In the movie Avengers: Infinity War, Dr. Strange uses the Time Stone to explore 14 million possible futures to find the one with the best outcome, the one where the Avengers could defeat the evil Thanos. Dr. Strange has to experience each possible future and consider the decisions that were made to see how each future ends. If only he had Evolutionary AI.

Sadly, of all 14 million possible futures, there’s only one in which Thanos is defeated. In all the others, the team fails, and billions of lives are lost. While some futures are likely better than others, only one produces the optimal outcome for the team. These possible futures exist (in the movie and for all of us) because of all the decisions that can be made – a nearly infinite number of them –  which then lead to new conditions and even more decisions.

Decisions, Decisions

Each of us makes decisions continuously, and we work in businesses that also make decisions continuously, often based on incomplete information, and always based on our limited bandwidth. Academic studies, including one from the University of Queensland, contend that human decision-making suffers greatly when we have to consider as few as four variables. Research even suggests that when humans need to consider optimizing across five variables, their decision-making is not much better than chance.

Consider the daily staffing requirements for a home goods store in Dallas. The store manager can assess the day’s weather forecast and the major sporting events on the docket, but she probably doesn’t know about the car accident on 35E that is clogging traffic in the vicinity, the power outage nearby, or the emotions of people tweeting within five miles of the store. These are all factors she needs to know to properly (and proactively) serve her customers.

We’re generating more data than ever before via our daily business transactions and social media interactions across a multiplicity of devices. In fact, more data was generated in the past year than in the past 5,000, and the rate of data generation is only expected to increase.

As we run our businesses – from optimizing staffing, inventory and patient care Rx’s, through approving insurance policies, maintaining equipment or just what to get for lunch today – we all need a Time Stone to see every possible outcome and choose the best one. Until now, that was purely the realm of science fiction.

But what if we could predict the future? What if we could simulate every possible choice, even the ones we would never even think to try, to find the best choices for the very best outcome?

Applying AI to Foretell the Future

Evolutionary AI enables exactly this. This branch of artificial intelligence leverages two powerful concepts:

  • Creating a virtual representation of your business. While modeling is not new, using evolutionary computation to generate and optimize models of the business is, and it’s a very powerful way to create a model without requiring large numbers of data scientists to design it.

    This approach uses business performance data and past decisions to create a model and refine it to reflect the most accurate view of the business. And because it evolves, the model changes as businesses change. Adaptable models are required to accommodate new regulations, modifications to bills of materials, demographic shifts and all the other factors that impact future decisions.

  • Optimizing your business through simulation. Most business decision-making today is really an experiment on real people, machines and customers. We try a new website design in hopes that it encourages customers to buy more stuff, but sometimes it introduces more scrolling or clicking and turns off users. We try to increase compensation just enough to ensure employee retention, but all it takes is a new competitor with better child care and more flexible hours to lose valuable skills.

    What if you could test every decision on a virtual copy of your business and see the impact on revenue, retention, conversions and other important metrics? And, what if you could test the decisions you never even considered to generate better outcomes?  

    For example, through online and in-store behavior, retailers can glean an enormous amount of data on consumer wants, needs and activities. Evolutionary artificial intelligence can discover new solutions and make recommendations for e-commerce site design and in-store product displays that maximize revenue.

Together, modeling and simulation deliver the power of evolutionary computation. Using large numbers of parallel simulations of the business and possible decisions, organizations can find the best possible outcomes. And by learning from these simulations – in parallel as they play out – they can uncover optimal outcomes more quickly and cost-effectively than any other method.

Continuous Change, Continuous Evolution

The future of business is based on systems that are always modeling, always simulating and always recommending the best choices for decision-makers. It will increasingly be considered irresponsible to make important choices without first consulting with an evolutionary artificial intelligence system.

When you need to consider four or more variables, an Evolutionary AI system can identify the best outcomes and even make recommendations you’ve likely never considered. To unlock this potential, business leaders need to channel their Avenger super powers and apply AI to drive new levels of business results.

To learn more on this topic, visit us at our website.

Bret Greenstein

Bret Greenstein

Bret is Senior Vice President and Global Head of Data at Cognizant, leading Cognizant’s Global Data Practice. He is focused on helping... Read more