Companies are using AI more than ever — but they're missing the tech's biggest opportunities

What you probably already know: A recent report from Deloitte’s AI Institute polled nearly 3,700 professionals across industries and found that while AI deployment has become widespread across enterprise organizations, the harder work to redesign the systems that use AI has barely begun. Nearly half of companies surveyed have added AI without touching the workflows or roles it sits within. Only 12% have redesigned at scale. For Amanda Rankin, a Formidable member and enterprise transformation executive with nearly 25 years of experience building enterprise transformation practices at some of the largest companies in the world, none of this is surprising. But it is alarming.

Why? “If the foundation’s cracked, then you risk automating a cracked foundation,” Rankin says. The tendency, she says, is to prioritize the pieces that are easiest to see, fund and measure. “But if that’s all you’re looking at, you miss the cross-functional capabilities and organizational conditions that make the whole system work, and that’s where real transformation can happen,” Rankin says. She knows. She watched this pattern play out before.

Rankin spent the better part of a decade building and scaling an enterprise transformation practice for Amazon Web Services, consulting with Fortune 100 companies on cloud adoption during one the most transformative periods in modern technology. The lessons she learned about data integrity, governance and the cost of skipping foundational work were hard won. Now, she says, companies are making the same mistakes — only magnified.

What it means: The Deloitte report suggests the real test is not whether AI speeds up existing steps, but whether it changes what is possible in a workflow. AI can expose cracks that have existed for years, such as poor data quality and governance, fragmented systems, unclear accountability, inconsistent compliance standards and organizational silos. “Most organizations have a lot of process in place, but it’s process for process’ sake and layers and layers of workarounds have happened over many years in most of these instances,” Rankin says. “And so they skip to the automation because they think the problem is the number of steps or the time. But if all you do is automate what’s there, you’re missing an opportunity to leverage what AI was really built and intended to do.”

She points to the mass layoffs sweeping through technology organizations, eliminating the very people who understand why those workarounds exist in the first place. The institutional knowledge of why systems were built the way they were, who they were built for and what they were actually trying to accomplish — all that knowledge walks out the door when experienced employees are let go. What remains is a team tasked with automating processes no one fully understands. With governance, the numbers are equally stark. In Deloitte’s survey, just 12% of respondents reported they were using the most mature state of AI where the tech runs end-to-end and humans audit outcomes rather than having people approve each step. Rankin says closing that gap depends less on policy and more on people who know when something is wrong.

What happens next: Rankin is particularly skeptical about the performative AI metrics she sees currently dominating board conversations. “You can be using it all day long and have worse results,” she says. “Using it and the actual adoption of something are not the same thing.” Her advice for leaders is less about technology and more about intellectual honesty. Stop pretending there’s certainty when that doesn’t exist. Stay focused on outcomes rather than optics and be willing to do the foundational work that so many organizations are glossing over or skipping entirely. Start with the desired business outcome, she says, and work backward to understand the full system required to produce it. “AI gives us an extraordinary opportunity to rethink that system, not simply automate pieces of the one we already have,” Rankin says. “And that requires us to think differently about technology, organizational design, leadership and human experience together.”