How Your AI Deployment Might (Actually) Live Up to Its Promise
Every week, it seems we see a new report projecting rapid growth in AI deployments. A few recent reports project:
- By 2019, 40% of digital transformations will utilize AI.
- By 2020, 75% of businesses executives will actively implement AI.
- By 2025, the AI market will grow to $100+ billion annually.
These are big numbers, and they seem speak to a real need for AI. But while the reports are happy to project huge growth in AI deployments, they are less clear about what organizations will get back from these investments. Within the reports, the benefits executives say they want to derive from AI often sound thin.
In one report, 79% of executives wanted to deploy AI because they felt it would “make their job easier and more efficient”. In another report, 80% of executives stated they wanted AI because they felt AI “boosts productivity”. And in a final recent survey, 72% of business leaders stated AI would simply offer a “business advantage” (often without stating what that “advantage” might be).
Now, we are not looking to set up a strawman here. When a new, hot technology hits the scene, executives often have to fluff a little around the subject. But the fluff isn’t helpful if you are considering an AI deployment, nor if your business stakeholders ask you about this technology and want to know how it might bring value to your organization.
In order to get past the fluff and see where AI projects might live up to their promise, we dug a little deeper into the available reports to uncover some of the common real-world applications of Artificial Intelligence that business leaders and technology executives actually plan to deploy.
Three Big Planned Applications of AI
In these reports, most organizations planned to deploy AI in one of three categories:
- Administrative Process Automation: PwC found that business executives are primarily looking to deploy AI to “alleviate repetitive, menial tasks, such as paperwork (82%), scheduling (79%), and timesheets (78%).”
- Customer Service Automation: Gartner projects that 85% of customer interactions will be handled without humans by 2020.
- Data Analytics: Looking at actual 2016 spend on AI among 1,600 businesses, InfoSys found that companies invested an average of $6.7 million in AI that year, and the majority of that investment went towards big data automation (65%) and predictive or prescriptive analytics (54%).
Each of these represent an area where business executives are seeking concrete value from AI deployments, and—depending on your context and your stakeholders’ needs—may offer an appropriate area of AI application to bring to your stakeholders in the business. However, while each area offers potential value, we want to give a little special attention to the third area: data analytics.
The data analytics area of AI is of special interest to IT leaders for a few reasons. First, the first two areas—administrative process automation and customer service automation—largely focus on cutting costs, while bringing AI to data analytics focuses uncovering new strategic value to add to the business. Second, data analytics already falls into the wheelhouse of activities IT performs. And third, because even though data analytics is immediately relevant to IT, most IT organizations currently experience a lot of challenges with their data analytics program for reasons that will sabotage any AI implementation if left unaddressed.
The Problem of Data Analytics and AI for IT
It’s no secret that most IT groups struggle to get value out of their existing data analytics systems. As McKinsey found, companies across multiple sectors—from manufacturing to healthcare—have currently captured less than 30% of the potential from their existing data and analytics investments. KPMG found that only 40% of business executives trusted the customer insights their existing analytics delivered.
We’ve written in the past about why this might be. From our experience, IT groups face one key challenge when they attempt to get value out of their analytics: they put together a great system that produces brilliant insights, but because their organization is not set up to take concrete action on those insights, the value of those insights—and of the analytics system that produced them—goes to waste.
The bad news is, while an AI deployment can help to analyze more data for even greater insights, it does not solve this fundamental issue of an organization’s ability to take action on those insights.
The good news is, other people are finally waking to this problem and writing openly about it. As expressed in a recent Harvard Business Review article, titled If Your Company Isn’t Good at Analytics, It’s Not Ready for AI:
“Management teams often assume they can leapfrog best practices for data analytics by going directly to adopting artificial intelligence and other advanced technologies. But companies that rush into sophisticated artificial intelligence before reaching a critical mass of automated processes and structured analytics can end up paralyzed.”
The author then notes that, in order to actually succeed with AI-driven data analytics program, an organization requires two things. The first is a strong existing analytics capability, and the second is the ability to actually act on the insights derived by that analytics capability (emphasis ours):
“By contrast, companies with strong basic analytics — such as sales data and market trends — make breakthroughs in complex and critical areas after layering in artificial intelligence. For example, one telecommunications company we worked with can now predict with 75 times more accuracy whether its customers are about to bolt using machine learning. But the company could only achieve this because it had already automated the processes that made it possible to contact customers quickly and understood their preferences by using more standard analytical techniques.”
What more can we add? Take these points to heart as you consider deploying an AI solution. Ask yourself, “Do I have an existing value-adding analytics program that AI can improve?”, and ask yourself the even harder question, “Do I have the organizational processes in place to actually take action on the game-changing insights I want to derive from this AI deployment?”
Before any AI deployment can live up to its promise, you must be able to answer “Yes” to both questions. Until then, the justification behind many AI deployments may remain thin.
Marc Schiller is the founder and Managing Partner of Rain Partners. Schiller is a leading voice and thinker on IT leadership and management. His book, The 11 Secrets of Highly Influential IT Leaders, broke new ground regarding the most significant management challenges facing IT leaders today—and how to address them.
Contact Marc at firstname.lastname@example.org, or call:1.914.290.4575