Artificial intelligence has come of age. There is hardly an industry that is not affected, hardly a business function that will be left untouched.
The McKinsey Global Institute estimates that AI has the potential to add between $3.5 and $5.8 trillion in economic value to the global economy through a combination of improvements in productivity, product and customer experience. Some industries will benefit more than others, of course — industries such as retail, logistics and healthcare will claim an above-average share — but the value to be captured will be significant for all.
The uptake of AI thus far is promising. Most companies have started to adopt it. They have appointed Chief Analytics Officers, hired data scientists and engineers and started to go after an initial set of use cases. In a survey of Chief Analytics Officers, FICO found that almost nine out of 10 respondents have started building proofs of concept for AI use cases, and almost seven out of 10 have deployed AI use cases in production. What is more, there is buy-in from the very top: There was unanimous agreement that boards of directors at least somewhat accept the importance of AI. And just over half of the respondents reported having an open line of communication with their boards.
Still, there are challenges to capturing the full potential of AI, specifically when it comes to consistently operationalizing and scaling AI use cases across a company: Only one in four of the Chief Analytics Officers surveyed by FICO had established a standard, enterprise-wise approach for this. Over 40% had difficulties showing a meaningful return on investment for their AI use cases, over 60% struggled with the integration of the new technology with legacy systems and over half reported being held back by ineffective change management.
This feedback from the field renders clear there is work to be done. Getting started and getting your feet wet has to be the first step. But the value of AI will come from operationalizing and scaling the adoption of the technology across a company. Companies need to set up structures and processes to do this, and they need to make it a habit to continue to refine them as they gain experience.
The tremendous economic value of AI, as predicted by the McKinsey Global Institute, will therefore not be evenly distributed. Companies that successfully operationalize and scale AI will capture an unfair share.
Harnessing The Power Of AI At Scale
We had to overcome the same challenges at my own organization. Our journey of AI adoption began about two-and-a-half years ago with the formation of a new Data & Analytics Office. The initial focus back then was on providing high-quality data to the company: Owning the master data as a single point of truth, institutionalizing data governance and making self-service capabilities available to business analysts across the company. Based on this data, the Data & Analytics Office began to develop and provide a collection of business intelligence tools for the finance, sales and post-sales organizations. These tools have become a tremendous success and are now widely used.
Offering access to high-quality data and providing business intelligence services has been — and will continue to be — mission-critical. But we realized that significant additional value would be created by AI use cases that build on top of our data. After all, data by itself does not create value unless it is put to use. Business intelligence puts data to use by informing business processes, but it cannot change business processes. Only AI can change business processes by streamlining or even automating them. It has the potential to more comprehensively and consistently create value compared to data and business intelligence alone. The Data & Analytics Office has hence more recently embarked on a mission to establish an AI center of excellence. While data and business intelligence continue to be core responsibilities, a new one is now to scale the adoption of AI across the company.
Being successful as an AI center of excellence is a different ballgame than being successful with data and business intelligence. The reason is that AI use cases are embedded in — and often materially change — business processes, while data and business intelligence are not. For example, in healthcare, physicians rely on AI-based clinical decision support systems that make diagnostic and treatment recommendations. In retail, AI dictates how many units of each product are stocked at each warehouse to minimize stockout events while reducing inventory-holding costs. And in logistics, AI optimizes the routing of freight based on expected traffic conditions and weather forecasts, improving delivery times, on-time arrival rates and throughput while curbing greenhouse gas emissions.
Similarly, there is a wide array of industry-agnostic AI use cases specific to individual business functions. For example, in sales, AI is increasingly used to score opportunities, effectively determining on which opportunities sellers spend their time. In customer support, AI suggests resolutions for certain tickets, while resolving other tickets without human intervention. In finance and accounting, robotic process automation helps automate repetitive, detail-oriented tasks, increasing both accuracy and productivity.
Changing business processes required new capabilities that our AI center of excellence had to develop. The journey that our stakeholders in the business functions go through when they adopt an AI use case spans from identifying and prioritizing the right opportunities to making build-versus-buy decisions and finding the right partners. It spans from assembling effective cross-functional teams to developing and trialing solutions to embedding these solutions into workflows and training employees on how to use them. Our AI center of excellence had to lead our stakeholders through this journey, end-to-end.
Because this journey is so essential to successfully operationalizing and scaling AI, we gave it a name: the “stakeholder journey” — similar to the customer journey that is top of mind for our colleagues in the go-to-market organization. Our stakeholders, after all, are the internal customers of the AI center of excellence.
In an upcoming article, I will illuminate in greater detail the new capabilities that we incubated for the AI center of excellence. I will show how these capabilities enabled us to become a trusted partner for AI for our stakeholders, and how they created a flywheel that turned success with one AI use case into demand for more across the company.