When did you last open a generative AI chatbot? Yesterday? Three days ago? If you can't quite remember, you're in exactly the same position as the majority of users worldwide. According to a recent analysis published by technology analyst Benedict Evans, a substantial share of people who use generative AI chatbots open the tool only once or twice a week. That's nothing like the frequency of email or messaging apps we use every day. There is an enormous psychological distance between checking your smartphone's lock screen dozens of times a day and opening a particular app once or twice a week. The former is a reflex. The latter is a decision.
So the question becomes this: if this tool is really "the technology that will change the nature of computing," why do people have to make a decision to use it?
The Crack the Numbers Reveal
Evans's analysis distills the state of generative AI adoption into two questions: Is this a problem of time, or a problem of product?
ChatGPT crossed 100 million monthly active users within two months of launch, making it the fastest-growing consumer service in history. By that number alone, the growth is clearly explosive. But look at the ratio of daily active users to monthly active users — DAU/MAU — and the story changes. The higher that ratio, the more a tool is something people reach for every day, and generative AI chatbots score noticeably lower on this measure than social media or messengers. Many people have created accounts; far fewer open those accounts daily.
Explanations for this gap split into two camps. The first is the optimist's case: "It's still early. Smartphones felt awkward to everyone at first, too. When the iPhone launched in 2007, did people stare at their screens all day from day one?" In this view, technology transitions take years to fully take root, and we're at the beginning of that curve. The second is the product case: "This tool hasn't yet blended naturally into people's daily flow. It works like a specialist instrument you pull out for particular situations." These two interpretations don't exclude each other, but which one you weight more heavily changes your strategy entirely.
An interesting contrast comes from workplace adoption. Tools embedded directly into existing workflows — like Microsoft's Copilot — see far higher usage frequency than standalone chatbots that require opening a separate tab. It's a structure where you don't go to the tool; the tool is already there. This difference is not merely a UX issue. It shows that before a tool can become a habit, the environment has to let your hand reach for it without a "conscious choice."
Why It Hasn't Become a Habit — and Why the Tool Isn't the Only Culprit
Here's something worth pausing on. Separate from the question of how polished the tool is, there are structural reasons on the user's side, too.
It helps to think of competence in three layers: skill, knowledge, and attitude. If the difference between someone who opens generative AI once a week and someone who opens it daily were a matter of skill or knowledge, training or tutorials would solve it. But observation says otherwise. Plenty of people who know ChatGPT inside and out don't use it habitually. Meanwhile, there are people with modest technical skills who use it every day.
What makes the difference is attitude — the fundamental stance you take toward the tool. Daily users treat AI not as "a tool you ask for help" but as "a space to unfold your thinking." They use it to pose questions, rough out drafts, stress-test dubious assumptions, and flesh out ideas. For them, opening a chatbot window is like opening a notepad. No special decision required.
Occasional users, by contrast, treat AI like a specialist you consult "when a special question comes up" — the way you'd call an accountant only at tax time. In that pattern, frequency never rises. Because you only open the tool once questions have piled up, there's no reason for the interval between sessions to shrink.
However much the tool improves, without this shift in attitude the adoption curve climbs for a while and then plateaus. Evans's question — "a problem of time, or a problem of product?" — is actually missing a third option: a problem of user attitude.
Why This Puzzle Matters for Solo Entrepreneurs
A large company can slot Copilot into its existing tools. The IT team deploys it, and workflows get redesigned automatically. Solo entrepreneurs and small teams have no such structure. You have to build the environment yourself.
Which means it's worth taking a cold, honest look at your own usage pattern. How many times did you open an AI tool in the past two weeks? When you did, what prompted it? Did you reach for it when you got stuck — or did you pull it out first, at the start?
The first condition for building a habit is removing friction. Simply pinning the tool to your browser bookmarks or start page changes your perceived frequency. It has to physically sit somewhere always in view for your hand to reach for it "without a decision." Installing the desktop app and leaving it running, or making it the first tab you open when work begins, works too.
The second condition is redefining what it's for. If you're drafting a project right now, try making it a routine to ask the AI, "Where is the weakest part of this idea?" Before writing a client proposal, get in the habit of asking, "Name five things this client would actually worry about." Frequency only rises once you have routines you can run even when no special question exists.
The third — the hardest, and the most important — is lowering your expectations of the output. One reason solo entrepreneurs use AI only occasionally is the judgment that "if this is the quality I get, I'm better off writing it myself." That judgment isn't wrong. But the expectation arises because you're trying to use AI as a final-output generator. Use it as a catalyst for thinking, a counterargument generator, a fast research assistant, and the standard changes. Speed and diversity of perspective start to matter more than polish — and by that standard, AI is useful surprisingly often.
In the end, before you change the tool, you have to change how you approach it. As AI rapidly fills in the domains of skill and knowledge, the attitude you bring to it has begun to create the real competence gap. The difference between the once-a-week user and the daily user may look trivial now, but six months from now, their ways of working will likely have diverged entirely.
Perhaps this is the heart of the puzzle Evans posed. Generative AI's adoption stall is partly because the tool falls short — but also because people haven't yet changed their attitude toward it. While waiting for the product to be finished, the people who finish their attitude first will pull ahead in this transition.
A solo entrepreneur has no IT team to restructure things for them. Changing your attitude — that's the only infrastructure you've got.
