2015 Retrospective of Intelligent Assistants and Look Ahead to 2016

20162015 saw lots of buzz about artificial intelligence, deep learning, and automation. There were also many exciting developments in the world of customer-facing enterprise intelligent assistants and consumer-focused virtual personal assistants. I wrote about as many of these developments as I could throughout the year in this blog.

I recently published a guest post on the Opus Research site in which I share my key takeaways for the intelligent assistant space in 2015. Dan Miller of Opus Research published a companion article with his look ahead to 2016 and his very interesting take on how companies should be wary about making false choices when they implement self-service technologies in the New Year.

Companies that implemented or built upon existing enterprise intelligent assistance solutions in 2015 reaped big benefits.  In 2016, it’s time to keep pressing ahead or, for those who haven’t started yet, get caught up before it’s too late.

Talk to Miss Piggy: Chatbots Take Over Brand Promotion

Techcrunch’s Drew Olanoff broke the news a week ago that Miss Piggy of the Muppets was holding live one-on-one chat sessions with anyone on Facebook Messenger. That’s right. Miss Piggy! What’s it like to chat with the Muppets diva extraordinaire?  Well, I didn’t get to experience it firsthand.

Miss PiggyChatting with Miss Piggy was only available for a limited time as a promotion for The Muppets show that airs weekly on ABC. But based on examples in the Techcrunch article, Miss Piggy’s conversation was indicative of her infamous narcissism and over-appreciation of her high school French: along the lines of “Bonjour! This is such an honor… for you. How are vous?”

The Miss Piggy chat was driven by Imperson, a company enabling “bi-directional communication” with fictional characters. Olanoff also cites Disney Consumer Products and Interactive Media Labs as participants in the technology, though I’m not exactly sure of their relationship with Imperson.

On their website, Imperson lists both movie release campaigns and immersive brand experiences as use cases that benefit from chat sessions with fictional characters. Potential fans of an upcoming movie can chat with superheroes or creepy zombies and get a little thrill from the authenticity of the interaction. Imperson employs interactive script writers who design dialog that reflects a character’s fictional persona and stays entertaining..

I’ve written in the recent past about the growing popularity of chat as the new customer engagement channel. Not too long ago I wrote about work that Pandorabots is doing with the 2nd largest advertising agency in Japan to enable chat-based surveys on messaging service Line. Japanese brands want to engage consumers on the platform that they like best, which is text-based messaging. Like Imperson, Pandorabots helps to script the dialog that makes those conversations with chatbots fun and engaging.

Christopher Mims coined the term chatvertising in his WSJ article of July 2014. Mims was looking specifically at services on UK’s messaging app Kik that enabled young consumers to interact with branded characters. Mims suggested that messaging apps, and the conversations that occur there, are the future of social media and even of interactive marketing. It seems his observation wasn’t off base.

A Few Calming Words in Defense of Hello Barbie

Ever since Mattel and partner ToyTalk announced Hello Barbie, privacy advocates, parental groups, technology pundits, security gurus, and the media have been heaping vitriol upon the little doll with the push-to-talk button on her belt buckle. Now they’ve even mustered up a class action lawsuit. Based on the ferocity of the attacks, one would almost think Mattel was trying to sell toddlers a working replica of Jack the Ripper.

Hello BarbieHello BarbieI’m still trying to understand what it is about Hello Barbie that elicits such a gut-level aversion in people, because it’s clear the reaction is primarily emotional. The main complaints levied against the doll are that it violates a child’s privacy, is riddled with security holes, and is likely to crush a child’s innate creativity.

Oddly, I rarely if ever read complaints about children wearing Disney magic band tracking devices, spending hours immersed in iPad videos and apps, or having their essays scored by robo-graders.

Not that there’s anything innately bad about any of the above. We’re just living in a new world and it takes some getting used to.

Hello Barbie isn’t perfect, but it isn’t evil. Like her early precursor Chatty Cathy, Hello Barbie is Mattel’s attempt to push available technology as far as possible in the direction of giving children what they’ve always wanted: a doll that actually talks to them. It seems everyone is upset with the toymaker because this time they’ve come darn close to delivering on the promise.

Is that really such a bad thing? Youngsters lucky enough to be playing with Hello Barbie this Christmas will be of age to enter the workforce in 2027. Since there’s a high probability (99.999%?) that in 2027 employees in all sectors will need to understand how to interact seamlessly with language-capable intelligent devices, why not get an early start?

Children can learn from disappointments with Hello Barbie’s current conversational limitations and watch as future iterations of the doll mature in capability. As far as privacy concerns go, digital natives probably reveal more personal info about themselves in 20-second social media interactions than they will in hours of scripted play with a talking doll. These children will be adults in a world where they’re surrounded by recording devices and where everything they say may be searchable. As much as we may want to, we can’t stop this from happening. The best we can do is prepare our children to navigate successfully through this new digital landscape.

Are there real risks and concerns with talking toys? Absolutely. I’m more concerned about what the doll is scripted to say than I am about who’s listening to the conversation. But there’s a difference in discussing risks and blindly jumping on a bash Barbie bandwagon.

Take a deep breath and think before succumbing to the knee jerk reaction to retweet the latest alleged Hello Barbie mishap. Or put down Siri for a minute and ask Alexa to play you some relaxation music. There now. Feel better?

Jacada Visual IVR Helps Customers Discover the Benefits of Self-Service

We’ve all seen the growing interest that companies have in engaging customers through SMS and messaging-based platforms. We also know the trend towards customer self-service, which brings benefits to both the customer and the companies that serve them. But how about leveraging the old-fashioned IVR call tree to nudge customers into a more satisfying self-service experience?

That’s exactly what Jacada, the Visual IVR provider, helped one of its large customers accomplish.

Visual IVRJacada recently sent me a thick deck chock full of data from various customer case studies. A large credit card company had deployed a very popular app that was getting high usage rates. But the company’s data showed a sizable percentage of customers still tended to call live support specialists. Further analysis indicated that the majority of people calling in for support had questions related to their account balance or billing questions.

Jacada came up with a creative approach to help these customers understand the benefits of self-service. When customers called into the bank’s IVR, they gave them the option of pressing 1 to transition into a Visual IVR experience. What started out as a customer service experience from the 20th Century suddenly transformed into a customer helping themselves by using their smartphone. Nice!

Within the first 3 months of its introduction, a whopping 25% of the IVR callers actually took the bait and switched into the Visual IVR session. 66% ranked the service as “very good” or “excellent.” And ultimately, 69% completed their service experience without engaging a human call agent. Those are significant results.

One of the things that I really like about Visual IVR is the way it can lead a person into a rich, user-friendly self-service experience without them even having to think about it. Another example from the Jacada case studies dealt with a postal service, The service is using Jacada to embed Visual IVR into SMS so that customers can track packages directly from a messaging session.

That kind of integration between a technology people are comfortable with (texting) with one they may not be as familiar with (self-service) is a good thing. It helps people discover the many benefits of self-service while enabling companies to meet their goal of lowering support costs. Everybody wins!

Niki and WhatsBot – Smart Assistants to Chat With

Virtual digital assistants that you interact with through SMS or messaging apps are all the rage. A group at Techcrunch’s London Hackathon jumped onto the texting-bot bandwagon with the creation of WhatsBot.

Text BubblesAs a hackathon creation, WhatsBot’s capabilities were understandably limited. By adding the bot’s phone number to your contact list, users could include WhatsBot in a group chat. The bot would suggest convenient meeting spots based on the locations of all the group’s members. As an added bonus, the bot would chastise users if their texts contained profanity.

WhatsApp blocked the unofficial bot assistant soon after it was launched. Apparently it violated WhatsApp’s terms of use.

A team out of Bangalore, India has launched a completely different, but also chat-based assistant called Niki. Niki isn’t integrated with popular messaging apps yet, but that seems to be on the radar. Niki assists users with purchase transactions. Users in India can currently text with Niki to ask the assistant to add data to their mobile plans or to order a cab. Niki will continue to add new capabilities.

See my full story on Niki, the smart-purchasing assistant, on Opus Research.


Cortana Is Now Your Personal Enterprise Data Analyst

Cortana took another bold step into the enterprise today. I learned about the move from Frederic Lardinois’ TechCrunch article covering Microsoft’s announcement of Cortana’s integration with Power BI.

Cortana Integration with Power BII’d written previously about the Cortana Analytics Suite, which enables organizations to apply machine-learning algorithms to their data to produce valuable insights. In comparison, the Power BI and Cortana integration seems to be lighter weight and more readily accessible. Microsoft offers a scaled-down version of Power BI for free and it can easily be linked to datasets from Excel spreadsheets to enterprise systems.

Power BI takes data from your sources and presents it in meaningful dashboards, reports, and datasets. When Cortana is integrated with Power BI, it expands the tool’s capabilities to provide something akin to a personal data assistant. Power BI already had a natural language search feature called Q&A. Now Cortana leverages Q&A to answer questions intelligently. That means you can ask Cortana data-related questions instead of painstakingly creating and executing ad-hoc queries.

For example, you can ask Cortana to show you the average spending of customers in Florida by city. Without Cortana, you may have had to build your own query to mine this information from various datasets. With Cortana, your voice-enabled assistant creates the query for you on the fly and presents you with the results. That’s a significant time saver. It can also help you avoid fat finger errors.

You can ask Cortana questions with either voice or by typing. Cortana’s new abilities as a data analyst are impressive. Thankfully, we humans are still in charge of coming up with the questions that need to be answered.