Dr. Connor Robertson on Why Some Entrepreneurs Are Using AI Wrong, and What to Do Instead
The most common way business owners use artificial intelligence is also, according to Dr. Connor Robertson, the least powerful one. A person opens a chat window, asks a question, reads the answer, and closes the tab. In Robertson’s view, that is not leverage. It is a slightly faster version of a web search. Real leverage, he argues, means AI is doing productive work inside a business while its owner is doing something else entirely.
Robertson, an entrepreneur and strategic advisor based in Pittsburgh, has built much of his recent thinking around this distinction. The problem with the search-engine approach to AI, as he describes it, is that the business owner remains the bottleneck at every step. They have to initiate the conversation, write the prompt, read the output, and then act on it manually. Nothing is running without them. The tool may be doing more than it once could, but the process still runs through the owner every single time, which means it has not actually reduced their workload. It has only made each individual step faster.
True leverage, in Robertson’s framing, looks different. It means AI is running without someone having to start it each time. He points to examples such as a system that monitors an inbox and flags high-priority messages on its own, one that drafts follow-up emails and queues them for a quick review rather than a full rewrite, or one that produces a weekly report and posts it to a team channel before the owner is even awake. The defining feature, in his view, is that the system is built once and then keeps producing, rather than requiring a fresh request every time.
To explain how far most businesses are from that point, Robertson describes what he calls three stages ofAI adoption. Stage one is manual query and response, the back-and forth chat window most entrepreneurs still rely on. Stage two involves templated prompts and saved workflows that cut down on repetition, letting someone reuse a strong prompt rather than write a new one from scratch each time. Stage three is fully embedded automation, where AI operates as a background process rather than a tool someone consciously opens. According to Robertson, almost all of the compounding value in AI sits in that third stage, yet most business owners never reach it, because stage two already feels like meaningful progress and the effort required to go further can seem hard to justify.
Making that shift, in his account, follows a fairly specific sequence. He suggests starting by identifying the three tasks that consume the most predictable, recurring time in a given week. From there, the next step is documenting exactly how each task is currently done, step by step, since a process cannot be automated if no one has actually written down what it involves. Once that is mapped out, Robertson recommends building a prompt template or an automation trigger that starts the task on its own, without the owner initiating it manually each time. After deploying it, he stresses measuring the quality of what it produces and refining the system until it matches, or exceeds, what was being done by hand. Only then, in his framing, should someone move on to automating the next task.
Underlying all of this, Robertson argues, is a shift in mindset rather than simply a shift in tools. He describes the difference as architectural. Thinking ofAI as something to pick up whenever a need arises keeps a person locked into stage one, no matter how sophisticated their individual prompts become. Thinking of it instead as infrastructure, something designed once and relied on continuously, is what actually opens the door to stage three. One approach demands attention every time it is used. The other, once built, keeps running while its owner sleeps.
Robertson frames this as more than a productivity tip. In his view, the entrepreneurs who treat AI as infrastructure now, rather than as an occasional tool, are the ones positioning themselves for the biggest operational advantage over the next few years. The gap between the two approaches, he suggests, is not really about who has access to betterAI models. Most people are working with similar tools. The gap is about who has taken the
time to build systems that run without them, and who is still, every day, sitting down to start the conversation from scratch.
About the Author
Dr. Connor Robertson is an entrepreneur, author, and strategic advisor based in Pittsburgh. He is the founder of Elixir Consulting Group and host of The Prospecting Show. More about his work is available at drconnorrobertson.com.

