Instead of spending hours on tasks that can be automated, business owners can design systems that work in the background. This frees up human time for decision-making, strategy, creativity, and customer relationships. Paul Cheek, MIT Sloan senior lecturer and software engineer turned serial founder and author, calls this a move from human time to system time.
However, getting to this point requires more than typing into a general-purpose chatbot for quick answers. It involves understanding the places AI can create real value, areas it should not be used, and how to bring employees along in the process.
Below, Cheek shares how business owners can move past isolated AI tools and toward real transformation.
Rethink the system, not just the tools
Cheek's argument is clear. He believes “AI tools plus AI training does not equal AI transformation.” When employees get tools that help them summarize emails or speed up research, individual productivity is only slightly nudged. It does not move the needle for the whole organization.
Cheek encourages business owners to go further, and consider building systems that quietly work to solve problems, even while they sleep.
The reason, he explains, is that retrofitting new software onto an old workflow will not get a business where it wants to be. The workflow itself has to change. In his book, “No One Works Here,” Cheek compares adding AI to an unchanged organizational structure to fitting a Ferrari engine into a horse-drawn carriage. “It would tear the carriage apart.”
Prioritize the AI opportunities that will have the biggest impact
With so many AI tools and potential applications available, it can quickly become overwhelming. Cheek recommends evaluating opportunities through four big questions:
- What is the business impact?
- What is the risk?
- Is it feasible?
- Should AI be doing this?
The final question is particularly important. Just because AI can perform a task does not necessarily mean it should.
Business owners should consider if using AI aligns with their company’s values and whether customers would be comfortable with the technology taking over a particular interaction. For a small retailer, for example, replacing a familiar employee with automation could undermine the personal relationships that differentiate the business.
Small businesses have a competitive edge
Startups can build quickly, but finding and acquiring customers from scratch can be difficult and expensive. Large companies already have customers, but complex processes, tangled data systems, and compliance requirements can slow them down.
Small businesses sit sweetly in between. They already have customers and established relationships, but often have fewer bureaucratic barriers than larger companies. “Small businesses have the most potential right now to win in the AI era,” Cheek said. “The question is whether or not they will develop the AI literacy amongst their teams.”
That window is not permanent. The businesses that build AI literacy now are the ones positioned to capitalize on their head start, while the ones that stand still risk watching it close as competitors of every size catch up.
Small businesses have the most potential right now to win in the AI era. The question is whether or not they will develop the AI literacy amongst their teams.Paul Cheek, software engineer, entrepreneur, and senior lecturer at MIT Sloan
Develop AI literacy through experimentation
Cheek pushes owners to start implementing something he calls “tinker time.” He advocates for companies to experiment with AI tools in low-stakes situations, ideally outside of core business processes at first. It is also important to test different tools to see where they succeed or fail.
Nevertheless, literacy cannot stop with the owner or a single AI-savvy individual. Cheek says the real value shows up only once an entire team is fluent. He recommends structured training paired with informal "show and tell" sessions, where employees demonstrate what is working, and equally as importantly, what is not.
Expect a slower start before things speed up
Before it saves time, AI usually costs more of it. Cheek has a name for that dip: the "latency J-curve." A task that takes an hour to do manually might take six hours to build into an AI system. That upfront cost feels like a step backward. But once everything is built, that setup becomes a permanent asset, running on its own, and saving time indefinitely.
He also cautions against over-indexing on short-term ROI. A productivity gain in year one may not show up in revenue right away, but the payoff compounds over time.
AI can inform the call. It cannot make it.
Speed is only half the goal. Cheek is just as focused on keeping critical thinking alive as decisions move faster, and he does it two ways.
First, he asks AI to critique decisions he has made, using it to pressure-test his own thinking rather than replace it. Second, he steps away from AI periodically to think deeply about the parts of the business no algorithm can fully grasp — the relationships, the context, and the judgment calls that belong to him alone.
“AI is not going to understand every unique complexity of any individual small business,” he said. AI can surface patterns, flag risks, and accelerate execution. But the call itself — the one that shapes the direction of the business — still has to come from a human who understands what no model ever fully will.
CO— aims to bring you inspiration from leading respected experts. However, before making any business decision, you should consult a professional who can advise you based on your individual situation.
CO—is committed to helping you start, run and grow your small business. Learn more about the benefits of small business membership in the U.S. Chamber of Commerce, here.
Enroll today!
Sign up for Small Business B(AI)sics, our free AI training course, designed for small businesses like yours. Learn the basics of AI in just 30 minutes, plus get resources to help you apply AI skills at your business today.