Mike Murchison is CEO of Ada, an AI-powered platform transforming how brands around the world interact with the people that love them.
Early in my career, I’d sometimes see a dozen printed-out emails stacked on the corner of an executive’s desk. It was a telltale sign that they had a different relationship with technology than I did. But it told me something else, too: that holding too tight onto old patterns of work is the surest way to hamper your productivity.
I don’t see those email stacks today. But I’ve noticed a far greater problem. The same lack of trust in new technology keeps many of us from making the most of our AI tools.
There are whole categories of work that AI can take off our hands if we let it. In fact, a full one-third of our current work lives could be automated by 2030. Yet, we still waste countless hours transcribing meeting notes and composing pro forma emails. And too often, we freeze when there’s a chance to hand real responsibilities over to an AI.
I believe that technical hurdles aren’t what’s holding us back. The true blocker is something much more basic, even primal. It’s our fear of letting go.
Stuck In AI’s First Gear
I see this every day in my field, customer service. The original chatbots were engineered to provide preapproved answers to frequently asked questions. To be clear, this “deterministic” model still has its upsides—especially when it comes to delivering carefully scripted messages.
But the technology has evolved, and the new AI agents aren’t deterministic. They’re probabilistic—just like us. They reason. They learn from experience. As a result, to let these agents do what they do best, we must let go of control and allow them to come up with original answers that we haven’t spoonfed to them.
And that’s the hard part. Many teams are stuck in limbo, afraid to let things rip. So, I see companies take the halfway measure of pairing human agents with AI assistants. These copilots suggest lines of script or point their human counterparts toward a resolution—“whispering in the agent’s ear” as opposed to handling the work itself.
Just like micromanaging doesn’t work well with people, it doesn’t work well with AI, either. We’re essentially stuck in first gear—leveraging some humble gains in productivity but just a fraction of what can be achieved.
The alternative, of course, is to let go. To give AI the autonomy and latitude to independently address customers and provide a solution. To treat AI as another team member to be onboarded, trained and optimized.
This doesn’t just apply to customer service, either. I’m talking about a transformation that’s needed across every part of the customer journey. The companies truly maximizing returns on this technology are getting their AI out of first gear. Here’s how.
Step One: The Mind Shift
I’m convinced that letting go of control begins with a simple shift in focus. Instead of remaining obsessed with inputs (all our busy tasks), we need to refocus on outputs (what we actually deliver).
There are countless examples. For example, instead of watching a full meeting recording to understand the takeaways, have an AI summarize the transcript for you.
Still, we take stubborn pride in those inputs. It’s a holdover from the Industrial Revolution when a worker’s value was measured in hours spent on a dreary assembly line. Today, we’re in the midst of a fundamental shift to an output-oriented culture—one that aligns value with end goals and celebrates human ingenuity more than human labor. AI represents a chance to accelerate progress toward that goal.
Step Two: Find Your Entry Point
Once you’ve broken down mental barriers to letting go, it’s time to tap AI’s full potential. At an individual level, that could be as simple as using ChatGPT to turn notes and bullet points into a first draft of your report in seconds. At a departmental level, certain parts of the company make an obvious entry point for leaning into AI, including marketing, sales and customer service. Look for any area rife with repetitive and administrative tasks.
I’ve had a front-row seat to the myriad ways customer service can be transformed by handing over control to AI agents. A field that’s plagued by highly repetitive work and low employee satisfaction suddenly boasts new possibilities. One company, for example, found their employees were so freed up they could create a new “golden glove” team that brings extra value to their top-tier clients.
Step Three: Recalibrate Your Concept Of Risk
It’s universally accepted—but almost as universally overlooked—that people are fallible. At work, errors resulting from issues like attention loss and fatigue cost companies millions.
We’re often happy to frame human error as learning opportunities. Yet, we’re often less willing to extend the same grace to AI. But allowing for errors and learning from them is essential for getting the most out of AI. These are tools that thrive on feedback and coaching.
And the risk may well be smaller than you think. We’re making rapid progress in reducing damaging hallucinations. Plus, AI platforms and third-party tools provide guardrails that increase accuracy.
The trick is deciding where you can afford that risk. For instance, if someone asks an AI agent a question that concerns their personal finances, an extra level of oversight might be warranted. Once you set up guardrails around riskier subjects, you’re free to explore AI’s potential in other, less dangerous arenas.
Today’s AI can reason. And working with an entity that can reason—whether it’s human or not—means granting it the freedom to make choices, to complete tasks without direct oversight. It means laying down the million little responsibilities we thought we had to carry ourselves.
Those who win the AI moment will take that leap early on. Pretty soon, those who can’t stop themselves from micromanaging their AIs are going to look like the guy who’s drowning in printed-out emails. It’s time to let go.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
