The Pressure to Use AI Was Everywhere. The How Was Nowhere.
One morning, he got a single Slack notification: “Starting today, AI writing tool accounts are active for the entire team. Please make full use of them to boost your productivity.” There were no usage guidelines, no sense of which tasks to try first. Two months later, he said, he could count on one hand the number of times he had logged into the account. He didn’t know how — or, more precisely, he didn’t even know what would count as a “how.”
The scene maps onto a phenomenon reported by BBC Business: a growing number of companies are urging or requiring employees to use AI while never telling them how to. The tools arrived, but there was no path for folding them into the daily flow of work. The organization wanted speed; the employees lost their bearings.
The Accounts Arrived. The Design Didn’t.
There’s a visible split in how companies roll out AI. Some buy the licenses, issue the accounts, and wrap it up with a memo that reads “AI use encouraged from here on.” Others first design which tasks, aimed at which deliverables, used at what level — and only then bring the tool in. Most of the organizations now in disarray belong to the first group.
This gap isn’t simply a lack of training. It’s less that people don’t know how to operate the tool and more that they don’t know what they’re supposed to do with it. When a marketer is handed an AI copywriting tool and has to decide for herself “where in my current workload does this even fit,” that isn’t a reduction in work — it’s a new cognitive load. Not many people have the spare capacity to make that call in the middle of the workday.
Surveys of UK companies show the same pattern repeating. Many managers at organizations that have adopted AI tools say their teams “aren’t using them properly,” yet only a minority have drawn up formal usage guidelines or a plan to integrate the tools into their workflows. The investment was made; the design that would turn it into real results was skipped.
It’s well established how decisive the first few weeks are when someone new joins an organization. Six months on, there’s a clear divide between those who were guided in a structured way — what work they’d own, who they’d collaborate with and how, by what standards their performance would be judged — and those who weren’t. A new hire given nothing but a desk on day one rarely discovers their own role. Tool adoption is no different. The day the account is created is the day onboarding begins. How you build a relationship with the tool in those first few weeks determines whether it ever makes its way into your work.
How Pressure Breeds Going-Through-the-Motions Use
The word that stands out in the BBC coverage is “confused.” Employees aren’t resisting or refusing; they simply don’t know which way to head. That distinction is no small thing. Plenty of executives blame failed AI adoption on employee resistance, but on the ground the situation looks far more like an absence of direction.
When there’s a goal but no path, people make the safest choice. More often than turning their backs on the AI tool outright, they settle for token use: handing a draft report to the AI and then rewriting it from scratch, submitting the output without review, or trying it out lightly only in areas unrelated to their actual work. Neither leads to the productivity gains the organization was hoping for.
The stronger the pressure for short-term results, the sharper this pattern becomes. When AI use lands on the performance review while there’s no time to learn it on the job, employees prioritize the deadline in front of them over learning the tool. Usage numbers climb, but the actual work doesn’t change.
Here we should also consider the opposing view. AI tools are designed to be far more intuitive than the previous generation of enterprise software, and YouTube and online communities are full of free learning material. Some argue that an overly structured rollout program actually blocks employees’ self-directed exploration and shuts down creative uses the organization never anticipated. There are plenty of cases in which someone who learned by wrestling with the tool came to understand it more deeply than someone who only followed the guidelines. The observation that exploration teaches more than a manual isn’t wrong.
But for that argument to hold, its premises have to be met. There has to be time to explore, a culture that makes mistakes okay, and a structure that lets the fruits of that learning be folded back into the work. Pressure poured out without those three things isn’t self-directed exploration — it’s neglect.
What Korea’s Solo Planners and Middle Managers Should Be Asking Now
In Korea, another layer of pressure sits on top of this problem. The decision to adopt AI comes down from the top of the organization, but responsibility for actually running it falls to the on-the-ground manager or the solo planner — the one-person operator who plans and executes alone. They end up in the position of having to design “how should our team use this” for themselves.
The first thing to check in this situation is whether the team’s purpose for using AI is genuinely concrete. Phrases like “efficiency” or “use encouraged” are not purposes. The aim has to be tied to a measurable outcome — something like “cut the time to draft a proposal to under 30 minutes” or “automate the data-gathering step of the weekly report.” Adoption without a purpose leaves only the license cost and produces no change.
Next, check whether you’ve actually mapped out — together — where in each team member’s workflow the tool can fit. If you’re working solo, you need to experiment directly to find which step of your daily or weekly routine an AI assist genuinely works in. Even without a company-wide rollout, small integration experiments at the individual level can start right now.
Set your tolerance for failure in advance, too. When you use an AI output as-is and the result disappoints, keeping that from hardening into a verdict of “this is useless” starts with adjusting expectations first. The first few attempts have to count as the cost of learning the tool. If you’re a manager, you have to create that room for your team; if you’re a solo planner, you have to grant it to yourself.
The gap between the day the account is issued and the day it’s truly integrated into the work — organizations that designed in advance how to fill it, and those that didn’t, end up standing in completely different places six months later, having paid the very same license fee. How you designed those first few weeks decides whether adoption succeeds or fails. This is not a technology problem. It’s a design problem.
