NEW YORK, NY--(Marketwired - April 19, 2017) - Artificial intelligence (AI) is the new buzzword being applied to seemingly every business challenge you can name. AI is an area of computer science that uses "intelligent agents" to develop capabilities in robots and computers to perceive and react like humans, and therefore to augment and/or replace humans, especially in doing tasks that are predictably rote, frequent and high-volume. Popular applications of AI can be found in language translation, self-driving cars, disease diagnosis, fraud detection, and delivering solutions to clients via chatbots.

It is important, however, to note the limitations of AI. Unlike humans, AI has no consciousness, cannot feel genuine emotions, and is incapable of empathy, nor can it generate or deal well with novelty and other uniquely human attributes.

"Every time I hear 'today it's all about artificial intelligence,' I know that I am about to be sold another allegedly disruptive and very expensive tech project that will overpromise and under-deliver," the CEO of a multi-billion-dollar luxury brand lamented at a recent meeting in New York with Luxury Institute CEO Milton Pedraza. "My team's challenge is that we are not expert enough about AI to know how to best use it, and how to maximize our return on investment. Everyone who sells algorithms claims to be AI-driven. We don't know what or whom we can trust to serve our clients' interests."

That sentiment sums up the current dilemma confronting luxury goods and services leaders who are trying to determine the most effective AI projects to undertake, and to choose partners that they can trust to complete projects safely and profitably. Unfortunately, results rarely live up to the claims made by consultants. Remember Y2K and CRM? This is bigger, a lot bigger; and so are the claims. If you believe the hype from AI vendors, your sales -- and the GDP of the United States -- should double by year-end, even though most of your employees will be replaced by chatbots.

You may need a fraud detection AI application to help you navigate this terrain. AI leaders claim that they have "super powers," that AI is the new electricity, and that there is no industry that AI will not disrupt. The humility is overwhelming, but they may be right, at least about the last two points. Still, legendary Silicon Valley investor Peter Thiel, who should know better than anyone, says that he doesn't just walk, he runs when he hears buzzwords and lofty promises bandied about by tech start-up founders and salespeople because it's usually an indication of undifferentiated technology, destined to be commoditized, that delivers zero competitive advantage. AI is the latest and greatest buzzword. It is also over-hyped and over-rated, perhaps more than CRM, Big Data, analytics, and the Kardashians.

Regardless of your opinion of Peter Thiel, his insights and track record should give you pause, and he is not alone in his views among technology investors. In a February 2017 video interview with Fei-Fei Li, head of AI at Google, Michael Abbott, general partner at Silicon Valley's premier venture capital firm, Kleiner Perkins Caufield & Byers, estimated that he meets with 15 new start-ups a week, and 10 of them use the term AI in their pitch. When he probes entrepreneurs on their AI claims, Abbott finds that most of them do not understand what they are saying, and often what they are doing does not truly constitute true artificial intelligence or machine learning.

If top venture capitalists are skeptical of those who label themselves AI experts, you should be too.

Bringing Business Expertise To Artificial Intelligence

Luxury Institute CEO Milton Pedraza is no stranger to the core drivers of AI, which include clean, structured data, processing technology, genetic algorithms, and, most importantly, domain expertise. Not by coincidence, those are the same core elements on a smaller scale, that drive customer relationship management. In 1997, Pedraza, for years a huge fan of Complexity Theory and the Santa Fe Institute, began to successfully execute CRM at Citigroup. Based on his results, he was selected to lead Citi's global CRM project. With massive amounts of client data, he hired a Boston-based data-mining firm to build predictive models using genetic algorithms. Pedraza also ran the global CRM project for $20 billion+ Cendant briefly before becoming an entrepreneur in 2003.

In 2004, knowing well the power of genetic algorithms, Pedraza hired IcoSystem, a modeling firm that uses neural networks and genetic algorithms, to assess the viability of his own luxury automobile shared access start-up. Pedraza came away with the understanding that most technology is a commodity and that the only true ways to differentiate in luxury are your unique, relevant products and empowered brand ambassadors with highly developed expertise and mastery in emotional intelligence.

Pedraza's experience illustrates that CRM and AI have more in common than many luxury and retail executives perceive. Both CRM and AI run on the same rocket and fuel, as Stanford's Andrew Ng likes to put it. Unfortunately, CRM after more than two decades has yet to deliver on its over-hyped promises, and while CRM technology is a great tool with real benefits drawn from the investment, there is no CRM system on the planet that has ever delivered sustainable competitive advantage to a brand. Most projects still under-deliver. Recently, a multi-billion-dollar luxury brand based in Europe that was reputed to be the most tech savvy discontinued its relationship with the largest CRM provider due to unsustainable costs and low measurable benefits.

While the entire luxury and retail industry and society may benefit astronomically from AI, for your brand, it will simply become table stakes. Apple, Amazon, Facebook, Google and Microsoft can accumulate massive amounts of proprietary data, and they are investing billions in cloud infrastructure to store and process that data. In AI, if you are dependent on vendors and don't have the massive proprietary data, processing power and expert staff of a Google, you will become the next trial and error test case. It makes sense that if AI is the new electricity, then by definition almost everyone can plug into the benefits.

Like electricity, AI is imperative for any brand, in any industry. In luxury, AI will be even more critical than it is in price-driven, commodity categories, where human beings will not be a part of the front-line equation like they are in luxury. AI will likely enhance most luxury back-office and front-line jobs rather than replace them completely. It remains to be seen whether AI will deliver a high ROI to luxury goods and services brands, or, whether, like e-commerce, it will simply add costs and force brands to compete with economically irrational competitors such as Amazon.

To help luxury goods and services leaders make optimal decisions, Luxury Institute recently conducted a study to identify current myths and realities of AI. Luxury Institute researchers, aided by a retail expert with a master's degree in mathematics and statistics from Columbia University, pored over documents and reviewed dozens of courses and videos, including those of Silicon Valley legend, Andrew Ng, seeking out and distilling the most objective insights.

Here are seven critical facts your AI vendor won't tell you, but you absolutely need to know before you sign on the dotted line:

1. AI needs access to massive amounts of data.

Insiders confess that the one thing that has driven AI's current vibrant state is not so much better algorithms, but the vast amounts of data that are now being generated every second by the Internet, and within proprietary company databases. Still, many individual companies suffer from scarcity of clean, structured data. If your data collection discipline is dismal, as is often the case, especially on the front-lines, then your ability to provide massive nourishment for the machine learning required by hungry algorithms is limited, no matter how clever they are. You need to conduct a data audit immediately, and assess the amount, relevance, and quality of your data. Identify which data are critical for your business to thrive. Begin to collect the critical data now. Client books, with their critical data, must be digitized in the era of AI. Clean, accurate data is the critical asset today, and if you don't have it, your business is destined to expire.

2. AI will only work if you reconfigure your organization to deliver coordinated, seamless client experiences.

To create a seamless client experience, you must reconfigure your organization to act as one customer-centered, highly adaptive organism. Despite a major crisis under way in the luxury and retail industries, many leaders fail to reorganize and optimize the client experience. One of the major reasons that CRM has failed to live up to its potential is the schizophrenic silos that still exist between operations, product management, marketing, public relations, customer service, sales, social media, and e-commerce. With AI, automating your silos individually will only generate even more disconnected client experiences that will drive customers to competitors. You must unite any part of your organization that touches the client into one expert team that works to optimize the full value for the customer in time, money, and happiness. If you don't do this first, don't invest in AI. Disconnected, optimized, stand-alone AI silo solutions that are automated will only amplify the dysfunctional nature of your client experience.

3. AI algorithms are as effective and as biased as the data scientists who create them.

AI is selling the myth that algorithms are intelligent and objective, but nothing could be further from the truth. Algorithms are designed and built by humans. These humans have major biases. The data set that is selected, labeled, structured and used to train a genetic algorithm can be biased by the experiences and mindset of the AI team. Take note that the AI industry, probably more than any other industry, is comprised of "bro-grammers," mostly single Caucasian and Asian males. Females and other minority group members are often excluded from or marginalized on their teams. AI's most common method today, machine learning, is about 90% "supervised learning." It is called "supervised" learning for a reason. The training examples and weightings upon which genetic algorithms learn iteratively are provided and tweaked by humans. They may have advanced degrees, but that does not guarantee that they have social skills, and, as research shows, they especially lack empathy. Before you set any algorithm free to engage with your luxury clients, you need to torture test it to ensure that you are treating employees and clients, especially women and minorities, who are the present and future of your brand, like cared-for, respected, and valued human beings.

4. AI is new, so new it direly lacks data scientists with expertise, especially business domain expertise.

Today, AI is in its practical infancy, and so are the data scientists who staff most of the AI companies. First, unless you have access to massive amounts of data and massive processing power (e.g. you work at a high level at Google, Amazon, Apple, or Facebook) no matter how many PhD degrees and boot camp lessons you have accumulated, you have minor experience. You do not have expertise in machine learning, and even less expertise in any industry or business domain. Even the best in data science are nothing without massive data and processing power. Most data scientists today are in the same shape as someone who has learned how to drive a Ferrari in a simulator, and who may actually own a Ferrari, but has never had the fuel or a challenging racetrack to develop expertise and mastery of driving the automobile. Before you hand your data over to a team of data scientists, conduct full due diligence, not only on how many successful projects the data scientists have executed, but their actual contributions to each project. Client references will be even more critical with AI vendors than any other business purchase you will ever make. Finally, you must provide the diligent adult supervision that brings reality and domain expertise to any AI project.

5. Genetic algorithms are "black boxes" that often cannot explain why they responded to a client in a particular way.

One of the most fascinating attributes of genetic algorithms is that their behaviors, like a human brain's, are not always easy to understand. Sometimes they can respond in ways that are negative, unethical, or illegal. The way algorithms work is that they behave a bit like high-speed procreation of a species inside a computer. They are given a training set of data, accompanied by the right answers, and they increasingly improve their ability to tweak the strength of each connection, until they get better, and eventually, reach a probabilistically, if not perfectly, correct solution. This process works like our brains, weakening and strengthening connections until they get it right and can learn to classify data, even if it is new. Unlike in a conventional database, the connections an algorithm establishes are encoded in the strength of the connection, not in a specific location. This makes it difficult for data scientists to audit and explain why an algorithm may have acted in a specific way. For example, it may have rejected a top-tier client's request, or the request of a daughter of a top client with a different last name and buying pattern. While a human being could make an exception judgment in real-time, an algorithm is restricted by its own connections. In some cases, algorithms have veered driverless cars off the road, or excluded people for loans illegally, based on connections that are not obvious to, or explainable by, the programmers. Make sure you understand the risks involved in letting algorithms do your thinking. Unlike the baffled data scientists who can walk away, you will be held fully and legally accountable, and you will be responsible for coming up with explanations and solutions.

6. Your online AI interactions will be subject to mischief from hackers and other interaction risks.

AI's big claim is that its data scientists have created algorithms that learn to learn. That is indeed a huge achievement, but it can also be a huge liability. Look at the example of Microsoft, which in 2016 put Tay, a chatbot, into the public domain to engage in conversation with people on Twitter. Within 24 hours, the chatbot was taken offline because it had been gamed by clever and malicious users who "taught" Tay to be racist and even to promote Nazi ideology. The lesson is that when you release your continuously-learning algorithm into the public domain, such as your website, and social media, you will expose it to a world where it will learn from whatever it interacts with over time, good or bad. Unless you are constantly and meticulously vetting the incoming data and refining the algorithm, the results it produces may be far from what you desired. This is a very expensive caveat. In the case of Tay, the culprits were mischievous random people, but you can bet that hackers worldwide are salivating at the opportunity to wreak havoc on the learning algorithms of companies that engage online, which means all of us. One thing you will need to figure out with AI is who is continuously watching the algorithm interactions to identify and prevent mischief that damages client experiences, and your brand, or worse, creates massive security risks.

7. AI is not a substitute for true human engagement.

AI vendors will tell you that their algorithms can think much better and faster than humans, and unscrupulous vendors will claim that they can, and should, actually replace humans. When you hear that, take a deep breath and think like an intelligent human being. For rote, mechanical, frequent, high-volume tasks, it is true that AI will replace humans. In fact, most company rules, policies and procedures today have hollowed out many jobs into dehumanizing, scripted, soulless, mechanical interactions that should be automated. Luxury Institute believes that genuine, empathic, trustworthy, and generous human experiences that deliver emotional value and joy, will never be replaced by AI. If you choose to join, or compete, directly with Amazon, then you should automate most tasks and transform your brand into a utility. In luxury, however, AI will be a human enhancer. It will augment the power of human beings and help create humanistic jobs that we cannot imagine today. Develop the refined skill within your luxury leadership of being able to differentiate the tasks that you want to automate with AI versus the experiences that you will deliver with the inimitable power of inspired humans. When your competitors are duped into dehumanizing their experiences in a commoditized race to the bottom by unscrupulous AI vendors, resist benchmarking them into oblivion.

Vast Hype, True Potential

AI is the new electricity. Electricity has been mostly a force for good in the world. In its early days, however, much remained to be learned about using it intelligently and safely. Artificial Intelligence is at the same stage. It will be a force for phenomenal good, yet, as the Internet amply demonstrates, it will unleash many unintended consequences, some very damaging to brands. One company that is on its journey to using AI intelligently is custom suit maker, Knot Standard. Even without AI, using the Luxcelerate system, Knot Standard continues to drive sales growth at a phenomenal clip. The brand's leaders recognize that its superb quality in custom-made clothing, along with its inspired associates, are the secrets of its success. At the same time, they are incorporating AI into their ability to perfect client measurement, improve quality, deliver better product faster, and in augmenting the relationship-building efficiency of its associates with a product recommendation engine.

AI will become necessary and ubiquitous, which is why it will never be the source of a brand's competitive advantage.

For more information and additional insights visit, or contact Luxury Institute CEO Milton Pedraza directly with questions (

Contact Information: