Why Artificial Intelligence is the Anti-Strategy for FP&A

Why Artificial Intelligence is the Anti-Strategy for FP&A

By Michael J. Huthwaite, Founder and CEO of FinanceSeer LLC 

artificial intelligenceArtificial Intelligence (AI) is everywhere these days. At the time of this post, it is perhaps one of the hottest topics in the entire Tech sector, with venture capital lining up to invest heavily in anything that even sounds like an Artificial Intelligence play. Quite simply, the potential application for AI is so far-reaching that it’s almost difficult to find industries where AI won’t have a significant impact. As a result, it’s clear that AI will be around for a long time.

Despite the recent explosion in Artificial Intelligence it is important to note that this concept is not new.  In fact, AI has been around for quite a longtime, but it’s taking off now due to the proliferation of several key technical factors which include. 

  1. Cloud Computing/PaaS (access to faster computing capabilities)
  2. Big Data (ability to store and comprehend large sets of data)
  3. IoT/Search/Mobile/etc. (ability to collect data at the Source)

In order to get a better understanding for AI, it’s important to realize that although access to faster computers and Big Data technologies was a challenge in the past, it’s widely accessible to most companies today.  As a result, the real foothold into meaningful AI comes down to data collection.  This is something that companies like Google have been doing for years (via Search).

Imagine what unique data a company such as Whirlpool could collect if they know what time of the day you open the refrigerator, what sections of the refrigerator were accessed and how often items were replaced.  This is information that no one else can collect.  Collecting this type of information is what will fuel AI for years to come. 

AI as a Strategy?

Artificial Intelligence is going to have a big impact on FP&A.  But there is one area that I am very concerned about and that is the perception that Artificial Intelligence will have with regards to strategy. 

I’m worried that companies will sit back and begin to rely solely on AI to make short-term automated efficiency choices rather than focusing on making strategic decisions that require upfront investments in order to break down barriers or boundaries so that real progress can be made. 

Over the past 15 years, we've seen a shift of focus from long-term value creation to short-term goals due to advancements in technology and that’s not even largely attributed to AI.  Imagine what lies ahead? 

Of course, I’m not against the benefits of AI, but they largely relate to operational efficiency and this is by definition not a strategy. 

Why AI is an anti-strategy

Current State Algorithms vs Future State Algorithms

Historically, the world has been driven based on static algorithms such as E=MC2, A2+B2 = C2, π=C/d, etc.  At times, these formulas may have required us to operate in a vacuum, but they have undoubtedly served us well.  
Now we are entering the era of Artificial Intelligence/Machine Learning where the “current-state” algorithms that drive the world are constantly evolving enabling us to take advantage of real-world (vacuum-less) issues that constantly adapt based on environmental and social factors.

This new power will undoubtedly create an enormous amount of efficiency, but I don’t believe it directly impacts strategy which is more focused on identifying optimal “future state” algorithms.  

In Financial terms, future state algorithms might mean evaluating various investment/acquisition targets that would fundamentally change the business model in the future.  This is difficult for AI to identify because the AI process relies on current state algorithms.  Sure, AI could play a role in helping guide strategic decision making, but I believe strategy will always remain a largely a human driven exercise.

Strategy is often about relying on less data

Artificial Intelligence, for the most part, relies on large volumes of data to make accurate inferences.  This data is often transactional data and therefore is information that is immediately relevant.    This data can be particularly helpful in making short-term tactical decisions that often focus on efficiency, but it tends to be less suited for Strategic Planning.  

As we start to evaluate longer term time horizons we tend to look at data at more aggregate levels and infer either some sort of trend or simply establish a high-level target.  These trends and targets can be discussed or debated, but the focus is no longer in the realm of big data. 

CEOs, CFOs and other executives may want to leverage AI in their analysis, but in the end, they often make their strategic decisions based on the realistic ability to influence the future based on the understanding of a handful of key assumptions and not on reams of transactional data. 

Competition could result in a fight to the bottom

I find it interesting that the companies that seem to rely on AI the most (Amazon, Google and Apple, for example) all struggle to remain six months ahead of each other in ultra-competitive markets.  Take the Mobile phone market for example where each new release seems to be focused on trying to one-up each other. 

Is this really creating strategic value if your competitive advantage is in jeopardy every few months?

By relying on AI as a primary answer to competition, companies stand a real chance of shortening the life cycle of their products.  This race to the bottom can be avoided if companies continue to focus on maximizing strategy rather than solely optimizing efficiency in competitive markets.

Conclusion

Artificial Intelligence is going to be a highly impactful technology in the coming years.  Yet, like most forms of automation, the major benefits are going to be efficiency related. 

The strategy is a function best suited for the human mind. Of course, technology can play a role, but the only way technology can replace human thinking is if we as humans choose to step aside… and I, for one, am betting on the humans. 

 

 

The article was first published in prevero Blog