could theoretically make the experience of buying tickets to an NBA game or a concert much easier, too, simply by asking a series of questions.

Digital personal assistants also know a user’s preferences, such as cuisine, location and favorite teams, as well as context – ie whether a consumer is just leaving home in the morning or working late at night. As a result, they can make recommendations and perform tasks, like suggesting takeout and ordering dinner for a user stuck in the office. Or reminding a user about a colleague’s birthday, finding a donut shop on his or her way to work and placing an order for birthday donuts that are ready for pickup upon arrival.

However, brands have to know the power and limits of speech. Ie, a bot may not be the best avenue when it comes to helping a user change a password if the password is long and cumbersome.

“Speech forces the bot to keep communication to a minimum,” Virji said.

And from here, simply put, brands and marketers need to bring on the bots, Virji said.

Indeed, Bing has made something of a big bot push lately, adding relevant chatbots to search results and encouraging developers to build their own bots.

Not surprisingly, Virji recommended tools like the Microsoft Bot Framework and Cortana Dev Center to build them, as well as app-building service Cognitive Services.

Virji’s chatbot checklist includes: a minimum number of steps, ease of use – and easier than the alternative – as well as intuitiveness. Ie, a user naturally knows what to do with the bot.

First, however, brands should start with whatever problem their consumers are facing – as well as how bots and skills can change their business for the better – and then choose the API.

In order to easily test bots, Virji recommended taking a site’s FAQ page and plugging it into a tool like, which pulls in the FAQ page and “spits out readymade bot that can answer lots of questions”, Virji said. This can help brands save time on customer service.

Digital Assistant Microsoft Bots

More from Digital Assistant

View all


Industry insights

View all
Add your own content +