customer experience applications for ai

How Well Do You Understand AI Applications?

In the last week I spoke about AI at the Argyle Forum webinar and at the ConnectID Conference in Washington, DC. Technology is emerging and we need to find a way to integrate it. In order to get approval for AI initiatives from our CFOs and/or boards and to get AI adopted by our customers, it must meet a specific customer experience need or fill in a journey gap.

As Jessica Groopman summarizes, AI can be broken into 3 major streams: big data, vision and language. These streams come with various applications that have clear business cases that we can learn from.

AI Big Data Application

Ancestry uses big data to explore DNA and compare it to hundreds of thousands of records it already has in its database to provide customers with insight about your genetics origin. Even my genetics specialist suggested last week that my husband and I check our ancestry there. Information is power, and knowing more about who we are and where we come from is a big gap in our journeys as individuals. Ancestry utilizes big data and creates a product and a customer experience that makes sense to the end user.  This is why their product is being adopted by customers.

AI Language Application

At Waverly Labs  Andrew Ochoa and his team are using machine translation to break down language barriers around the world. With an ear piece and an app, users can travel anywhere in the world and hold a real-time conversation with locals without learning the language of the country they are visiting. That is real game changer in communications. Imagine finally having a real conversation with your waiter in a small French bed and breakfast and ordering the best local dish. Or speaking to your mother or father-in-law without having to learn the mother tongue of your spouse (although understanding each other better could deteriorate that in-law relationship :)). These real customer experience gaps drove 22K people to fund Waverly Labs on IndieGogo and helped Andrew Ochoa raise $4.5M.

AI Vision Application

Facial recognition is one of the most common applications of vision AI. Yesterday at the ConnectID Conference we saw a variety of trials that prove the use of facial recognition at the airport to enhance efficiency and security for the customer and his/her journey. Airlines and airports all over the world are working on refining the value proposition of these deployments with the seamless biometrically enabled journey as the end goal.

Last year we wrote about eBay and their big bet to make AI the center of their product design. E-Bay is building the ultimate personalization, making it possible to design an outfit that does not exist!  We are all learning, but one thing is certain, chat bots are not the only application in AI and if you are a competitive customer experience professional, you better learn all the answers AI has to offer before you find yourself with a stupid chat bot in your contact center while  your competitor used AI to reimagine the customer experience, embracing a technology that drives them straight into the future.

Subscribe to DoingCXRight and get our exclusive Tips to Turn from CX Novice to CX Expert.

How to Prepare for AI: Dispatches from CR Summit, Charleston

On the eve of the #CRSummit in Charleston, customer experience leaders from various industries held the first AI Committee meeting. AI is a challenging topic to cover since it has varied customer experience applications depending on a brand’s growth cycle, customer base and business challenges. Companies like eBay place big bets on AI, while others use natural language processing (AI) only to build smart chatbots.

Regardless of their approaches all companies have one thing in common – they all need to prepare for AI implementation by having a comprehensive data strategy with flexible architecture and a lot of storage. This is the missing piece for most companies. Organizations have different reasons for lacking intelligent data. Some brands are too young and have homegrown systems that need major overhauls to even scale for the growth of the companies. Others have more robust data repositories, but have been built without the customer as the common unit.

There is a third scenario: companies that have third party CRM systems that also host the data. This makes it almost impossible to have the end-to-end data to use for building personalized experiences. It is important to learn the necessary foundations so when you meet with sales reps you can recognize the option that will fit your technology needs.

Another foundational and somewhat counter-intuitive aspect of applying AI is the need for humans. The biggest misconception about AI is that it will “remove jobs”. Meanwhile, customer experience leaders are all struggling to persuade CFOs to fund new teams of data scientists, people who would tag existing data, or people who watch for the “triggers” to use the data. Once this is done, brands will need data councils to add new elements or to design new uses of AI. Companies will always need more people to manage AI effectively.

Lastly, we all want to build solutions that will save operating costs today and enable a future customer experience transformation. So when we build, we need to think about scaling and building further, with the customer at the center.

There are still questions we need to answer. How do we begin the process (especially without much funding)? How can we make AI a reality and not just talk about it? What are the tradeoffs (if any) that we will have to make in the process?

Stay tuned for the next post on AI.

Why I Don’t Love Chat Bots

Today, we tackle the value proposition that chat bots are more valuable to companies than customers. I reject this.  The ROI simply is not there, especially since better customer experience is not there. I have experienced both fully automated bots and “augmented” service agents interactions using the chat channel, and neither delivers on the promises of chat bots as the game-changing resource we all need.

For our blog we use a photography subscription service. They push chat support heavily. When I used it, it was slow. The person either did not know English well enough or was multi-tasking several chats, but it felt like he was not present. On occasion, it felt like he was not answering the question I was asking, but rather providing a generic response. The base for effective communication is connection. When I felt unheard by the “support,” my frustration almost led me to drop the service (there are plenty of options for photography sources). Suddenly the chat offering threatens to cause the loss of a customer… and the organization PAID for it, for its integration, monthly support fees etc. I do not like that. It makes no sense.

empathetic response ai photos

AI presentation photo by Liliana Petrova, CCXP

Let’s talk about the fully automated chat bots. Allegedly, this is where companies see the real efficiencies. Again, no real value for customers unless the automated chat function allows them to fully self-serve. The problem is, the fully automated chat bot today is stupid. It can perform only very distinct functions.

The customer faced with a chat immediately tenses up with anxiety. He/she knows that the chances are pretty high he/she will have to channel switch pretty soon. How many bot interactions have you had that allowed you to solve your problem without switching to a human? Chat bot is supposed to deliver customer-centric experience. It is supposed to be customer directed, but the customer feels anxious that the minute the bot gets “stupid,” the customer loses control of the experience. Suddenly, the power is with the bot.

What is the future for chat bots? Disturbing is a word that comes to mind. Soul Machines is the company to check for the preview. They have developed a digital human that has a BRAIN built on a virtual nervous system. Now that is equally exciting and terrifying. It reminds me of the AI empowered female voice that sounds like a human one (the digital human is also female…just saying). It will take some time before the digital human agent becomes the norm, but one thing is for sure, the next version of “stupid bot” will have empathy.

According to Forrester, 60% of us prefer not to use a chat bot at all. I bet you if you ask why, you will also hear that the bots just stupid. So how can chat bots become smarter? We are back to the data conversation. Chat bots are as smart as the power and scope of the underlying data they can pull/learn from. In the case of the JetBlue customer experience team, without an integration to our reservation system, the chat bot cannot change or re-book a passenger’s ticket. Until then. customers will keep reading “please speak to an agent.” That is hardly a way to feel the love.

AI

How Smart Do Humans Want AI To Get?

Last month we covered the basics of AI and what it is in theory. Today we will talk more about the existing practical applications of AI and reflect on what that means to us. We will also share some caution about AI, which, similar to IoT does not have existing laws to follow due to the speed of its development. It is a fact that there is no real regulation on IoT products today and there are cases filed against familiar brands like Bose.  Last month I bought my team Google home minis to inspire a culture of innovation. Sometimes I wonder… what exactly did I buy them?
 
Currently, one of the most impressive real expressions of AI is the voice-generating AI that is indistinguishable from humans. Now that is a real capability that can generate significant efficiencies in call centers (given that you have the clean data and the cloud enabled applications to power the “brain” of the AI).
 
The AI voice re-imagines the how of the interactive component of the customer experience. The innovation part is that, to the customer, there is no real change in the emotions they experience during the call. The customer still hears a friendly female voice. This AI application does to the call center what the touch screen did for cell phones in 2007.
 
Keep in mind that as you converse with the friendly female voice, the AI that powers it is learning – it is getting smarter and smarter. The question is how smart can AI get. A better question, how smart do humans want AI to get? We all have heard the story of Frankenstein. If we are not careful, we might build our own replacement.
 
If you think I am exaggerating, listen to Elon Musk who has been cautioning us against AI for years. “I think we should be very careful about artificial intelligence. If I had to guess at what our biggest existential threat is, it’s probably that… I’m increasingly inclined to think that there should be some regulatory oversight, maybe at the national and international level, just to make sure that we don’t do something very foolish.” We should most certainly embrace the future. However, we need to make sure we keep the control. With regulation and legal framework speed we might build the guardrails too late. Let’s also not forget that the master computer voice in Terminator, like those used in call centers, was female. I wonder if that is just a coincidence….
 
AI is not capable only of generating voice. It is able to “read” and to analyze human voices in a manner we have never seen. The term for this is “voice profiling.” Apparently, humans cannot detect the data/information in our voices the way AI software can. So the next time your bank asks you if you consent to record your voice for “additional security” on your account, think twice (too late for me… the TD Bank AI already has filed me away). The advancement went as far as creating a face to the voice in December. “Your voice is like your DNA  or your fingerprint.” according to Rita Singh. Thanks to today’s advanced algorithms and the large computing power available, our DNA is out in the digital space, ready to be “profiled.”
 
The reach of AI goes beyond customer service and personal banking. AI is now making hiring decisions today. Companies use AI technologies to shift through the large volume of online resumes they receive. Ostensibly, AI creates a leveled ground for all candidates and alleviates existing biases during interviews. An industry has emerged predicting either candidates’ skills based on online clues in the case of Fama or on their actions through Entelo that “guestimates” who is likely to leave a job. The situation gets creepier when we start talking about companies like HireVue. On the surface, HireVue’s solutions sound innovative  – “databased, unbiased talent decisions.” In reality, HireVue has the potential to read our faces when we speak and determine whether we are not truthful. No more saying at an interview you are leaving your current employer to pursue further development. HireVue will know you did not get along with your boss, or that you hate banking. A world in which we cannot have our internal lives sounds like a scary and lonely place.
 
What is next for the brave new world? According to the cofounder and CEO of Nvidia Huang it is “… the ability for artificial intelligence to write artificial intelligence by itself.” In other words humans will no longer be needed or even capable of running the new technology since it will be beyond our capacity to comprehend or manage. I don’t want to go back to Terminator and the Transformers, but it is hard not to see the resemblance in the plots. On the bright side, one can make a fortune in the next decade investing in those underlying technologies that will be running themselves or us in the future. This is how Huang made his fortune. We can always follow in his footsteps.
eBay

CX Bold Moves: eBay’s Vibrant Marketplace of the Future

Two years ago RJ Pittman and eBay made a big bet and made AI (artificial intelligence) a core discipline of the company, similar to brand marketing and sales. The organization was able to focus and build upon existing technologies to create a truly personalized experience on eBay. This is the definition of a big bet, and the reason why eBay is part of our CX Bold Moves series.

In the retail world of RJ Pittman, the conversational commerce does more than assist us in buying an outfit, it uses predictive analytics and AI to design the outfit if it does not exist anywhere in the world.  Now that is the ultimate personal relationship any brand can have with a customer. Once eBay is able to consistently deliver this value proposition to the customer and to scale it for the 3Bn online users today, Amazon will finally have a match – and become less ubiquitous.

How did eBay get here?

“Don’t start from the tech, start from the experience and start with transformative experiences that your customers will feel…” is how RJ Pittman summed up his approach at the Forrester conference in San Francisco this fall.  eBay is an excellent case study of a brand that followed the principles for self-service that we laid out last week – the brand will enable the customer of the future to post and sell any item without any effort or friction. Through computer vision, deep science looking at images, real time pricing, conversational commerce with interactions, and a world price guide, eBay is building an AI-enabled ecosystem that will have the power to create many real personalized experiences for all of us.

What about the eBay brand?

“The brand is the product is the brand” are the words of RJ Pittman that sum up the future of marketing. Today marketing is less about ad campaigns and more about a brand’s products and the customer experience that accompanies them. This year JetBlue achieved $33M of ad spend equivalency with its launch of facial recognition technology at the gate. This is much more brand exposure than an ad campaign. Without the existing social media and mobile technologies, that would have been impossible. But in the world in which we live, the brand is more about keeping and living the promises you have made as an organization and less about what those promises are.

This is an important note of caution for all those who have brand marketing and customer experience under two different executives. How are you ensuring that there is a real alignment between those two legs of your customer relationship?

What did eBay really do?

eBay applied technology to build a dream no one has imagined yet. The innovative brand did not merely optimize its responsive website (some of us actually even call responsive design digital revolution:( ), it invested heavily in a conversational commerce that will define how customers purchase in the future.

This disruptive strategy is not a science project. It is the birth/built of a vibrant marketplace of the future. Rik Reppe said on stage in San Francisco that courage, flexibility and imagination will make or break our efforts with AI. Listening to RJ Pittman it felt like he has all three of them. Do you?

Is AI Really The Answer?

Earlier this week we shared some of the pitfalls of implementing  self-service and highlighted the importance of strategic and empathetic implementation.  AI (artificial intelligence) is one of the self-service tools in the customer service professional toolbox today. It is also one of the new buzzwords, together with blockchain (for the curious ones  – my favorite explanations of blockchain are this video and this article).

The primary current positioning of AI is in call centers. The value proposition is that with AI, companies will empower customer service employees to make better decisions/recommendations, thus increasing employee engagement. Additionally, through AI, organizations will achieve significant ROI by automating the role of the customer agent (in JetBlue’s case, the crewmember) and scaling customer support without incremental headcount.

So what exactly is AI? Do you really understand what this technology can and cannot do? If you do not, keep reading as all working professionals today should understand at least the basics of AI. If you are like me, you probably get 100 sales emails every week telling you that you are running late and must leverage AI in your call centers. But do you really need AI? What problem do YOU need to solve? And is AI the best way to do that for your company?

Last month I was invited by Execs In the Know to join their AI Advisory Committee with the below mandate:

“The final group output will come in the form of a report. The exact nature of the report will be determined and framed by the group, but may include areas such as:

  • A summary of ways AI is enhancing CX channels
  • Best practices on where and how to start
  • Trends and technology in AI to improve service
  • ROI of current AI customer service initiatives
  • Perspectives and predictions about the future of AI and customer service”

This is no small mandate and it is encouraging that we have a group of professionals who are examining these questions before we all get ahead of ourselves with AI and compromise the ROI we all want so much.

The most common use of AI is ML (machine learning).  Basically, this is data mining and predictive learning on steroids that enables a computer to make decisions and interact with a human. With that basic understanding I already have a few questions to all the companies that are calling, emailing, inmailing etc., to offer me AI enabled solutions. Who, or rather what, is enabling those solutions? Is it my company’s data? Because if it is, I have a lot more work to do internally before I respond to those sales pieces.

Erik Brynjolfsson and  Andrew Mcafee provide a comprehensive explanation of what AI is and what it is not in their publication The Business of Artificial Intelligence. In it they state that AI technology is ready for implementation in the business world and that “[t]he bottleneck now is in management, implementation, and business imagination.” I do have the business imagination, but I also am taking my time to know exactly what capability I am buying with AI. It might be cheaper to streamline processes and fix existing software tools or integrations to enable my employees to deliver excellent customer service, instead of paying for yet another software integration that makes decisions based on my bad data (since I would have prioritized the purchase of the expensive AI solution over cleaning my existing data).

Although it is clear that AI will be a solution that needs demonstrated ROI and employee adoption for success while it is learning, it is also clear that we cannot wait too long to get comfortable with this new technology. As Erik and Andrew say, one thing is pretty sure: “[o]ver the next decade, AI won’t replace managers, but managers who use AI will replace those who don’t”.

That is the exact reason I chose to join the AI Advisory Committee. More to come!