Five common AI mistakes—and how to avoid them

For executives tasked with making the big bets that translate to shareholder gains, the challenge of monetizing Artificial Intelligence (AI) can feel overwhelming. Faced with a profusion of exciting use cases, leaders need to pick one surefooted path through the forest of options, while also managing the adoption of enterprise platforms for boards and shareholders who want to see short-term results.

But help is at hand. While these are still the early days of the AI Era, it isn’t dawn anymore. There are lessons to be learned from the pioneers and early adopters who have already put transformative technologies such as machine learning, conversion AI and generative AI to work in their organizations. Based on our own experience advising top-tier communications, media, and technology clients worldwide, here are the five most common AI adoption mistakes we’ve seen companies make, and our best advice on how not to make them yourself.

Common mistake #1: Getting carried away.

AI encourages our imaginations to run wild—and that’s a good thing. But creative license cannot come at the expense of solid, old-fashioned business decision-making. We worked with one company—already beset by weighty challenges, from resource limitations to a complex, long-tail operation—who prior to our arrival had devoted valuable time and attention to considering a proposal for, of all things, an AI-generated avatar.

How to course correct: Set targets that can deliver quick results. A different company we partnered with, a prominent telco, sought breakthrough growth in digital revenue. To maximize engagement and conversion rates, we guided the client to adopt a suite of AI tools to personalize customer experience in real time, while providing real-time shopping assist based on purchase propensity. Careful execution of this plan resulted in the telco exceeding its top- and bottom-line targets.

Common mistake #2: Failure to integrate.

Corporate capabilities don’t exist in a vacuum, and AI is no exception. We’ve seen too many companies be proactive in their AI investments, only to then treat these new tools and platforms as “standalone” systems from which they expect value to flow as if by magic. Generally speaking, companies that fail to integrate have been disappointed.

How to course correct: The key to integrated AI adoption is to build capabilities that will one day be at the heart of a company’s future operations, but which also yield immediate results. Maintaining this dual focus is no small feat. One approach we’ve found helpful is to create a future-state “capability map” for the enterprise. The capability map inventories all of a business’s tools, competencies, and abilities that will be impacted by the AI initiative, and addresses what’s needed to reach the desired future state.

The result is a holistic vision for AI, and a step-by-step plan for its implementation. We partnered with one client on a capability-mapping exercise that began by outlining the company’s operational structure and business processes, then reimagined its omnichannel customer-care operations. With the capability map as its guide, the company positioned itself for a unified, integrated approach to AI, and we used the map to build out AI capabilities across all functions: reducing departmental siloes, and democratizing the company’s enterprise-wide knowledge management system.