34
portfolio_page-template-default,single,single-portfolio_page,postid-34,stockholm-core-2.3.2,qode-social-login-1.1.3,qode-restaurant-1.1.1,select-theme-ver-9.0,ajax_fade,page_not_loaded,popup-menu-slide-from-left,,qode_menu_,wpb-js-composer js-comp-ver-6.8.0,vc_responsive
Title Image

Robotics

Robotics

 

 

 

 

 

 

 

About the above: An excellent example of the uses and conundrums of internet robots, or bots.  This guy codes an Instagram bot in front of you. And if it gets past the Facebook digital guards, you can quickly pump your IG numbers! 

 

 

Definition and background

 

The evolution of the field of robotics over the past 20 years has been breathtaking. We’ve gone from IBM’s Deep Blue, a closet-sized computer that beat world chess champion Garry Kasparov in 1997, to today where literal microscopic bits of code and electricity sweep across the internet in milliseconds; measuring, monitoring and executing programmed functions. 

 

For communicators, these tiny bits of code, or “bots” are the only robots that matter.

 

Dunham and Merrick describe internet bots as parts of “a software application that runs automated scripts (tasks) over the Internet.” These tasks can be simple or complicated, benign or malicious. Without Google bots, there would be no SEO and a way to determine which sites are more popular. Without Facebook and Twitter bots, some would say there’d be no Trump. Without bots we cannot easily identify online threats, but shopping bots clog email boxes with junk. Bots buy up the good seats in milliseconds at Ticketmaster, and they helped take down the Iranian nuclear weapons program back in 2010.

 

Bots, at this point in their evolution are Jekyll and Hyde, depending on the user. One reason why people rightly fear AI is because many of these early adopters have been predatory.

 

Before we get into the types of bots, let’s look deeper at their usage. Whether you choose to believe it or not, bot traffic surpassed that of humans in 2016. I could write 1000 words to explain it, or I could just show you this 2016 infographic on bots and traffic from web analytics firm Incapsula.com:

 

If bots are something everyone does to each other, what if you don’t want to play the game? If you want to understand your traffic, you should identify the bots that interact with your site. You should establish what human activity looks like on your site and channels as a way to identify bot activity now.

 

SEO bots have a pattern. So do bots that you allow on your site due to contractual agreements. Humans, even the most shopping-addled, don’t hit the same product URL 50 or more times over a day. Bots do. Borrowing from best practices established by the people who fight digital fraud, human activity comes from a specific number of devices at fairly reliable times and places. Bots don’t share those patterns.

 

Chatbots, the other “bot”

 

Chatbots went through a fad phase in 2018. There was a big push to recruit new waves of casual and small business developers to deploy channel-specific chatbots on their digital properties. They offered easy integration to Facebook Messenger and embed code for websites. Some of these platforms focused on customer service. Others focused on SMB services like direct communication, appointment setting and prequalifying customers.

 

Overall, the category projected this growth due to Facebook putting a Messenger chatbot API on the market. To entrepreneurs and social media marketing companies, this meant business success by recruiting as many new users as possible to a dashboard-style product.

 

The year started strong but sputtered a bit after Messenger put a hold on third-party products. Naturally, the momentum stalled. About that time, the conversation and enthusiasm turned fully toward computer audition. Why type when you can speak to your phone—or your Alexa? By the end of the year, the mini-wave of chatbot enthusiasm waned, while the number of Alexa, Siri and Google Home users skyrocketed.

 

None of the above means chatbots are dead; rather it’s another lesson about fads. The fad subjected the non-enterprise side to the brutal survival test of being the next “it” product. Some technologies and companies survive, others don’t.

 

In the case of chatbots, the enterprise side is still working away, providing good use cases on how to properly use the technology. Like customer service.

 

Why are chatbots good for customer service? First, there is a willingness to engage. Depending on the complexity of the problem, it has been shown that if a customer can solve a problem without talking to another human, they will. Companies are looking for better ways to prepare CSRs with information and options for customers. All three are interested in understanding when to switch a customer to human intervention from chat.

 

Was 2018 supposed to be the start of a WordPress-style wave of adoption among freelancers, bloggers and small businesspeople? The fad was driven by press around and from developers of  relatively cheap Messenger bot platforms like Botsify and Chatfuel, just two of a handful. These platforms offered people a chatbot without much thought or effort; just customize the template and go. The problem was the results matched the offer, not much was well thought or properly supported. Unlike the enterprise side, conversation flows were only as good as the person developing them. If the designer didn’t think through the range of inputs and replies, the bot broke, making it useless. The idea of plugins for maps, email, etc. added features and flexibility, but their reliability was hit-or-miss.

 

While that was happening on the social side, the enterprise side kept developing significantly better chatbots, improving the AI, and exploiting new ways to connect to customers.

If you’ve used Expensify Concierge to track your travel and trip expenses, you were interacting with a chatbot. If you’ve used “gifts when you need,” or GWYN while at 1-800-Flowers.com, you used a chatbot to help select an item. If a recent college grad says they’re talking to Mya about a job, they are talking to a chatbot.

 

Each one of these products represents the next iteration of NLP-powered chatbots. Pypestream, a New York City-based AI startup calls it “Conversational AI.” It represents what happens when companies begin to use NLP to mine conversations for cues in real-time. As you submit expenses to the Expensify app, the Concierge feature uses NLP and machine learning to understand your patterns. Eventually, the system will automatically categorize expenses without your assistance, flagging you if there is a question.

 

Mya is the most interesting use case. It is an NLP-powered AI assistant for recruiters and HR departments that revolutionizes their role. The technology serves as a helper to both candidate and company through the employment process. On the candidate side, Mya guides people through the application and serves as their single contact. Candidates can interact with Mya for status updates. Mya can advise prospective applicants of other positions of interest. Candidates screen through Mya, and the information is helps set up a shortlist. Mya serves as a central scheduler, working with both sides to find the right time for an interview.

 

In the meantime, Mya reads every job description, resume on file and piece of related information. Mya mines this existing database and suggests optimal candidates for upcoming needs and handles outreach to them if asked. Mya also periodically reaches out to new and existing employees asking for feedback about the process and company.

 

Mya uses “natural language understanding” or NLP that is tuned for semantics, entity recognition and multiple interest classification. Which means it understands context, interjections, evolving answers and can keep conversations on point. While acting and sounding like a human. Most of all, its ML algorithm keeps it optimized.

 

 

Robotics through Porter’s Lens

 

Bots are simple applications that perform simple tasks. In fact, they are the delivery device for most everything important to communicators. The best way to think of them are as enablers when used with other AI technology.

 

Cost leadership—is there a way to use bots to flatten costs enough to become the low-cost leader in a sector? An army of bots, used properly, can be focused to learn, understand, monitor and react. Is there an angle within those broad purposes that one can build economies of scale around? If you buy a lot of media, programmatic is one way to accomplish this. Bots could spread across the web to monitor ad costs and placements. The information can be used to influence future bids. But is there enough savings to rightfully claim cost leadership?

 

Can a company use chatbots to accomplish the same thing? Is there potential harm in using an enterprise tool for 90% or more of customer service functions? Are those savings enough to provide a significant enough savings over the competition? At first glance, bots and chatbots may not be enablers of cost leadership.

 

Differentiation—can a company use bots or chatbots as a way to become unique in the minds of consumers? Chatbots-R-Us? We-be-Bots?

 

Focus—using technology to stand apart from the competition on the needs of a specific customer or customer type is more of a natural fit for bots and chatbots. Mya would be an example of using NLP-powered chatbots to focus on the needs of candidates and HR professionals.

 

 

Couple things

 

A chatbot needs to focus on something specific, lead identification, appointment setting, etc. Inevitably, someone will see a chatbot as a Christmas tree, something for everyone. They will be wrong.

 

Chatbots could be a reason to reach out to prospective customers. Approaching them with a better product, thoughtful strategy and an offer to centrally manage the process should be of interest. It allows you to discuss more about their business in new areas; HR, business processes, customer service as well as PR & marketing.

 

 

 

Date

March 4, 2016

Category

Robotics