Forget low-hanging fruit; use AI to gather fruit on the ground

Adopting new technology is challenging. It requires matching technology maturity to real problems and overcoming human resistance to change. Even in the best-case scenarios, putting new technologies directly into core business processes has always been risky. When technology touches customers or employees in major value streams, the stakes grow.

McDonald's learned this when it introduced AI to its drive-through order boards . Customers met every glitch with glee, filling TikTok with videos of embarrassing AI failures. Within a few months of the pilot launch at 200 locations, McDonald's shuttered the project. Likewise, hospital giant HCA experienced similar challenges when it rolled out an AI tool to improve nurse scheduling . By many accounts, it produced superior schedules, but nurses felt they had lost control of something that directly impacted their lives and patient care. The scheduling tool quickly became a focal point of other workplace frustrations. After the union organized protests, HCA pulled the plug.

At a recent Digital Strategies Roundtable , CIOs and AI executives cataloged their experiences with AI to date, and we noted that many successes shared a few key characteristics, often side-stepping the pitfalls of annoying customers or employees. First, some of the most successful projects addressed long-tail opportunities that had been ignored. For example, power management giant Eaton developed a configurator that allows it to bid on smaller projects . Reducing the cost and employee time required to develop a customer solution allowed the firm to serve smaller customers that had not previously been serving.

Likewise, another roundtable firm developed an AI negotiation agent to reduce procurement costs. The largest contracts representing the majority of procurement spending had rightly been the focus of procurement specialists. Those relationships were carefully managed and assigned human negotiating teams. While much more numerous, smaller vendors were ignored. They didn’t warrant the cost of management and negotiation. Embedding AI into the procurement system allowed simple negotiation with a legion of small vendors, yielding impressive savings.

Another category of successful applications that roundtable members reported involved reducing errors. This group of success stories was broad, touching many industries in the room. Clear examples were using AI to monitor invoicing, rebates, or contracts with the objective of flagging errors. In the past, the labor time and cost of monitoring these processes outweighed any potential savings. AI opened the possibility to harvest these savings, and turned the tables on the ROI question, achieving a strong return on investment.

Beyond addressing the shifting cost curve, the projects illustrated an important organizational element of success. Rather than simply improving or automating existing processes to achieve new efficiencies and reduce work, these projects addressed new opportunities. They were not simply low-hanging fruit, but more analogous to fruit that had already fallen to rot on the ground and were not being gathered at all. Finding economical ways to gather this fruit yielded significant gains without disrupting core processes, changing roles or threatening the organization with job reductions. This translated into less employee resistance and change management, yielding an even faster return.

Eric Johnson is the Bruce D. Henderson Professor of Strategy, Ralph Owen Dean, Emeritus, at Owen Graduate School of Management, Vanderbilt University.