Common Misconceptions Executives Have About AI
Artificial intelligence is no longer the sole domain of tech companies like Google, Facebook, IBM or Amazon. Recognizing the potential of exponential technologies like A.I. and bots, creative agencies like Ogilvy and consulting firms like McKinsey and Accenture now proudly feature A.I. departments.
The message to executives is clear: it’s time to understand and leverage trends in automation and artificial intelligence.
According to McKinsey’s Michael Chiu, “research suggests that as many as 45 percent of the activities individuals are paid to perform can be automated by adapting currently demonstrated technologies. In the United States, these activities represent about $2 trillion in annual wages.”
Andrew Ng, Chief Scientist at Baidu and Stanford professor of machine learning, confirms that “if a typical person can do a mental task with less than one second of thought, we can probably automate it using AI either now or in the near future.”
Breakthroughs in deep learning have driven major advances in machine perception. Computers can now reliably detect and classify objects in images and video, transcribe and translate speech as well as humans, and even generate art, music, and movie soundtracks.
A.I. companies like Clarifai, Ditto, and GumGum are leveraging these new technologies to help brands understand content, identify brand mentions, and calculate earned media spend from sponsorships. Other companies like Affinio, Motiva, and Reflektion improve marketing intelligence, automatically optimize campaigns, and streamline customers’ retail experiences.
Last year, Salesforce acquired an A.I. startup called MetaMind to integrate into Salesforce Einstein, which promises to automate many sales, CRM, and ERP processes for your business.
The use cases for A.I. are here and signs of positive ROI are appearing, but before you get swept away in the hype, let’s talk about the 5 most common misconceptions about A.I.
1. A.I. does not have to be elusive
The type of A.I. that investors care about is not necessarily the same that enterprises care about. Investors look for 100x returns on capital, so they heavily scrutinize a founding team’s technical pedigree, industry expertise, margin defensibility, and broad market potential.
On the other hand, “executives don’t need (or want) a lesson in computer science, they want to know how this technology can be used as a tool to help them achieve their business goals,” points out Ophir Tanz, CEO of GumGum. GumGum is an applied vision company which has partnered with Fortune 100 brands to curate visual content and optimize brand marketing since 2008.
Tanz highlights just a few of the many applications of computer vision for brands: “retailers can leverage visual search and increase revenue through shoppable imagery, sports teams and rights holders can deliver more accurate valuations of broadcast and social exposure, social media can be scoured and activated like never before.”
Normally when brands sponsor major sporting or social events, they can’t easily calculate the earned media lift or the ROI on their investment. GumGum’s vision technologies can identify when brand logos have appeared in social media images or sports videos, making such calculations possible.
While the technical expertise required to successfully strategize for and implement A.I. technologies for major brands may seem daunting and out of reach, partners like GumGum combine industry expertise, full-service agency services, and A.I. expertise to help brands ramp up.
Zachary Jean Paradis, Vice President of Customer Experience at SapientNitro, has helped many brands in industries ranging from financial services to CPG get started with their A.I. strategy. He emphasizes to his clients that A.I. is “not a single thing, but rather a series of methods and technologies that allow you to mimic human intelligence.” The key question to answer is “what intelligence am I trying to mimic?”
His recommendation to brands is to start with offerings from foundational mainstays like Google, Microsoft, and IBM, and Salesforce and then layer in specific vendors, such as Cycorp and Luminosofor natural language understanding or Clarifai and Sentient for computer vision. In many cases, brands can minimize the amount of bespoke code they need to write.
2. A.I. is not a silver bullet or cure-all
As transformative as A.I. is for many industries, the technology is not magic. One mistake non-technical brand executives make is “assume that artificial intelligence is some kind of silver bullet,” according to Ryan Detert, CEO of Influential. “Some executives think A.I. is a sentient being, like in Terminator. They ask if it can think and tell them what to do. We have to explain that A.I. is simply a better way to turn data into actionable insights.” Such misconceptions are not necessarily the fault of the executives.
Many startups capitalize on the knowledge gap in A.I. to hype up marketing fluff such as “executive brains” that can “predict the future” and “automatically increase revenue.” No wonder executives are confused as to what A.I. can and cannot do.
Detert and his team help brands improve the performance of sponsored social media posts by matching campaign and brand content to the right influencers based on personality, context, and timing. Normally, sponsored posts by influencers suffer from a 20–30% drop in engagement rates. By leveraging A.I. powered by IBM Watson, Influential is able to drive gains of 20–30% across ad recall, positive sentiment, social engagement, click through rates, and ROI. The key industries who benefit from this targeted influencer marketing are consumer product goods (CPG) and entertainment businesses.
We’ve previously shown you how easily developers can leverage Watson’s cognitive services, but non-technical teams without the requisite software development, data science, or machine learning capabilities are often mystified by the process. “Due to lag of time and the ever-moving landscape, most corporate companies move slowly,” Detert notes.
In the cost-benefit analysis of whether to build or buy, enterprises typically move faster with an experienced vendor or consultant. After all, if McKinsey has to write guides to teach executives the basics of software development, perhaps fast-moving technology projects are best managed by experts.
3. A.I. does not have to be expensive or onerous
Building competitive A.I from the ground up requires expensive specialized talent and volumes of proprietary structured data. Luckily for most brands, this is not yet necessary.
Many brands like Disney, Uniqlo, and the New York Times, have successfully experimented with chatbots on Facebook, Slack, or Kik while others have dipped their toe into voice-based technologies by releasing Alexa Skills for the Amazon Echo.
Paradis of SapientNitro points out that plenty of chatbot enablers exist, ranging from IBM Watson, Nuance, Microsoft Bot Framework, Google’s API.ai, and Facebook’s Wit.ai.
While the tech industry does not perceive brand usage of mainstay vendors as “real AI,” such experiments nevertheless solve important business problems and are essential for executives to stay educated and competitive in digital.
In the liquor and spirits industry, companies are legally required to put up an age gate to protect minors from digital content. No matter how much you optimize the birth date input form, manual input by users leads to massive drop-off rates, particularly on mobile.
That’s why Anheuser-Busch, the world’s largest beer producer and the company behind Budweiser, implemented a voice input option alongside the regular form. In testing, these new designs improved the consumer experience and conversion rates. In development, the technology needed to proactively handle mis-spoken or mis-understood words while also equating “December 23rd 1994” to “23rd of December 1994” and many other permutations.
According to Lucas Herscovici, VP President of Consumer Connections, Anheuser-Busch was able to go from idea to implementation in less than 3 months. The bulk of the time was spent validating with legal and compliance and negotiating with vendors, while the actual technical integration took less than 3 weeks.
By corporate standards, that’s a fast turnaround time.
4. A.I. will not replace executive decision making
“When we first introduced websites, they were a train-wreck. They didn’t work well technically or design-wise,” Paradis reminds us. Similarly, “A.I. is in a young adult, awkward teenage phase. Some technologies will do very well out of the box, some will be a challenge.”
From C-suite executives to front-line managers, business leaders will need to identify where A.I. can and cannot make an organizational impact and continuously “prototype, prototype, prototype” in order to “raise their A.I. IQ,” in Paradis’ words.
McKinsey estimates that the benefits of implementing the right automation and A.I. technologies can be 3–10x the cost. Thus, “the ability to staff, manage, and lead increasingly automated organizations will become an important competitive differentiator.”
5. A.I. will not cost-cut you into leadership
As with any technology wave, there are leaders and laggards in A.I. Paradis of SapientNitro observes that “the single industry that’s leading end-to-end is financial services.” Companies he’s worked with in the space are leveraging A.I. and bot technologies for customer engagement, process automation, fraud and risk mitigation, business analysis and improved executive decision making.
Not all industries have seen the same benefit. “In telecom, we’ve seen the introduction of chatbots, but they didn’t perform to the level that was expected.” Similarly, Paradis sees challenges for consumer packaged goods (CPG) and quick-service restaurants (QSR).
CPG brands are challenged not just by A.I. but by competing in an era that is driven by customer experience when they don’t own many of the customer touch points. Many of them own the marketing experience, but not the retail experience, which means they are missing critical purchasing behavior about their products. At the same time, many retailers are introducing private label products that are directly competitive.
CPG products are also generally too simple to turn into connected products. Adding sensors or bluetooth connectivity would make their products prohibitively expensive without adding much utility for consumers. While companies like Google and Facebook generate petabytes of data from their consumers’ every move, CPG companies are in the dark with consumer usage data.
A.I. can still be used by CPG companies to improve marketing insights and strengthen product innovation pipelines, but the data gaps across the industry “make application of data and A.I. very limited,” says Paradis.
The same applies to QSR who have historically been slow to collect the requisite data to power A.I. initiatives. Many QSR such as McDonald’s run on cash-driven businesses, making consumer purchasing trends harder to track.
Self-service kiosks can close inefficiencies and reduce costs, but “A.I. will not cost-cut you into leadership,” warns Paradis. Instead, companies need to radically re-think how A.I. can free up humans to deliver on the highest value activities.
“The majority of sales of McDonald’s in Europe go through a kiosk. There’s only a single pay register.” Paradis goes on to illustrate how executives can think in terms of what A.I. can enable rather than what A.I. can cut: “But there are hosts and hostesses to welcome you at the door, walk you to the kiosks, explain your menu options, and bus your table. Instead of eliminating humans, McDonald’s can deliver a more amazing experience.”
Learned something? Hold down the the 👏 to say “thanks!” and help others find this article.
This post was originally published at www.topbots.com on February 13, 2017. Love what you read? Join the TOPBOTS community to get the best bot news & exclusive industry content.
