Bob Ivkovic, Principal of IT Architects
January 15, 2018
Artificial Intelligence or AI has been around since I was born, and I’m no spring chicken. When I started my career as an Oracle DBA back in 1983, we had an AI group in the IT department. I’m not sure they ever got off the ground, or anyone else had for that matter, because we all had a different understanding and expectations of what AI was supposed to deliver. Whatever it was, I was ready to create some relational tables in Oracle to collect all the data that we would need for AI. This didn’t go anywhere simply because there wasn’t a preconceived and common value proposition of AI. I’m now 35 years into my career and see that many companies are still challenged on the value of AI. However, I do know that AI has come back with a vengeance and it is being refactored into the enterprise technology landscape. The question is, what can organizations be doing now to prepare themselves and maintain their competitive advantage in the marketplace, or even in the global economy? AI is definitely not for the faint of heart, but more for those who want to take the bull by the horns, where the bull is a company’s information technology portfolio and the horns are the AI capability that can be introduced as a layer of IT encompassing all of an organization’s processes, data, and systems, in order to optimize business operations and decisions.
It would be unfair of me to get too deep into this topic of AI without giving an example. If I had artificial intelligence built into my brain, I probably would have given you that example right at the start rather than describe my humble career beginnings. So, let’s cut to the chase. Did you ever get a phone call from the credit card department at your bank? If not, you’re probably one of the few who had heard the story from someone who did. The credit card department may call you because they’ve been notified about an irregular transaction or transactions in your account. I got called once because I made a purchase overseas while I was still on the North American continent, and an hour earlier I made a purchase at the Starbucks two blocks from my house. The bank must have figured out that it’s hard to be in two places at the same time. Another time I got a call because I made purchases at three different gas stations within an hour. The guys at the bank were smart enough to realize that no vehicle requires three fill-ups within an hour. Now, if I filled up my car, went back home and took my wife’s car for a fill-up, and then returned again and took my son’s car for a fill-up, that would be a different story. But, no matter how you look at it, it’s an irregular pattern of activity. It wasn’t the bank people who caught this crazy activity, but computers with artificial intelligence. I truly believe that if it wasn’t for AI, I’d be broke right now.
AI is a Household Reality
AI has become a household reality since I took my first job in IT. It seems to be prevalent in all facets of society, not only banks but airlines, hospitals, travel companies, government, etc, which have all jumped onto the AI bandwagon to serve their customers and dependents. Many software vendors have made it their mandate to optimize the AI algorithms required to keep their customers happy (i.e. provide the competitive products and services that meet their customers’ demands). It is no secret that these companies must take care of their customer needs by leveraging AI. This is probably the same reason companies are putting all their IT in the cloud and let the IT experts manage their systems while they focus on their business. I don’t mean to get off topic, but cloud computing is a major enabler of AI, which we’ll talk about some other time. However, just let me say this: modern cloud computing and optimized algorithms have made AI a fast, efficient, and inexpensive approach to problem-solving. This has become ever so important in the digital marketplace where customers have less interaction with people and are buying everything through the internet by talking to machines – which is really AI without getting into the technical details.
A lot of us who have a frame of reference when it comes to AI [meaning we know something about it], will describe it as machine learning. In other words, machines learn things and respond to situations. And the more they learn, the better they respond. They can learn from people or even other machines. Remember, the bank example? Banking computers with AI capabilities have learned to respond to different patterns, and they learn about new patterns every day and are able to respond better and quicker as they learn more – just like human beings. They are pros when it comes to identifying patterns and then upselling us on more of their services. I’d say some of the AI computers at banks have earned a PhD by now. Their accreditation with a doctorate in AI means that they possess machine learning algorithms to handle any situation. Banks are well aware of AI potential and have become big fans of the AI technologies they have implemented as they strive to be more accurate and more deterministic. Their banking customers are big fans, too.
AI Vendors Rule!
Remember that I told you that I was an Oracle DBA? Well, even Oracle has expanded its horizons and has become a big proponent of AI with the advent of Oracle AI Platform Cloud Service and Oracle Mobile Cloud Enterprise, which incorporates AI. Rest assured that Oracle has AI in most of its business enterprise platforms such as Oracle Human Capital Management Cloud and Oracle Financials Cloud. For instance, when a customer asks the banking chatbot [a chatbot is a computer program which conducts a conversation via auditory or textual methods] to transfer $1000 from account A to account B, AI not only recognizes the customer’s intent to transfer funds but also extracts pertinent information such as the “From” and “To” accounts, the currency, and the amount necessary to construct the transactional statement to execute in the banking application. Technology companies like Oracle are evolving their AI algorithms with cognitive capabilities such as image processing and real-time video processing, ultimately increasing the possibilities of an AI-powered conversational user interface.
AI technologies such as chatbots are only one of many scenarios when it comes to AI and machine learning. Let’s take a common business application like Human Resource Management, which relies on AI technologies to perform almost every HR function serving a company and its employees. Nowadays, many of us are employees of companies with very sophisticated HR applications that guide our careers and promotions, determine work schedules that take into account childcare and vacations, evaluate performance reviews, and connect us with training courses to sharpen our skills, and even plan our retirement. AI technologies appear in everything from image recognition leveraging neural networks to health diagnostics accessing expert systems to data analysis and so on, which have been very instrumental in the evolution of machine learning. You may have also heard about advanced search engines, which learn about a customer’s buying patterns and offer suggestions about other related products that may interest them. If you don’t know what I’m talking about, just go buy or search for merchandise on Amazon and eBay, and see what they offer the next time you log in to their websites. They may just be able to offer you exactly what you need, or at least some comparable alternative that you never knew existed but has wetted your appetite. These are good examples of eCommerce engines with built-in AI using big data, pattern identification, image recognition, and even machine learning to present us with personal and suitable choices. After all, everyone wants our business and the merchant needs AI to optimize our shopping experience. Catalogues have since been replaced with AI and are a thing of the past.
AI is No Longer the Future, but the Present
Like I said at the beginning, AI isn’t new and has been around for decades. What’s really new is how far it has come. The reason I know this is that I was able to do all my Christmas shopping last December without stepping out of the house. To be honest, I did it at all on my work computer at the office during lunch so that I wouldn’t tip anyone off at home. And I have Google, the King of Internet Search, to thank for it. Think of Google as a massive AI engine and whenever someone types a query, Google learns from it and optimizes its search for the next guy. I’m not sure how Santa is going to compete with all this. AI has become a critical technology in solving problems in medical diagnostics, factory automation, pharmaceutical research, text analysis, stock selection, and even playing games such Dungeons & Dragons. If you haven’t played it on your computer, I suggest that you don’t unless you want another addiction in your life. AI just puts it over the top. Technology advancements have also played a role in progressing AI. You see, back in the days, AI wasn’t scalable to make it feasible for mass deployments. Thus, it was left out of the mainstream for all intents and purposes. However, with the break-throughs in microprocessors, graphics, network and database technologies, internet advancements, social media, cloud computing, etc, AI has become a reality we can no longer live without. The emergence of the internet has been a key enabler of AI with an exponential increase in digital information being generated, stored, and made available for analysis. And let’s not forget about cloud computing which brings unlimited processing power, storage, and bandwidth to the AI table allowing companies to deploy applications they couldn’t before. Lastly, the proliferation of algorithms is making AI a possibility in many business applications – Caffe deep learning, Jupyter Notebook data cleansing, Keras, neural networks, NumPy high-performance array processing, scikit-learn data mining, TensorFlow machine intelligence, and the list goes on and on.
AI has reached a required level of maturity and is needed more than ever before by a digital world. It is also easier for companies to implement by embedding AI functionality into their existing systems and making their applications faster, more feature-rich, and more valuable to their business. It’s ironical but Oracle has become a major player in the AI space with its Oracle AI Platform Cloud Service after I gave up my Oracle DBA career. I admit, I’m a bit of an Oracle bigot considering where I started. I’m quite impressed with its cloud implementation and support for the popular algorithms and frameworks, not to mention it’s flexibility using algorithms and frameworks such as Caffe, TensorFlow, and DL4J, while it runs directly on existing Spark/Hadoop clusters for big data computation and analysis. And then, of course, there’s all the database stuff that integrates it all. Machine learning has been part of the Oracle database, management, and security products for a good part of its corporate existence. It was wonderful being part of that corporate and technology evolution.
Anyone without AI experience is well-advised to hire an AI consultant or experienced AI developer to help figure our which AI models or algorithms are best for a specific application, since this can easily become cumbersome or impossible to experiment with all the different algorithms. It becomes even more problematic if a vendor’s AI platform supports only a single model. Thus, a multi-algorithm approach is key, especially since there are so many frameworks, while companies don’t always know which is best and are not always resourced to maintain all of them. This is why companies and their management are big proponents of open frameworks where AI is concerned. It is important to find vendors who are committed to developing and testing all popular machine learning algorithms within its cloud service, so that the functionality will continue to grow and mature over time. Customers have high expectations of vendors offering AI-enhanced software-as-a-service and platform-as-a-service products, and rely on them to continue to incorporate additional AI capabilities to ensure cloud-delivered and on-premises services are more reliable, more performant, and more secure. AI has finally proven itself to companies with complex data analysis and big data requirements, while it emerges in increasingly diverse industries. We now live in a society that depends on AI, and companies that don’t make a calculated use of it will become extinct.
Mr. Bob Ivkovic is a Principal with IT Architects in Calgary, Alberta. IT Architects (www.itarchitects.ca) is an information consulting firm specializing in business process optimization, system evolution planning, and the deployment of leading-edge technologies, including Artificial Intelligence. If you require further information, Bob can be reached at firstname.lastname@example.org or 403-630-1126.