AI: Forget the hype. Show me the ROI!

The media is filled with numerous articles about the power of artificial intelligence (AI) and how the technology is transforming entire industries with tectonic shifts and disruptions. Driver-less trucks, robotic fast-food chefs, virtual customer-service agents, and even AI-powered virtual lawyers are either on the horizon or projected to arrive soon. And I believe that many of these fantastical prognostications may well come true

However, what is written about less is the real progress businesses are making with practical applications of AI. A recent Deloitte survey of 1,100 US executives from companies considered to be early AI adopters found that 82 percent report a positive return on their investment.

And so it is with this in mind that I have written this guide. The goal is to help you plan and evaluate an AI initiative in the context of your business and what it means for the bottom line. Or more simply put, as Cuba Gooding Jr.'s character proclaims in the film Jerry Maguire, how AI can “Show me the money.”

In this article, I consider AI to be an umbrella term encompassing advanced technologies such as machine learning, natural language processing (NLP), and robotic process automation (RPA). Machine learning uses large numbers of data samples to determine patterns, as opposed to traditional programming techniques that make use of prewritten lines of code. NLP includes such applications as speech to text, chatbots, and automated video annotation. RPA technology provides a means to quickly automate repetitive processes such as opening new bank accounts.

So let us begin on our AI journey and explore how AI can help improve your bottom line.

Are you ready for AI?

Before you take flight on your AI journey, an honest assessment of your organization’s readiness to embark on an AI initiative is essential. Making this effort will save you time and money. I consider the following three categories to be important in assessing your readiness: (1) existing key metrics, (2) data quality, and (3) organizational resource requirements. Let us examine each of these readiness factors:

Existing Key Metrics: If your organization does not already have metrics in place, now is the time to define and implement them. Get a snapshot of the health of your current operations, and track these metrics for a period of time. If possible, analyze historical data to determine trends before implementing an AI system. Otherwise, you will not be able to objectively measure the success of the new system.

Without having metrics measuring your business processes in place, it will be difficult at best or impossible at worst to quantify the impact of an AI initiative on your business. Examples of relevant metrics might include error rates, time to resolution, transactions per hour, quality measures, and others depending on the nature of your business. By measuring them against benchmarks, you can establish what to expect from your AI initiative.

Data Quality: During a pre-flight check, an aircraft pilot determines if he or she has enough fuel and also confirms that the fuel is clean and free of water. Data is the fuel for many AI efforts and key to their success. Do you have the data you need? Is it clean and normalized? Is there enough history to train the models you wish to implement?

You must explore these questions before you can move forward with a successful AI initiative. In some instances, the data you have may be from a legacy system in which there is plenty of history, but the quality is less than optimal. In many cases, this will be a challenge for implementing a proposed initiative but will not be a deal breaker. It may just require more sophisticated modeling and tuning to get the desired accuracy and result.

Organizational Resources: Do you have the talent needed for your project and the technology required to support your initiative? The good news on the talent front is that only a short time ago, an organization needed to have data scientists on board to build any meaningful AI solution.

Today, there are many cloud-based solutions that can help you get started without having in-house data science expertise. Amazon, Google, Microsoft, and numerous new startups such as DataRobot, BigSquid, and Talla all have offerings making AI projects orders of magnitude easier than ever with built-in algorithms, web services, and other tools.

There are also purpose-built solutions for functions and vertical markets like Condati for marketing or Health Fidelity for healthcare to help you put AI to work more quickly without the commensurate development time and effort. For example, some vendors prepackaged machine learning solutions that optimize marketing campaigns or predict customer churn. APIs enable you to efficiently operationalize AI services, such as chatbots or machine learning models, and benefit from the results more quickly.

One thing to note is that many times, members of analytic teams are eager to work on AI projects, and some are already putting in the time independently to educate themselves on AI techniques. These team members would love the chance to participate in a real-world AI initiative even if it means putting in the extra time outside normal work hours to learn and experiment.

That being said, you may still have to decide if you need to hire additional talent or outsource some of your development depending on your specific project and the expertise you need to complete the effort.

AI and your business strategy

It is natural to think that an AI journey starts with technology. It does not. It starts with your business strategy.

There is an exchange between Alice and the Cheshire Cat in Alice’s Adventures in Wonderland in which the Cat says, “If you don’t know where you’re going, any road will take you there. . . . Then it doesn’t matter which way you go.” When embarking on an AI journey, not knowing where you are going can be a costly adventure.

That is why your business strategy is crucial when launching any AI initiative because it shapes the direction and goals of your project. Without a sound strategy, you will be unable to define a clear destination, wasting time and money.

 One way to approach this challenge is to leverage Harvard professor and business luminary Michael Porter’s three generic strategies of cost, differentiation, and focus. His key concept is that a company cannot be all things to all people. It must choose among these strategies or strategic levers to maximize success and profitability. So you can either choose to be a low-cost leader or offer a differentiated service or product. Likewise, you can also decide whether to sell across market segments or to focus on a particular vertical market segment.

With a well-defined business strategy, you are in a position to leverage AI to magnify your unique strategic levers. For example, if low-cost leadership is your strategy, then AI could support this by reducing manpower requirements for customer service with chatbot-powered agents.

Likewise, you could also leverage AI to employ differentiated service by offering faster service or more in-depth knowledge. The objective here is to have a clear vision on the strategic levers you want to affect using AI.

To bolster your success, I highly recommend you develop an AI road map to flesh out your vision, articulate its impact on operations, determine resource requirements, and analyze the organizational requirements and impact.

Conclusion

As I have noted, your strategy is a critical component of setting a course toward a successful outcome because it will serve as your lodestar to achieving a desired ROI from your AI initiative.

Once you have locked down your strategy, you are in a position to reap significant benefits. Cost reduction is a common benefit from implementing AI. However, a less obvious benefit of AI is that it can remove barriers to growth. It can reduce the need to hire and train additional staff, enabling a company to grow faster with far less friction. For a business looking to achieve scale, this is a critical benefit and one that you should consider as you evaluate a proposed AI initiative’s viability.

AI adoption is still in its infancy—it is estimated that only about one in twenty enterprises have adopted AI. Early movers will achieve powerful and enduring competitive advantages while the laggards who delay do so at their own peril. I hope this guide helps you start your AI journey soon and reap significant rewards.

Authors note: I would like to thank my friend and associate, Jay Mason, a leading AI and big data practitioner, for his wise counsel and contributions to this article.