Artificial Intelligence and Business Value
Digital technologies are pervasive. Nearly 5 billion people in the world now have a mobile phone connection and more than 7 billion mobile phones are in use (some people have more than one phone). Approximately 2.5 billion of the phones are smartphones. Cell phone penetration is now approaching that of electricity — about 88% of the world’s population, or 6.7 billion people have access to electricity.
At the same time, advanced digital technologies, particularly artificial intelligence, are getting embedded in all sorts of physical “things,” from cars to toasters to thermostats. Every smartphone has some element of artificial intelligence, meaning that AI is touching at least 2.5 billion of the world’s population. Every corporation has been scrambling for the past three years to make sense of the digital onslaught and to ensure some semblance of a strategy is in place. During this time, various people and sources have endeavored to project the value that corporations, people, and indeed the world will derive from digital technologies, and more specifically, artificial intelligence. This is all part of a natural cycle in techland of creating excitement and FUD (fear, uncertainty, and doubt), followed by the predictable feeding frenzy of investment, leading then to some great breakthroughs along with some equally great failures.
Here are some of the business value claims (paraphrased) I have read for some of the key technologies of our day:
What are we to make of these proclamations? Is this a fugazy Rolex, or is it the real deal? From a distance, you can’t really tell a fugazy Rolex from the real thing; it’s only on closer inspection that you can tell if it’s a fake. So, let’s take a closer look.
Early in my career, I worked on supply chain transformation project at a Fortune 50 manufacturing company. In the first stage of this program the team executed a process to develop a plan for the transformation. During this first stage, we mapped existing processes, determined the to-be vision, identified solutions for getting from point A to point B, quantified the value for getting from point A to point B, and developed a time-phased value-delivery plan. This delivery plan included a payback analysis based on delivered economic profit (also known as EVA or ROIC). The overall plan was then delivered to C-level executives for approval. This approach is now fairly standard in the software business. (while this approach was novel in the software business at the time, companies had been using similar approaches for decades, at least since the development of the DuPont model in 1915).
Even though I was to later work on dozens of transformation programs, there was a funny thing that occurred at the beginning of this early transformation program that stuck out at the time and I have always remembered thereafter. The client program lead, who was smart, seasoned, and somewhat sarcastic, got the team together to discuss the status of the value discovery part of the process. He made a funny comment, “just remember, cumulative value from each of the improvement areas cannot exceed revenue.” In the feeding frenzy for improvement, many projects go after the same improvement buckets; for example, inventory is a common target for supply chain projects. What the program lead was trying to say is that if the company added up all its individual departmental and corporate inventory improvement project targets, they might just exceed the amount of inventory on the balance sheet. Thus, it was important to scrutinize closely the values of individual improvement targets and to ensure that there was not double-dipping across projects.
And the same goes for multi-trillion dollar improvement claims for digital technologies.
One of the problems with the grand value proclamations is that they don’t have a lot of definition or science behind them. Some people say increased sales is business value; others might say cost reduction. Both are business value, but a more formal definition would also consider costs. A common formal definition of business value is economic profit:
Economic profit = net profit after taxes minus the cost of capital.
In other words, it’s my revenue minus all my operating costs minus the cost of all the capital I need to deploy to create my revenue. This is a fairly standard definition that has been around the business world for more than 100 years. Let’s use this thinking to see if any of the trillion-dollar value statements pass the sniff test.
Let’s say the world’s economy is a business. If this were so, the world’s business would have about $82 Trillion in revenue and about $6.2 Trillion in net after-tax profit. The global growth rate is approximately 3.7%. Now, let’s say that business value means profit (since the cost of capital deployed is very difficult to get for the global economy), and that AI is going to create about $1 Trillion in incremental annual value. This means that AI would increase net profit by 16%; if the incremental value were $2 Trillion, then AI would be responsible for a whopping 32% improvement in net profit. Now, let’s say there is no margin expansion, but AI creates such value through sales increases. For $2 Trillion in incremental profit, AI would have to create $26 Trillion in additional GDP. This would require a huge uptick in the global growth rate, which is highly unlikely. What about $10 Trillion in incremental value? Fuhgeddaboudit.
You can probably see where this is going. The numbers don’t pass the most basic sniff test, at least not when using standard definitions of business value.
What might we be missing?
What is the business value of electricity? That’s like asking what is the value of water to the human body. A reasonable way to think about its value is by taking it away. If you take away water from the human body, the human body will cease to exist. If you take away electricity from the modern corporation, it likewise will cease to exist. The business value of something like artificial intelligence is different; it’s not like electricity at all. It’s not like water; it’s more like wine, or music, or dessert. If it ceases to exist, business will still go on, but life will be a heck of a lot less fun. In other words, in Maslow’s hierarchy of needs, the lion’s share of business value of AI and similar technologies will accrue towards the top of the pyramid, allowing for the achievement of ever higher levels of enlightenment. For example, in retail, higher levels of enlightenment come in the form of a highly personalized experience in product and fulfillment choice.
So, if this is true, then how does one place a value on it? The answer is simple — there will be winners and there will be losers. A $1 Trillion value pie will likely look something like this: $800 Billion gained by the winners, $800 Billion lost by the losers, and $200 Billion of value tide that raises all ships. In this simple example, 20% of the pie will accrue to economies and societies overall and 80% will be a zero sum game. My guess is that the 20% is closer to the real economic value than is the 80% that represents winners and losers.
Thus, if today’s business competition occurs largely at the top of Maslow’s hierarchy, as manifested in personalization, and if AI is a critical ingredient in providing these capabilities, then the value of AI is indeed huge. This is particularly true in mature markets, where consumers no longer view wine, music, and dessert as nice-to-haves. Nice-to-haves become must-haves at an ever-increasing pace. AI will increasingly determine winners and losers, even if it does not increase significantly the size of the overall pie. And, unfortunately, its value is likely to accrue unevenly, with those that are already over-served becoming increasingly so.
For businesses, this essentially means considering what is called the “do nothing scenario.” This is a mindset put forth in the Harvard Business Review classic “Innovation Killers,” by Christensen, Kaufman, and Shih. As the authors state “The first error [in payback analysis] is to assume that the base case of not investing in the innovation — the do-nothing scenario against which cash flows from the innovation are compared — is that the present health of the company will persist indefinitely into the future if the investment is not made.” This essentially means you must consider the possibility of going out of business if you do not invest.
Therefore, most of any multi-trillion-dollar value statements for AI and digital technologies will be about avoiding value destruction versus creating incremental value in the form of increased sales or profits. The value will be hugely incremental versus the do-nothing scenario, but very small when compared to the current baseline of revenue and profit.
If this sounds a bit like a circular discussion, it’s because it is. The previously enumerated value numbers are both right and wrong. They are wrong in saying that AI will create and deliver to society that much value; but they are right insofar as the winners are concerned. Taking the cumulative value of the winners and subtracting that of the losers and then subtracting the investment required will likely result in only an incremental positive to the world’s GDP growth rate. At one time electricity also determined winners and losers, but it did so while also significantly expanding the overall pie.
So, what’s the verdict? Is the Rolex real or is it a fugazy? The answer is neither. The watch is real and authentic, but it’s neither a real Rolex nor a fugazy Rolex. It’s a different kind of watch altogether — let’s call it a Cartier, and a real one at that.
Originally published at www.worldlocity.com on February 8, 2019.
Artificial Intelligence and Business Value
Research & References of Artificial Intelligence and Business Value|A&C Accounting And Tax Services
Source
0 Comments