Making Decisions with Social Data

“Data” is such a general concept–one can just about use words like “information” or “stuff” interchangeably. Of course, such data is widely available to brands who want to collect it, but it’s pretty useless unless they know what to do with it next.

The Big Boulder audience had the pleasure of hearing from panelists Gloria DeCoste from Nestle, Beverly Jackson from MGM International Resorts, and Brad Ruffkess from The Coca-Cola Company about how exactly such large global brands are using social data to improve their product experience for their customers.


When it comes to social data, it can be difficult to determine how to best analyze it in order to move forward and provide consumers with a more pleasing brand experience, especially for CPG brands such as Coca-Cola and Nestle. Do social fans arrive because they already love the products, or are they made aware of the products because of an existing social brand presence?

“It’s difficult for CPG companies to say, ‘We did this online and then a sale happened,’” said DeCoste. Due to the quick turnaround conversion nature of CPG brands, those attribution events are hard to properly document and understand. As the brand was able to scale and grow their penetration inside of the social sphere, they did eventually see the data tied to it: impressions online were converting to offline sales.

Jackson spoke from a different industry experience entirely, with a much more limited and targeted audience: “We saw that the people who were most likely to like or engage with our content also were not actually the people who were able to afford our brand or would ever book with us.” Jackson said that this led the resort chain to expand their thinking from “beds and heads” into a more fully developed portfolio approach: what are the other things we serve and sell and how can we leverage our social presence to increase awareness and conversions on those things?

A notable connection between all the panelists was their management of an actual portfolio of products as opposed to one singular product brand. “If a person drinks eight beverages a day, I want for them to drink products in our portfolio as many of those eight times as possible,” said Ruffkess. “Social data can give me the information I need to try and put those products in front of you, based on your other behaviors.”

These brands also attempt to harness the power of influencers by utilizing user-generated content. These kinds of posts are essentially “free content,” saving the brand the trouble of creating additional content, while also allowing the brand to reap the benefits of engagement or A/B testing: if this user’s photo of a fountain played really well with social followers, it can be used in other capacities to increase interest and engagement. When social listening produces insights about slang terms or phrases being widely used across the internet, brands can capitalize on this by implementing such phrases into their own campaigns, like Coca-Cola’s #ThatsGold campaign executed during the last Olympic games. Brands should, of course, be cautious about using fleeting slang terms, either inappropriately or too long after they’re “cool”: no one wants to be the brand using “Netflix & chill” incorrectly.

Examining social data can also help with knowing how to provide more personalized experiences for individual customers: does a customer love to bake cakes, but not cookies? Brands like Nestle can use these social insights to provide more highly targeted content toward customers who prefer one or the other, resulting in a higher conversion rate because of more accurately targeted content.

Product research & development seems to be an area shockingly left untapped by social data & insights. All of the panelists agreed: it would be an incredible move forward to begin using this data to determine what their customers like about their products, along with what they don’t like, in order to create new products for their portfolio. Unfortunately, none of the brands seemed to, as of yet, be implementing these social insights to this advantage.

“Marketing has always been about understanding your customer very deeply, at their core,” said DeCoste. The data brands are able to mine from their customers has seemingly unlimited potential, and even major brands are only truly scratching its surface.

Additional Takeaways:

  • While agencies are common players inside of major brand spaces, it’s important to know what is lost when another party is responsible for collecting and scrubbing your data before you get a chance to see it. What kind of interpretations are lost in translation?
  • Social data is an incredible asset where users freely give of their information, unlike other spaces. “My loyalty card at my grocery store is actually tied to my pager number from 1989,” said Jackson. “I don’t want people to have my phone number!”
  • When a brand has a portfolio of products widely differing from one another, social data can help direct the right products to the right people: family-friendly resort properties versus party properties, for instance.
  • “Our instinct was to put our name in every hashtag,” said Jackson of MGM. “But you don’t learn anything from your customers by forcing yourself on them. Be on the outside and listen first.”
  • The gaming industry is having to entirely change the way they operate. If millennials don’t know Wheel of Fortune like the generation before them, they’re less interested in playing a slot machine themed after that game. Millennials are less likely to want to sit at a slot machine siloed away from everyone else; the industry is having to accommodate by learning about how this generation wants to play games.

An Inside Look at Twitter

Out of all the conversation happening around Twitter and it’s future, one question seems to keep the Twitter-sphere buzzing: “Is Twitter Relevant?” For those still confused as to which answer is right, we can safely say “YES”. Chris Moody sat down with Twitter’s own Joel Lunenfeld, VP of Global Brand Strategy, to find out more about how the company is staying relevant in a changing society and give a glimpse into the world that is Twitter.


It’s hard to argue against the importance of a company that has essentially become the “cultural operating system” of the majority of the world, let alone a service that accomplished this in just over 11 years. Recently on Twitter, we’ve seen examples of this ranging from political figureheads holding unfiltered conversations and impacting a community, a global discussion about climate change, and even to Wendy’s bestowing a teenager from Nevada a year’s worth of chicken nuggets for receiving the most-retweeted tweet of all time. We suppose food does have a way of bringing people together.

As far as Twitter’s place in the world of consumers, there is a slight gap between how many people recognize Twitter as a brand (which is over 90%) and how people understand and operate within the platform itself. It’s unlike differing social platforms, many of which host users connecting with friends and family. In fact, within the past year, Twitter has refocused their public brand, which also included a change of their category in the App Store from “Social” to “News”. This plays along with the solidified idea that Twitter is a place where users go for Moments, to seek out trending news, and where open dialogue exists between users who had no way to speak with each other before.

Joel Lunenfeld also expressed that Twitter is constantly improving their platform. According to Lunenfeld, “there has always been more good than evil in the world but evil has had a larger marketing budget.” However, Twitter is breaking that trend and helping users gain back control over the narrative. We’ve seen this in times like the 2015 Paris attacks and the more recent Manchester attacks, where users come together and use hashtags like #prayforparis and #prayformanchester. Similar in idea, brands like Dove are also controlling a more positive narrative with their “#RealBeauty” campaign, which is optimistic, hopeful, and changing the messages said about young women in particular. Basically, all of this is only helping to create a better online environment. We look forward to Twitter’s continued future into cleaning up data from negative thoughts to spam accounts.

So, if you’re still asking yourself “Okay, but is Twitter dying? Everyone I know is saying that it is…” then let us put it this way: Twitter is the #1 tool for business, the company is made of people who truly care and believe in their mission and, most importantly, Twitter is a company that will outlive us all. Twitter “will change the mode of transparency for years and years to come”.

And this all makes sense, seeing as how Twitter is the largest public archive of human thought to ever exist. (And if that doesn’t give you excited goosebumps, we don’t know what will!).


Influencer Marketing

“Influencer marketing” is a buzz phrase we’ve been hearing in the marketing world for quite some time, but does it actually work, and to what end?


Devon Wijesinghe, CEO of InsightPool, would respond with an emphatic “YES”…provided you actually know what you’re doing with the data you receive.

Five years ago, it was common for brands or agencies to dip their toes into the influencer marketing space by asking for influencers in large pools: “We want mommy bloggers.” The problem with this old approach, Wijesinghe argued, is that it’s too broad–not all mommy bloggers are the same, much like no two mothers are the same or, if he is to be believed, no to women are the same.

“Don’t get me wrong,” said Wijesinghe, “the ‘Mom Mafia’ is important, but the old model completely misunderstands the way we target.” The former methods used to find influencers for a brand typically boiled down to age, demographics, and life stage. But audiences are complicated, he argued, and not all of an influencer’s followers will be interested in each and every thing they say.

The question in the new model, to continue using mommy bloggers as an example, is not who are the mommy bloggers, but how are they influential, and in what spaces? Are these bloggers focused on the day-to-day activities of child-rearing, or on cooking food children will eat, or on the brands they use in their house, or even on having a glass of wine at the end of the day and remembering that no one is perfect? When marketers get more granular in terms of the audience they want to attract, that’s when influencer marketing really starts to work its magic.

In becoming more laser-focused on a specific audience (mommy bloggers who have multiple children, are the primary caregiver, and prefer Cabernet to Chardonnay when they’re attempting to relax), marketers can begin to pull together data that will start to flesh itself out into a very specific taxonomy. Mommy bloggers who blog about being a mom in a rural or religious community versus mommy bloggers who blog about being a mom in a major city. When taxonomies become clear and specific, marketers can start seeing where the audiences get segmented based on the data: this audience only engages this mommy blogger when she talks about lipstick, while this audience engages her when she talks about making food for her kids. Clearly segmented audiences provide opportunities for highly targeted advertising with more focused voices: all keys to increasing awareness, conversions, and ultimately, ROI.

But what else can we use this kind of social data for in lieu of advertising? Wijesinghe stressed the importance of understanding the entire story before advertising even begins–a brand can think their audience desires one thing when, in fact, the data will tell a different story completely. He told a story of an automotive parts client who assumed that their audience of likely buyers would be men who liked NASCAR. When the data was actually collected and reviewed, however, it was found that their most likely buyers actually ended up being women who liked Formula One racing–talk about being off-base. Once the brand was updated on and embraced the data presented, they were able to switch gears, as it were, and advertising voices to cater to this new audience, which resulted in a 10X conversion rate.

When looking for influencers, Wijesinghe urges marketers to “go organic” whenever possible. Networks of influencers sound tempting: it’s easy for marketers to feel they’ve hit a gold mine of verified influencers interested in being paid for their endorsements of specific products. The problem with this, however, is the potential lack of true loyalty investment on the part of the influencer. For instance, if an influencer is endorsing a Mazda, but driving a Mercedes in their personal lives and is caught in the act, all trust and credibility is gone…and you can’t get that back. “Earned [influencers are] going to deliver what you really need in the end,” said Wijesinghe. “That way, [the endorsements] are just based upon the data as opposed to being based upon how much someone wants to get paid to talk to an audience about something they have no freaking clue about.”

“We have a tendency to make sweeping generalizations on race and gender when we really can’t,” he said. It seems, then, that the general theme of this year’s conference is emerging, and it’s simple: we can attempt to start with data, but we really need to start with a specific hypothesis of what we think is happening in our audience. Once we have a hypothesis, only then we can test it–the data that emerges will either confirm or deny what we thought to begin with.


Brands and Bots

Robert Stephens doesn’t enjoy doing things by the book. He dreamed up his first venture, Geek Squad, while riding a bike in a Minnesota winter (the same week that Marc Andreessen started Netscape). It sounds quaint now, but in 1994 the idea of providing computer help for average consumers was pretty odd.

By many measures Stephens’ new venture, Assist, is also pretty odd: the company is working to build a future where customers “never have to wait on hold”—and where they can accomplish any business interaction with bots and AI.

Stephens joined BBI board member Justin De Graaf on stage to discuss the future of automated commerce and care, and to share some of the wisdom he learned from Geek Squad—including the ten years he spent at Best Buy after his company was acquired.


So what’s on Stephens mind now?

  1. “Be happy bots are sh***y!” That means now’s the perfect time to work on them.
    In 2017, industry outsiders may find bots irrelevant or annoying; industry insiders might find bots limited in use and flexibility. According to Stephens: it’s okay if Machine Learning is more like “meh-chine” learning; that means it’s prime time for innovation.

    “We have a chance to kill the call center,” said Stephens. “That’s why we’re here!”

  2. When you collect data or feedback from customers, prove to them you’re using it
    In Stephens’ mind, one of the most empowering and freeing experiences of the early internet was package tracking—by equalizing information distribution between the customer and delivery companies, companies like FedEx could increase consumer trust.

    Stephens thinks bots can build a similar type of trust by providing data back to consumers, instead of just asking and then disappearing. For example, asking for product feedback? Tell customers when their comments were read, by whom, and what the company is going to do with it.

    In chatbot experiences, conversations should allow for the AI to acknowledge confusion and defer to the customer to put them in control. For example, a simple “I’m not sure if this is what you want, can you choose from these two options” can deliver a huge improvement in customer experience.

  3. Bots are the expression of a messaging platform’s API combined with the brand platform, so “make sure your brand platform cares about something!”
    The ability for a bot to conduct a transaction or provide customer service is limited by both a given messaging platform, as well as a company’s infrastructure. But Stephens’ firmly believes that one of the most important factors is non technical: a brand’s platform: the personality and passions that make it unique.

    Stephens believes this is particularly important with commodities. For example, if a consumer uses a chatbot on a messaging platform—like Twitter DMs, Facebook messenger, or Line—to order both an Uber and a Lyft, how will those experiences feel differentiated?

  4. The biggest predictor of innovation: “the curiosity quotient”
    During Stephens time at both Geek Squad and Best Buy, he found that the best quality in executives was the “curiosity quotient,” or how much people have a natural interest in their work and business. Stephens found, especially in corporate settings, that the quality of curiosity was more effective than any internal advocates for innovation, or external business threats that suggest an enterprise should change.

    For people who want to encourage their companies to invest in bots and AI, finding those with natural curiosity is going to be the best way forward.

    So choose your colleagues carefully! “The curiosity quotient,” Stephens said, “I don’t think you can create that in people. They just have it.”



Bots in the Enterprise

“Why should we care about bots?” This was just one of the key questions Tyler Singletary posed to Slack’s Amir Shevat during their fireside chat about “Bots in the Enterprise.” The two had an enlightening conversation that covered everything from the purpose of bots, to challenges, to how bots will impact human jobs in the future.

Through their discussion Amir noted there are two types of conversations as they relate to bots: topical conversation (which need lots of language processing) or task-led (which focus on making a purchase). Bots in the enterprise will mostly be task-led he noted, focusing on buttons & choices, not on processing actual conversations which requires greater AI capabilities.


Bots serve the purpose of exposing a brand to the user to ensure they have what was described as a “delightful engagement”. But what exactly does this mean? Amir notes that the engagement needs to be personal, not intrusive, and giving the right service to the end user. Another key set of considerations for bots in the enterprise are all around the attributes of the bot itself: brands should think about the “environment” (is the purpose of the bot for work or commerce), “gender” (most languages other than english only have only “he” or “she” gender specific tenses), and “personality” (a sassy tone might make sense for a consumer brand but might not be as relevant for a bank).

Amir went on to discuss how the current challenge with bots is not a lack of engagement per se, but rather a lack of awareness that people can and should engage with bots. In his opinion, they need not a giant volume of bots to choose from or engage with, but rather high quality bots which provide a great experience.


AI and Machine Learning in Social Data

Why is it that when Anne Hathaway trends on social media, Berkshire Hathaway sees a pop in the stock market? According to the CEO of Converseon, Rob Key, it is simply the result of primitive AI gone wrong. Hedge funds buying shares of Berkshire Hathaway based on an uptick in “Hathaway” mentions is a great example of how Artificial Intelligence gone wrong can sometimes mislead us.

Despite the recent rise in popularity, AI isn’t new. In fact, it has been around for decades. Millennials were interacting with “chatbots” back in the late 90’s with the rise of AOL’s Instant Messenger (AIM) and infamous chatbot, SmarterChild. As Madeline Parra, CEO and Co-Founder at Twizzo, pointed out, SmarterChild may have been a popular chatbot, but it’s biting sarcasm was too much for tricking your parents into thinking it was you.


AI and Human Interaction

That same challenge of blending AI and human interaction still exists today. As Joshua March, CEO of Conversocial, put it, “From a customer service perspective, a pure bot will fall over at some point.” Thus, the opportunity is really about combining the unique qualities of bots and human agents to create the best possible customer experience. The benefits of AI combined with human customer service can lead to faster response times and increased agent efficiency.

We’re talking about social data here, but really we’re talking about human communication. – Rob Key

Optimizing Content and Interactions

More and more complex customer service interactions are happening through chat. It’s clear that customers are increasingly choosing chat as a way to engage a company or brand. This has created a unique opportunity for AI to help support the customer service agent handling these conversations.

AI is moving beyond simple boolean rules to derive sentiment. The challenge now is, how do you correctly capture emotions through a simple sentence or 140 characters? Boolean queries offer a great starting point, but it must be handed off to AI for deeper analysis. The advancement of AI is enabling companies to capture expressions of feeling such as frustration and trust.

It can also surface important context, not just sentiment but the underlying reason for that sentiment.

Public vs Private Interactions

Twitter was one of the earliest to see organic customer service interactions take place on platform. As it is predominantly public, this incentivized companies to not only take notice, but try to resolve customer issues quickly and effectively. As these interactions move more towards private channels (e.g., DMs, Messenger), it becomes increasingly important for companies to have a strategy for public and private interactions.

As Jonathan Farb of ListenFirst Media explained, “Mistakes come from applying the same model and approach to both public and private context.” Private context enables companies to leverage customer data to help expedite the interaction. This opportunity for personalization will come from tapping into the company’s own data (e.g., CRMs). When done right, this personalized interaction can lead to an amazing experience for the consumer.

A Call for Standardization

Across technology as a whole, we’ve seen open source tooling made available to speed up development time and solve common problems. Within the AI space, that simply hasn’t happened yet. There is a lot of duplication of efforts going on. Everyone that does AI suffers from common issues, so having more of a community to manage what might be adult content for example, would greatly help out innovation of how AI gets adopted and rolled out.


AI – Fad or Future?


“Say Data one more time!” With a poignant throwback image, this was just one of the rapid fire slides Lux Narayan, CEO and Co-Founder of Unmetric Inc. gave during his “Pecha Kucha” style talk (a presentation format for delivering 20 slides for 20 seconds each) on AI and social data.

Lux’s presentation began with a (literally) illustrative example of two artistic images: one drawn by a human and one created via AI, which the audience had a hard time distinguishing between. This touched on the theme of the presentation around AI being able to help businesses & marketers tremendously, assuming they can ask the right questions to help make AI smarter.

Lux presented a balanced viewpoint, also touching on some examples of things going wrong, whether it was a chatbot taking on a life of it’s own (for the worse) or kids being able to proactively order a brand new dollhouse via a voice enabled digital assistant (without the knowledge or consent of their parents!).

Yet ultimately the benefits and applications outweigh the potential for abuse, which Lux displayed with his two “key triangles” for AI. With examples ranging from the New York Times using AI data to determine what stories to promote through their social media channels or airlines using AI to relieve & handle some customer service requests to their human agents, the applications are nearly limitless.


The Changing Interplay Between News, Government and Society

While we were enjoying Big Boulder 2016, a global event of far-reaching implications took place halfway across the world: Britain, against all polling predictions, voted to leave the European Union. Only five months later, the United States experienced our own upset with the presidential election, resulting in an outcome that even our best data analysts didn’t see coming. The aftermath of these events left a lot of people asking the same question: How were we all so wrong?

IMG_5422Deb Roy, Director of the Laboratory for Social Machines at the MIT Media Lab and Chief Media Scientist at Twitter, was already uniquely set up to study this very question. Using the collective power of the MIT Media Lab professionals, the team began collecting data in the months leading up to the election, attempting to determine the outcome. When those results were flagrantly different than anticipated, the team wanted to know why.

From August 2015 to the US Election Day in November, Roy’s team documented one billion tweets discussing the election specifically. They sought to map out what they called the “horse race of ideas,” filtering tweets through their deep learning network and separating them into topic classifications. Yes, the social media public was talking about the election online–but what do we talk about when we talk about elections? The deep learning network would “read” news sources across the political spectrum, from Huffington Post to Breitbart, and then “listen in” to the Twitter users as they chatted about those same conversation topics.

As the tweets were mined, the network narrowed the list of topics down to 19 particular conversations ranging from gun control to immigration to education and race. Using these tweets, the MIT Media Lab constructed debate briefs for presidential debate moderators leading up to the election, allowing moderators to comb through and select questions that would matter most to the American people.

The problem, they would eventually find, is that the topics that were the most important to American citizens were not necessarily the topics that were being discussed in common news coverage. Even more challenging was the discovery that users had a tendency to read only what they agreed with, following primarily their candidate of choice and, perhaps unwittingly, committing themselves to an insular “tribe,” as Roy called it.

With regard to media coverage, Roy made two important observations surrounding the role that journalists played in election influence. First, per a Pew study that year, more than 70% of Americans got their news from television. Whether the media loved or hated soon-to-be President Trump, he was the candidate who easily dominated television coverage across outlets, partisan or not. With 70% of potential voters taking in a 24-hour news cycle disproportionately covering one candidate, it’s not hard to understand how that candidate stayed top-of-mind for many.


Roy also noted that conversation is able to destroy brands and people, much like the conversation surrounding the very public snafus of brands like United Airlines or Pepsi…but the conversation is still centered around those brands or people. The old adage goes: there is no such thing as bad press.

The second major observation Roy discussed was the disparity in the type of topics covered by media outlets versus the topics that social media users seemed to care about. For instance, in the weeks leading up to the Vice Presidential debates, over 30% of media coverage revolved around the VP candidates; on Twitter, only 3% of conversations seemed to care about these candidates at all. Additionally, the data showed a major divergence in topics discussed by journalists online versus civilians: while campaign finance was heavily covered by the media, users online seemed disinterested in those talking points, preferring to focus on conversations about race.

So what does it all mean for the future of media coverage as it relates to the public?

First, we must start bridging the gap between journalists, pollsters, and the general constituency. Roy noted that 80% of journalists live in 1 of 3 major cities in America, not leaving much room for the topics that rural Americans care most about. “The data shows that people in rural Wisconsin don’t care about Russia,” he remarked. “They care about local issues that affect them.” Roy and the Media Lab are working on ways to show journalists and pollsters their own research bias through network data visualization. Additionally, they are beginning pilot programs to build networks of influencers in rural America whose voices are respected by citizens in those areas, and need to be heard.

Another point of interest is exposing social media users to their own “tribe,” and the information bias to which they leave themselves vulnerable. The Media Lab quietly released a Chrome extension in 2017 called FlipFeed, which allows Twitter users to see into the feed of a user completely unlike them. The user experience feels like ChatRoulette: a user can click “Flip My Feed” on their Twitter interface and the extension, using social listening tools to determine what tribe that user may belong to, “flips” to the feed of a Twitter user in a distinctly different tribe. The extension gives users a view into worlds they could otherwise ignore, which would only create wider expanses between people groups and political ideologies.

Finally, the question remains: how now should news be created and affected? Should journalists lead by following, listening to social media users and the topics that are important to them, and creating story topics from those conversations? Do we need to retire the old way of news creation in favor of listening to the data in front of our faces? Most importantly, what will this take and how will it affect publishers?

What we do know is that post-game analysis shows us the data we didn’t know existed, and we can’t afford to ignore it any longer.

Additional Takeaways:

  • Social conversation online is evolving so rapidly that data collectors have a hard time maintaining the accuracy of which topics are relevant; the accuracy of that relevance has a tendency to drastically decrease within a two-month period.
  • Polling data differed starkly from Twitter conversations as well: the Media Lab found that Twitter users were very interested in foreign policy, while polling data indicated the opposite.
  • The MIT Media Lab team found that if Twitter users followed only one of the 19 early primary candidates, those users also tended to lean toward voting for that candidate.
  • Surprisingly, most “Sanders tribes” also had two-way connections with “Trump tribes.”
  • Through looking at network data visualization, it was discovered that most journalists tended to follow Twitter users connected to nearly every tribe but the “Trump tribes.”
  • Citizens have begun to tire of “mainstream media agendas,” Roy notes. “There’s a real feeling that, my God, we’re being manipulated here.” When this happens, citizens simply create “new” media, giving credibility to previously discounted outlets and giving rise to “fake news” phenomena.
  • There were more people who followed both Trump and Clinton than just one or the other. Determining language and intent played a big part in parsing out this demographic: for instance, did they use the word “illegals” or “undocumented?”



AI and Cognitive Solutions

It’s hard to get people to learn about – and be interested in – any new product. It’s perhaps even harder for B2B products, which tend to be less intuitive (and let’s just say it: less fun to think about).
IBM’s Watson hasn’t had this problem. Though it has no user interface and isn’t a consumer product, the bundle of Artificial Intelligence and Machine Learning services known as Watson is arguably a household name (thanks in part to a successful run on Jeopardy).


Beth Smith, General Manager of Technology at IBM Watson, joined Big Boulder Chairman Chris Moody on stage this morning to discuss the state of AI.

What do the uninitiated need to know about AI in 2017? Here are four places to start:

  1. AI helps read through data humans don’t have time to review
    AI technologies like Watson are (rightfully) associated with the act of thinking – commonly referred to as “cognition” within the industry. But one of the first places AI tools can help is just by reading at scale: humans don’t have the time or capacity to review millions of pieces of information when starting a project – Watson does.

    Researchers at the University of North Carolina’s Cancer Center trained Watson on all their literature on diagnosis and treatment of the disease; they then had Watson review patient records and treatment plans. Watson made the same decisions as the Cancer Specialists 99% of the time. For 300 patients, the AI suggested alternative courses of treatment that it discovered with its big data literature review.

  2. You may hear about AI a lot, but tools like Watson aren’t yet used to their full potential
    As it stands, business leaders are aware of Watson and its peers, but few have grasped the field’s near limitless applications. So to-date, most AI implementations have focused on the most intuitive use: deepening customer engagement.

    Chatbots and other consumer-facing, digital experiences are where most people will experience Watson in 2017. And IBM has big goals to make sure Watson’s “talked” with everyone: by the end of 2017, Beth and her team are working to make Watson speak with a billion people around the planet.
  3. Businesses that succeed at implementing AI have both interested executives and folks in the trenches
    According to Beth, it’s not enough to have executives interested and aware of AI technologies: the people “on the ground” need to be fluent and interested, too.

    “Data Science teams seems to have exploded over the past two years,” said Beth. In her experience, this bodes well for adoption of tools like Watson. “Ideally, there’s people playing in the data sandbox already” with a solid understanding of how Machine Learning and AI work, so when more complex systems like Watson are implemented, there’s less to learn.

  4. AI doesn’t replace people – Watson still needs teachers
    Watson is a quick learner, but it still needs a place to start. When Watson gets deployed, it’s still  built on a human training the system.
    “Many successful companies have established a center of excellence to observe Watson, and offer it more training when it needs it,” said Beth. Though the tool is smart enough to notify its human handlers when it needs to learn more about an idea, interaction, or outcome, ongoing human observation is key for making sure AI technologies run as effectively as possible.

What’s next?

“We’re transitioning from programmable, rules-based computing to a cognitive era,” said Beth. What does that mean for humans? For one thing, technologies like neural nets don’t require programmers to anticipate every possible condition of an interaction. The upshot: businesses get more time to focus on growth and refinement – while patients and customers get better services, quicker.


Understanding Audiences


Facebook’s approach to sharing data is changing, says Kunal Merchant, Audience Insights Partnerships Lead at Facebook. Facebook is now developing new ways to provide user data, which was music to the ears of the few hundred social data experts in the room. We heard Kunal discuss the role of data at Facebook internally – “Data is Facebook’s religion… Nothing is published, developed, or acquired without data to back it up” – and externally – “Users are the most important thing by far. If we’re going to present our data to the public, user safety is the most important thing and won’t be compromised by anything or anyone or for any dollar amount. If it’s going to compromise user experience, Facebook won’t do it”.

Too much data, though, can be paralyzing and overwhelming, especially for start-ups and small firms that must sift through it manually. With increasing access to data, said Mr. Merchant, must come better tools to interpret  it. Facebook isn’t trying to simply provide or distribute the data – the “firehose”, he called it – they’re trying simplify data-driven decision-making.

Everyone wants the firehose, but what do you do with that?

Facebook, as we know, has changed the way advertisers reach their audiences, but not the content audiences are served. Throughout the next decade, Merchant said, content will catch up to targeting: “Big steps have been made in reaching the right people and the right times, but content needs to be better.” Of course, as audiences get smaller due to improved targeting, just as much new content must be created to remain relevant to these increasingly specific audiences. The Creative Platforms department at Facebook, Kunal said, is developing tools to make creating relevant content scalable and easier for ad agencies. This nascent industry, we heard, is only beginning its life-cycle.

If we’re treating social data as a real-time focus group, we’re being unfair to ourselves. Real focus groups take months and hundreds of thousands of dollars, and we’re fine with that. Anything with social has to be immediate. If anything, we need to spend more time examining what we have.