The effects of smart algorithms are manifold
Artificial intelligence (AI) is both overestimated and underestimated. The “intelligence” of the systems, which is still miles away from human intelligence, is overestimated. The impact of cleverly deployed AI algorithms and the speed at which AI processes, customer relationships and entire companies will change is underestimated.
Admittedly, there were already euphoric times for AI in the 60s and 80s of the last century. Analysts now call the long phases in between AI hibernation. In the euphoric phases of the past, similar AI dreams and nightmares were spun as today. Terminator and HAL from Kubrick’s epic film “A Space Odyssey” send their regards.
The big difference compared to today is that we now have the necessary technology and, with cloud and serverless computing, a very flexible infrastructure that can accommodate the extremely CPU- and memory-hungry AI applications. And above all, we have the data that AI needs to learn and recognize patterns that often remain hidden from us.
Three elements are driving development
At the same time, we ourselves have created the need for AI applications. Our world has become so complex, both on a large and small scale, that we will not be able to master it in the long term without significantly more extensive analysis and automation of processes. This applies to almost all areas of human life, at least in industrialized nations.
The three elements of opportunity (CPU + cloud), demand (digitalization, complexity) and motive (increased productivity) will ensure that there will be no more AI hibernation. On the contrary: the further development and adaptation of AI will be faster than the triumph of the smartphone. This will not happen with a big bang. Instead, AI algorithms and applications will soon be found in digital services and application software. User companies should therefore prepare themselves today for this development, which is already underway. Otherwise, they will not be able to realize very good business opportunities in three to five years.
Why is AI so important?
With such high pressure to act, it is not advisable to wait until more users have gained experience with AI technology in order to carry out a risk-free implementation based on the resulting best practices.
Because AI is not a single-task technology. It is not like accounting software, which is used by finance departments in companies to manage income and expenditure. AI is already playing a role now and will continue to play an increasingly important role in processing a wide range of tasks in the future. It is used wherever decisions need to be made on the basis of data and its patterns. This is more often the case than many people think: facial recognition is based on the comparison of data reference points, just like speech comprehension and language processing. There are also examples in the area of predictive maintenance, where stressed components can be replaced shortly before a defect occurs based on comparative data. All of these scenarios are now optimized using AI algorithms. Some things are already working with astonishing precision, but many are not yet optimal. Ethical questions regarding the development and use of AI have not even begun to be answered. But is it therefore right to wait until all open questions have been clarified?
“Get started, don’t hesitate” is the motto
No, it is not appropriate to classify artificial intelligence processes as an overrated young technology. The argument that it is better to wait until the technology has reached the necessary maturity for corporate use does not hold water with AI. There are many good reasons for this.
The increasing complexity of production, processes and decision-making parameters, which is pushing existing IT systems to their limits, speaks against this. Most processes in medium-sized and large companies are IT-supported, and the trend is rising. At the same time, more and more digitally enhanced products and services are coming onto the market. Both dramatically increase the volume of data to be processed. In addition, more and more options are being created for recording and processing data. According to a rule of thumb, the amount of data worldwide is currently doubling in less than two years. The speed at which technologies, products and processes are changing has also increased enormously in recent years. The term VUCA world is doing the rounds; a world characterized by volatility, uncertainty, complexity and ambiguity, i.e. rapid change, uncertainty, complexity and ambiguity. Traditional IT doesn’t cope well with this nothing-is-fixed model. It reacts too slowly because it is difficult to adapt, because different programs are difficult to integrate and, last but not least, because there is an increasing shortage of specialist staff. This type of IT is costly because you have to pay for and operate your own infrastructure and expensive software licenses. For the same reasons, traditional IT is not flexible. Standardization and automation help when it comes to speed. However, both approaches also make traditional IT landscapes extremely inflexible.
Flexibility thanks to intelligent automation
Artificial intelligence with smart automation approaches, for example with IoT analytics, test automation or AI security, should provide relief here. When it comes to flexibility and costs, cloud computing, software containers and managed cloud services can help. However, AI can provide particular support in areas where simple decisions need to be made. When it comes to AI, everyone thinks of granting loans, insurance risk assessments or autonomous driving. Some of this is still a long way off. However, AI has already found its way into areas such as distinguishing perfect from faulty components, translations, speech recognition and processing, searches and recommendations on the web, and the navigation systems we all use every day. The digital voice assistants Alexa, Siri, Cortana and the Google Assistant are bursting with AI. Amazon’s shopping experience would be unthinkable without AI and the many retailers on the platform are also already being intensively supported with AI without realizing it.
AI stimulates growth
Another good reason not to wait and see: many market researchers and management consultants expect AI to have a huge impact on the productivity of companies and on the growth of global economic output. Accenture estimates that AI could generate additional growth of around EUR 1 trillion in Germany alone by 2035. The prerequisite for this is widespread acceptance of artificial intelligence. For companies that want to participate in this growth, this means that they need to focus very quickly and intensively on AI in their processes and in their products and services. The first reports on the benefits of AI already suggest that Accenture’s figures are not speculation: In the IT security sector, 77 percent of respondents in a survey conducted by Market Cube on behalf of security provider Cylance stated that they had prevented data attacks more often with AI-supported security tools than without. In fact, 81 percent said they used AI to detect threats before their security teams. A good example to demonstrate growth and productivity gains and that it is worth taking a closer look at AI.
Many algorithms have outgrown their infancy
Especially as some of the current AI algorithms date back to the 1980s and are therefore not untested. Today’s processors, storage systems, connectivity in the gigabit range and data storage systems (data lakes, NoSQL, in-memory databases, Hadoop, Spark) provide the performance that AI algorithms need to function quickly. This in turn means that many algorithms have long since outgrown their infancy, are often open source and are therefore available free of license costs. This accelerates development enormously. This will make AI available to many companies in a very short time. If you don’t use them, you quickly lose touch. However, it should be clear that standard AI algorithms generally only solve standard challenges. In the business context, many individually developed AI algorithms are needed in addition to these standards.
And as artificial intelligence becomes more widespread, customers’ demands are growing. They always focus on what is high quality, comfortable and inexpensive. Companies that use AI in their product and service portfolio offer a better customer experience; market players that refuse to use these technologies will quickly lose out.
Making data volumes manageable – opening up fields of application
Another reason to abandon the reluctance to use AI is the amount of data, which is growing explosively. Analyzing and using them really exhaustively and quickly only works with the help of AI. Expenditure on this is growing correspondingly strongly in companies. IDC market researchers predict that AI spending will increase to 79.2 billion dollars by 2022, which corresponds to an average annual growth rate of 38 percent from 2018 to 2022. Retailers will spend 5.9 billion dollars over the next four years on solutions for automated customer service agents, shopping advisors and product recommenders. Also among the top 5 industries by spend are discrete manufacturers, healthcare and process manufacturers. According to IDC, the steepest increases in expenditure up to 2022 will be in federal authorities, personal and consumer services and education. Most of the money flows into the application areas of service bots, sales recommendations and sales automation as well as automated threat intelligence and prevention. In addition, the areas of predictive maintenance, diagnostic and treatment systems, fraud analysis and detection, intelligent process automation and intelligent advisors as well as recommendation systems are also recording strong growth.
The application scenarios for AI are diverse and the growth potential of new technologies is huge. Especially as they make a decisive contribution to keeping track of and controlling the exponentially increasing data growth. These are all important reasons for companies to rapidly expand their AI capabilities. The associated momentum is enormous and it is important not to miss the boat. Because one thing is certain: AI will catch on faster than smartphones have done.
To gain further insights into the topic of “AI”, we recommend our white paper entitled “Artificial intelligence in today’s corporate world”. Find out how you can stay one step ahead with AI.
