skip to main content

Ovum view

Summary

Artificial intelligence, as the flavor du jour, is overly hyped and often discussed in impenetrable technical terms. This has created confusion around a technology that promises to be the next revolution in productivity. If you strip away the hyperbole and focus on the realistic use of AI in the near term, its role in the enterprise becomes clearer: for the overwhelming majority, it is the automation of analytics.

For most enterprises, AI is the sum of data, analysis, and human-guided scope to act

"I'm sorry, Dave. I'm afraid I can't do that" or "My logic is undeniable" are the representations of AI that most people tend to think of when the word is mentioned. In truth, the immediate reality is far less unnerving, and focused on capabilities that most enterprises would recognize as valuable. AI in the enterprise is about cutting down on the mundane, repetitive tasks that steal time from people who could use that time more productively by being creative, or making everyday decisions more effective with some data-driven insight either by "copiloting" for the user or taking action (within predefined limits) in the background. In both cases, AI is a function of automation.

Putting aside some of the mathematical and technical complexity of AI for one moment – under the covers, complexity is not in short supply – and focusing on the basics, what does an AI-powered capability that achieves those two desirable enterprise outcomes depend on, and look like? That is a straightforward question. First, it requires access to data, ranging from small (familiar, transactional) to big (less structured and unfamiliar), in a computer-digestible format. Second, it requires an analytics engine with access to that data to interrogate it. Third, it requires business context, constructed of rudimentary business rules that help make sense of the data being ingested and analyzed. Finally, its efforts should be shared with the user in the most easily digestible way, or pushed into the background through automation. This could be through a timely pop-up in the user's application of choice, or through automated monitoring and refinement of processes, at speeds or scales beyond human capability. Taken together, these potential benefits of AI within the enterprise can be characterized as automation of analytics.

It's fair to say that the concept of automating analytics is nothing new; terms like "closed-loop decision-making" have been kicking around the industry for a while. However, for enterprises, which will be the buyers and therefore ultimate investors in this technology, the new world of AI can be simplified in a word: automation. Like so many ideas, the concept is often sound, but the technology to achieve it is not readily available – the customer 360-degree view being a classic example, and one, like AI, that only recently came to practical fruition. AI has real, practical applications in the enterprise that might not usher in the end of human endeavor (or worse), but rather free people from digitized manual labor to focus their efforts further up the value chain of work. Keep two things in mind: AI is about making work easier, and hopefully smarter; and the benefit of that is time gained that can translate directly into additional productivity.

Appendix

Author

Tom M. Pringle, Head of Applications Research

[email protected]