AI ROI can be calculated in tangible and intangible returns
“In the new world, it’s not the big fish that eats the small fish, it’s the fast fish that eats the slow fish,” warns Klaus Schwab, founder, and executive chairman, World Economic Forum. That’s a great visualization to motivate enterprise leaders to get serious about implementing AI technologies, and understanding AI ROI, so their company isn’t left behind. The problem is that according to Avanade’s Human-centered Artificial Intelligence Survey, 88% of global executives believe companies incorporate AI because it’s a hot topic; but they don’t know how to use it, much less achieve any level of ROI from it.
In some respects, David Schubmehl, who follows the AI field for research firm IDC seems to agree. He writes, “People are starting to kick the tires, looking to see how it can help their business and the bottom line. The ROI evidence is not yet clear, but it's starting to happen.” That said, IDC forecasts worldwide spending on cognitive and artificial intelligence (AI) systems to reach $57.6 billion in 2021.
AI’s biggest investors
Retail, banking, manufacturing, healthcare, and process manufacturing industries will continue to be the industries with the largest spending amounts throughout IDC’s five-year forecast. By 2021 their combined investments will represent nearly 55% of all worldwide spending.
In his MoneyInc article Uncovering the ROI in AI, Aaron Reich proposes that in any market, there are almost an infinite amount of problems that AI can solve. Moreover, it’s highly unlikely that any single AI product will solve every problem for any single business. He suggests that it’s better for companies to solve existing, specific problems first, using different AI technologies to maximize their ROI.
Cutting through the hype, how are AI applications helping enterprises? And how are their investments driving returns?
Uday Kamath, Chief Analytics Officer at Digital Reasoning, describes in a recent Gigaom submission, a use case that AI solved for financial institutions. “Today, smart machines capable of understanding the true meaning behind human communications are augmenting the work of human analysts at most of the world’s major investment banks.” The technology extracts messages that indicate misconduct, while entity resolution and knowledge mapping help analysts identify sources of human risk and hidden networks of collusion. Among the compliance organizations of leading investment banks, widespread adoption of AI-enabled analytics has taken place in less than three years. He points out, “It’s no exaggeration to say that, for these organizations, regulatory compliance would be impossible without the amplifying effects of AI.” It’s clear given the stiff penalties GDPR has brought about, that there’s significant ROI derived from this application.
Flying the AI-friendly skies
Kamath offers another scenario where the ROI could be demonstrably assessed. He says, “Imagine an airline being able to consolidate insights captured in tweets, at a call center, or in emails into a heat-map of problems and opportunities.” The AI could rapidly see issues with the quality of meals emanating from a particular supplier. It could see trends in requests for unserved destinations and glean early insight into the probable popularity of a new route. It could uncover issues ranging from confusion about security procedures, praise for great service, or ignorance of company policies. “Such an airline would turn the inputs of its employees and customers into a valuable asset for management,” he wrote.
The aircraft manufacturer Airbus is a big believer in the ROI of AI. It has an entire division devoted to it. The company offers an option on its aircraft to minimize ATC miscommunications since a missed air traffic call could spell doom for hundreds of passengers. In the European airspace, much conversation happens in heavily accented English. This makes it difficult for pilots and controllers to understand each other. Airbus, as part of its AI Gym platform, took on this problem using AI, machine learning and speech recognition. Airbus only recently industrialized the initiative but has been running it in concept mode. This is just one example of a product that you wouldn't think of, but it's something the company undertook to solve because it recognized the need. In this case, the ROI can be translated into human lives saved.
Not all calculations are as weighty. Calculating ROI in dollars is easier to do when there is an existing problem and budget attached to it. Once the AI solution is in place, it’s a math equation that involves comparing the fully burdened costs of both to each other over a given amount of time and adding in the expected revenue stream.
AI-first and AI ROI
It’s years away for most organizations, but the concept of businesses transforming into an AI first organization like Google has been morphing itself into since 2017, is not so far-fetched. Much like SaaS providers embraced mobile-first, enterprise leaders; once they understand the technologies can adopt an AI-first strategy. They can use the lessons learned in financial services, aviation and other industries similar to theirs to invest in solutions with proven results and a quantifiable ROI.