The Amazon acquisition of Whole Foods represents yet another confirmation of our rapidly changing business environment driven by opportunities at the confluence of technology and network dynamics. Although only the latest in a powerful trend initially impacting in this case the grocery industry, the business and technology issues driving the strategy are relevant to most and serves as a reminder that digital convergence is not confined to traditional thinking or industry lines.
A sharp devaluation of public grocery companies followed, so apparently many investors share concerns highlighted in the current HBR article “Managing Our Hub Economy”, which warns that “most companies will not become hubs, and they will need to respond astutely to the growing concentration of hub power”. The devil in the details for management is how to respond astutely. The answer is largely an AI OS.
A recent article at MIT SMR describes the complex operating environment: “The Five Steps All Leaders Must Take in the Age of Uncertainty”:
These ecosystems are nested complex adaptive systems: multilevel, interconnected, dynamic systems hosting local interactions that can give rise to unpredictable global effects and vice versa. Acknowledging the unpredictability, nonlinearity, and circularity of cause-and-effect relationships within these systems is a notable departure from the simpler, linear models that underpin traditional mechanistic management thinking.
One of the reasons for the attention in this latest combination is the integration of virtual networks with physical locations, which has been a priority for many companies, including Kroger, which shares many of the same zip codes as Whole Foods. A few days following the announcement Kroger Chairman and CEO Rodney McMullen revealed that he wasn’t surprised: “you could tell that Amazon wanted to do something from a physical asset standpoint and I think Whole Foods is a great fit for them.” Kroger is a well-managed company known for wise use of analytics, which is reflected in McMullen’s advice to investors: “you should assume that we look at any potential opportunities”.
“We needed a new operating system” – Doug McMillon, CEO of Wal-Mart Stores, Inc.
The question is will traditional methods be sufficient moving forward? The answer may be found in part by looking at the world’s largest retailer. Wal-Mart was viewed as one of the most stable companies before Amazon entered their core lines of business, eventually leading to the recent conclusion by Wal-Mart CEO Doug McMillon: “We needed a new operating system”. The company recently paid $3 billion for an e-commerce component of a new OS in Jet.com, which may seem excessive to those of us unaccustomed to managing a half-trillion USD in annual revenue, though represents a relatively minor investment if it works well.
Unfortunately for 99.99% of businesses, investing $3 billion in a native e-commerce platform is not an option, particularly one experiencing significant losses. Very few of the remaining .01% could consider doing so for a partial OS. Even Wal-Mart’s bold actions appear insufficient when we consider that the acquisition of Whole Foods was powered less by the core business of Amazon or Whole Foods than the bold manifestation of what was previously learned, resulting in an entirely new and much better business model in AWS (see article on spinning off AWS). Amazon’s strength is its ability to learn rapidly, recognize potential, and then convert and realize interconnected opportunity to new offerings in a fiercely competitive manner.
Competing in such a hypercompetitive and rapidly changing environment can be especially difficult for companies thinking and behaving in a linear manner. The retraining for me personally that began in our lab in the 1990s was a profound voyage initiating from a relatively high level. The technical training and transition involved a sharp learning curve that has only become steeper and more intense with time.
Native platforms are quite different than corporate networks that have evolved incrementally over decades. Understanding related opportunities and risk many years in advance is a critical challenge. One must peer through an asymmetrical prism constructed from tens of thousands of hours of total immersion and make bold bets that are well timed, particularly with AI systems.
Among many lessons learned is that the network economy is not only interconnected, it is also multidimensional and pre-programmed. When managed optimally and competitively the entire experience of the customer is an obsession with little deference for traditional lines.
Important considerations for an AI OS
1 – A competitive AI OS will be necessary for most to survive
Essentially all the evidence we see with mid-size companies to market leaders across most industries is that a strong AI OS is rapidly becoming the new competitive bar. If a company doesn’t have a competitive AI OS platform and the competition does, it will likely negatively impact the entire organization and each individual within it, as well as customers and partners. Google and Amazon are examples of companies that appear to be employing some functions similar to those found in our Kyield OS. While leadership and corporate strengths are critical, employing advanced AI systems is among the most important improvements any organization can make. The question really is how and when.
2 – An organization OS is not a computer OS
Many different types of operating systems exist. A few minutes of reading my book (condensed version) Ascension to a Higher Level of Performance will highlight the difference between the Kyield OS and a computer OS. The standard system is focused on universal issues common to all organizations, individuals, and networks. We have good reason to believe Kyield is among the world’s competency leaders in knowledge systems, which is a sub-specialty of artificial intelligence.
Our focus is a thin yet broad and very deep specialty with little overlap to most others, including AWS, Azure, and Google (Kyield OS integrates well with most others). Although executed with software, the Kyield OS is a ‘low code’ system compared to a computer OS and more dependent upon data and algorithmics. The system operates in the background with a simple natural language interface for corporate, group, and individual administration. The Kyield OS is transparent, non-intrusive, and interoperable so that any function can be added as needed in a highly efficient manner.
A recent note from a Fortune 50 CXO exemplifies the need from a slightly different perspective in response to our recently released HumCat offering—a new model for prevention for human-caused catastrophes, including cyber prevention.
I like your idea of an Operating System. I’m so convinced that the world is too complex and getting more complex every second that human beings cannot manage it in the right way anyway… Now, it is time (for the Kyield OS), otherwise we are on the hook of dark side of cybernetics—cybercrime or cyberwar and nobody can defend us.
3 – Reinventing AI system wheels is not wise
As I shared with a Fortune 20 team recently, while it may be extraordinarily easy to underestimate the amount of tradecraft and secrets for such an endeavor, it is nonetheless foolhardy to do so (hence the AI talent wars). Fortunately for our customers, we’ve done the bulk of the heavy lifting. It was two decades ago this year that Kyield was originally conceived in the lab as an authentic invention (Optimizing knowledge yield in the digital workplace).
Many research and consulting reports on AI systems are available, and they have improved significantly over the past two years (See reports by MIT SMR & BCG and Nordea as recent examples), but caution is warranted. Some consulting firms are still advising to start small and experiment in areas that no longer need experimentation. Although generally appropriate five years ago, it is increasingly dangerous today as the competitive gap due to AI systems is expanding rapidly.
A good example of an ongoing experiment was highlighted in the WSJ CIO Journal: “Swiss Re Bets AI Can Help Workers Cope with Complexity of Reinsurance”. The goal is admirable, achievable and sounds impressive until reading the subtitle: “The company’s 100 data scientists and AI experts are building software that can read documents on their own.” This is not a new technology. If the article is accurate it appears that Swiss Re is spending between 10-100 times more for a small fraction of the functionality found in our Kyield OS. Other options also exist for the specific function described that would likely be much more wise than a custom effort.
Our friends at Swiss Re are far from alone. Munich Re publishes an IT radar report (with a nice diagram) based on research that “systematically analyzes opportunities, trends and technologies, and provides an ongoing insight into which technologies could be relevant for Munich Re and our customers from a very early phase”. In the current 2017 report Munich Re places predictive analytics in adoption phase while advanced machine learning and robotics process automation in the trial phase. These and other recommendations may raise some eyebrows. Predictive analytics has been deployed for many years as has advanced machine learning for specific purposes. If one is competing with a technical leader—and increasingly most are, waiting too long can be a fatal error. The first mover position, however, is not always an advantage, so such decisions need to be situation-specific.
4 – Method and sequence of adoption
To date the super majority of investment in AI systems have been strategic resulting in a few notable successes. We have also witnessed large and costly errors, including in M&A, VC, internal development, and in system designs and business models by vendors.
Horizontal systems like our Kyield OS serve as an efficient platform to unify the organization and ecosystem. Ideally a native AI OS should be adopted first. Quite apart from significant IP liability risk, since our standard system is focused on universal issues for every type of organization, with improved productivity, security, and prevention, it is difficult at best to justify internal custom efforts that replicate any of this functionality. Strategic functions are best built on top and/or integrated with our networked platform OS so that the organization and ecosystem operate in an optimal manner.
All is not lost, however, for those who have experimented in overlapping areas. The modules within the Kyield OS creates the data structure needed for compliance and then populates across the network in a manner designed to execute the functionality within the system as efficiently as possible, including for accuracy, integration and financial efficiency.
As important as external competition can be in this environment, the degree to which displacement will occur is dependent on a number of factors. All things being equal otherwise, the outcome primarily depends on the incumbent’s actions, its people, systems, and processes. Even though some companies may seem well positioned, the fundamental economic and business environment is rapidly changing. To the best of my awareness, survival from this point forward will essentially require a strong AI OS for the super majority of organizations.
Mark Montgomery is the founder and CEO of Kyield, originator of the theorem ‘yield management of knowledge’, and inventor of the patented AI system that serves as the foundation for Kyield: ‘Modular System for Optimizing Knowledge Yield in the Digital Workplace’. He can be reached at email@example.com.