Her all-hands deck for the award winners misspells the winner’s first name. Her AI-use survey reads like ten paragraphs of AI-generated slop, full of typos and mysterious “levels” of AI adoption that nobody can actually decode.
That is not harmless style. That is a manager creating confusion and asking everyone else to treat it like leadership. If you are the employee stuck underneath that, yes, it is frustrating. You are not wrong to feel it.
Either way, the response should not be panic, gossip, or a passive-aggressive Slack thread with a robot emoji. It should be a sober, defensive way of working in a bad environment.
Do not let the boss turn nonsense into your problem
A normal employee should not have to spend half the week decoding bad spelling, vague AI worship, and requests that sound like they were written during a meeting with the browser tabs open.
If the manager is sloppy, overconfident, and impossible to follow, that is not a personality quirk to admire. It is a management failure. It wastes time, breaks trust, and makes ordinary work harder than it needs to be.
You do not need to solve her writing style. You need to keep the actual work from getting dragged into it.
Translate the request into something real
People who talk like AI experts often mean one of four things:
- they want to sound modern
- they want speed
- they want automation
- they want the problem to disappear without owning the consequences
That may be what she means. It does not make the ask good.
Your job is to turn that into actual requirements. Ask:
- What problem are we solving?
- What changes if we do this?
- What data is involved?
- Who owns the result?
- What is the smallest version that still works?
If the request survives those questions, it may be real. If it turns into a fog machine, it was never a plan. It was an idea wearing a badge.
Use the typo as evidence, not entertainment
The misspelled award winner name is funny for about five seconds. Then it becomes evidence. If the boss cannot keep a name straight in an all-hands presentation, you should assume the same level of care may show up in projects, approvals, and AI requests.
That is not you being petty. That is you noticing a pattern. A person who wants everyone else to trust their AI judgment should not be putting out work that looks careless enough to fail basic review.
The AI-use survey problem
A survey full of slop, typos, and mysterious "levels" of AI use is not just embarrassing. It is a warning sign. If leadership cannot define the thing they are asking the company to measure, then the survey is not a management tool. It is theater.
You end up with questions like:
- What does "level 3 AI usage" mean?
- Does anyone know?
- Did the person who wrote this know?
- Does the data answer anything, or just produce more slides?
That is how organizations mistake documentation for intelligence. They are not the same.
What a frustrated employee should do
-
Answer the business question, not the aura. If the request is fuzzy, ask what decision it supports.
-
Write everything down. Especially when the output is messy. Messy requests need clean records.
-
Keep the conversation concrete. Replace "AI adoption levels" with actual use cases. Replace "innovation" with tasks, owners, and outcomes.
-
Do not imitate the slop. If leadership writes like a confused model, that does not mean you should respond like one.
-
Escalate the cost of the mess. The problem is not just that the deck has typos. The problem is that the boss is wasting time, creating confusion, and making the team absorb the fallout.
The useful rule
The more your boss sounds like she is selling AI to herself, the more human your response should be.
That means: - ask for clarity - force specificity - document decisions - keep the work reviewable - refuse to let confusion become your normal
You should not have to absorb a manager’s slop and call it professionalism. You are allowed to be frustrated. You are also allowed to protect yourself by demanding something legible.
Get in touch if you want help turning vague AI enthusiasm into something operational.