Every course I teach on managing people covers the report who works for you. This one is about the person you report to.
Managing up means communicating with your own manager well enough that you actually get what you need: clarity, air cover, resources, a fair read on your work. Almost nobody gets taught this directly, and most managers are stretched thin enough that a badly framed update or a vague ask just disappears into the noise. I’ve watched genuinely talented people get overlooked for exactly this reason. Their work wasn’t weak. Nobody could tell what mattered in it from the way they described it. AI helps more here than in most places I recommend it: it translates your work into terms your boss actually cares about, tightens a rambling status update into something skimmable in ten seconds, helps you frame a pushback so it reads as a tradeoff instead of a complaint, and preps you for a 1:1 with an actual agenda instead of a vague sense that you should probably say something. Gallup’s 2026 workplace report found manager engagement has fallen to 22%, down from 31% just three years ago, which tells you the person you’re managing up to is very likely stretched thin and short on attention. This guide gives you the prompts I use for this, plus an honest read on what AI can’t do for this particular relationship.
Nobody teaches you how to manage up. The day you get promoted, you get a crash course on running a team, sometimes literally a training module. Managing the relationship with the person who runs you? You’re expected to just absorb that by osmosis. Most people never do. I see it constantly in the workshops I run: someone over-communicates and annoys their boss, someone else goes quiet and gets written off as invisible. A fair number of people manage to do both in the same week without noticing.
This matters more now than it used to. The person on the other end of your update is almost certainly more stretched than they look, and I don’t think most people account for that when they hit send. Gallup’s State of the Global Workplace 2026 report found manager engagement worldwide has fallen to 22%, down from 31% just three years earlier, and lower engagement among managers now accounts for most of the recent slide in overall employee engagement [1]. Your boss isn’t sitting around waiting to give your work full attention. They’re triaging. Your job is to make that triage easy.
Here’s the pattern I didn’t expect when I started teaching this: across the 2,000-plus people who’ve come through a Future Factors workshop, the ones who get promoted fastest are almost never the best communicators in the room overall. They’re the best at this one narrow, specific skill. Everyone else is spending real energy being generally articulate. The people who get ahead figured out, usually by accident, how to be understood by exactly one person under time pressure. That’s a smaller, weirder skill than ‘good communicator,’ and it’s the one this whole guide is actually about.
This isn’t the same conversation as ‘how do I run better 1:1s with my team’ or ‘how do I have a difficult conversation with an employee.’ Those are about managing down, and I’ve written about both elsewhere on the site (see our guide on how to use AI to prepare for a difficult conversation and our practical guide on how to use AI to be a better manager). This is the mirror image: you, communicating up, to someone who has their own boss and their own deadlines squeezing them from above.
Managing up isn’t flattery, and it isn’t playing politics. What it actually does is hand your manager the clarity they need to go to bat for you. Most people give volume instead, and volume is the one thing a stretched manager can’t use.
I call this the translation gap, and once you can see it, you notice it in almost every failed update: what you find interesting about your own work is rarely what your boss needs to hear about it. You might be proud of the elegant workaround you built, or the twelve stakeholders you finally got in the same room. I’ve watched people describe a genuinely strong quarter of work and lose their manager by the second sentence, because they led with process instead of outcome. Your boss wants to know what changed, what it’s worth, and whether they need to do anything. Most people only know how to tell the first story, the one about their own effort, because it’s the story they lived through and the outcome story requires actually stepping outside your own head.
This is genuinely one of the best uses of AI I teach, because translation is exactly the kind of task large language models are built to do well. Paste in your raw notes, your task list, or a messy project recap, and ask the model to rewrite it through the lens of what your specific manager cares about: revenue, risk, timeline, headcount, whatever it is. You already know what your boss reacts to in meetings. Tell the AI that, and let it do the reframing.
Try this prompt: “Here are my raw notes from this project [paste them]. My manager cares most about [timeline risk / budget / customer impact, be specific]. Rewrite this as a 4-sentence summary that leads with the thing they’d care about most, states the current status in one line, and ends with what I need from them, if anything. Cut anything that reads as a status update on my personal effort rather than the outcome.”
One honest caveat, and I say this in every workshop I run: the AI does not know your boss. It only knows what you tell it. Feed it generic instructions and you get a generic translation, which defeats the point. The more specific you are about what your manager actually prioritizes, the better this works, and that specificity has to come from you paying attention, not from the model guessing.
Most status updates fail for the same reason: they’re organized around what the writer did, in chronological order, when the reader needs the headline first. Journalists call this burying the lede, and I use that exact phrase when I’m training a room full of managers on this. In a status update to your boss, it means the one thing they actually need to know is sitting in paragraph four, after three paragraphs of context they didn’t ask for. Here’s the part that surprises people when I point it out: the updates that get ignored are usually the most thorough ones in the batch, not the sloppiest. Thoroughness reads as noise to someone triaging forty messages before lunch. Your boss doesn’t want the full picture. They want the one piece of it that requires them to do something.
The fix is a structure sometimes called BLUF, bottom line up front. State the outcome or the ask in the first sentence, then back it up with just enough detail to support it. AI is very good at restructuring a rambling update into this shape once you tell it the format you want, because reordering and compressing information is a task it handles reliably, even when your original notes are a mess.
Try this prompt: “Turn this list of things I did this week into a status update for my manager. Lead with the single most important update in one sentence. Then three bullets max: what’s on track, what’s at risk, what I need from them. No filler, no restating the obvious, and cut anything that sounds like I’m justifying how busy I was. Here’s my raw list: [paste it].”
This works whether you’re sending an async Slack update, an email before a 1:1, or a weekly report into a shared doc. The format barely changes. What changes is the level of detail your particular boss wants, which is something you learn by watching what questions they ask after your updates and feeding that pattern back into your next prompt.
Asking your boss for something, more headcount, a deadline extension, budget, a title change, is one of the moments people most often botch the framing on. The instinct is to lead with the problem: how hard it’s been, how unfair it feels, how much has piled up. I get that instinct. It’s also a bad strategy. Managers respond better to a clearly framed tradeoff than to a pile of frustration, even a completely justified pile.
AI is useful here specifically because it’s good at reframing a complaint as a decision. Tell it what you actually want, why, and what the tradeoffs are, and ask it to draft the ask as a choice you’re presenting rather than a problem you’re dumping. You still have to supply the honesty and the actual reasoning. What the model contributes is the structure that keeps your ask from reading as either an ultimatum or a vague plea.
Try this prompt: “I need to ask my manager for [the actual ask]. Here’s the context: [why now, what’s at stake, what happens if this doesn’t happen]. Draft this as a short message that states the ask clearly in the first line, gives one or two sentences of context, and offers a tradeoff or option if there is one, like ‘I can do X by Friday if I get Y’ rather than ‘I need more time.’ Keep it under 100 words and don’t soften it into a non-request.”
The same logic applies to pushing back on a decision or saying no to new work. Let’s be honest: most people either cave immediately or push back in a way that sounds like a complaint. A well-framed pushback names the tradeoff explicitly: ‘I can take this on, but X will slip, which one do you want prioritized?’ AI can help you find that sentence fast when you’re too annoyed in the moment to find it yourself.
A lot of professionals show up to their own 1:1 with their manager having given it zero thought beyond ‘I should probably mention the thing with the client.’ I can usually spot it within the first minute of a meeting when I’m coaching someone through this. Then the meeting drifts, the important thing gets forgotten, and you leave feeling like you didn’t get anything out of thirty minutes that were supposed to be yours. This is different from prepping to run a 1:1 with someone on your team. Here, you’re the one who needs the agenda, because nobody is going to build it for you.
Before your next 1:1, dump everything on your mind, half-formed, out of order, into a doc or a chat with AI and ask it to sort the noise into an actual short agenda. This takes two minutes and turns a vague sense of unease into three concrete items you can walk in and cover.
Try this prompt: “Here is everything on my mind before my 1:1 with my manager this week: [dump it all, unsorted]. Sort this into a short agenda: one update they need to hear, one decision or ask I need from them, and one thing I want their read on. Order it by what matters most, and tell me if anything on this list is better handled outside the 1:1 instead of taking up our time together.”
This also solves a subtler problem: it stops you from spending your whole 1:1 reciting status your boss could have read in an email, and frees up the actual face time for the things that need a real conversation, a decision, a gut check, feedback you want to hear in person rather than in text.
At some point your manager, or their manager, will hand you more than fits in the week, and you’ll have to negotiate what actually gets done. The trap here is the same one as with asks: leading with ‘I have too much on my plate’ instead of showing the actual tradeoff. Nobody wants to manage someone who complains about workload on repeat. Showing up with options instead is what actually gets you taken seriously.
AI is useful for modeling out the tradeoff quickly, especially when you’re stressed and can’t see past the pile in front of you. Give it your actual task list with rough time estimates and priorities as you currently understand them, and ask it to lay out two or three sequencing options with what each one costs.
Try this prompt: “Here is everything on my plate right now with rough time estimates: [list it]. My manager just added [the new thing]. Help me lay out two options for how to sequence all of this, and for each one, be explicit about what slips or gets deprioritized. I want to bring my manager a short, clear choice, not a complaint about volume.”
Bring the output to your manager as an actual conversation starter, not a script you read verbatim. AI didn’t decide your priorities here, and it never should. What it did was get you from overwhelmed to organized fast enough that you can have the conversation calmly instead of reactively.
Telling your boss something they’re doing isn’t working is one of the hardest conversations in a professional relationship, and most people are never taught how to do it. That’s not just a feeling. Radical Candor’s 2026 report, based on a survey of 600 US employees conducted between February and March, found that 61% of people say they often stay silent at work when they hold a different opinion than the room, and more than 70% of managers said they rarely or never had a chance to practice giving or receiving feedback before stepping into leadership [2]. That second number is worth sitting with: your manager likely never learned how to take feedback well either, which is exactly why the framing matters so much.
This is a place where AI genuinely helps. It shouldn’t write the feedback for you, since feedback that doesn’t sound like you will land badly no matter how well it’s worded. What it’s actually good for is giving you a low-stakes place to test phrasing before you say it out loud. Draft what you want to say, ask the model to flag anything that reads as accusatory or vague, and rehearse a version that’s specific and about behavior, not character.
Try this prompt: “I need to give my manager feedback about [the specific behavior, not a general complaint]. Here’s a rough draft of what I want to say: [paste it]. Tell me honestly if this sounds accusatory, vague, or like I’m avoiding the actual point. Then help me tighten it into two or three sentences that describe the specific behavior, the impact it had, and what I’d find helpful instead.”
Honestly, this only works if you actually say it. I’ve seen people use AI to draft the perfect version of hard feedback and then never deliver it, because the drafting felt like progress. It isn’t. The draft is a rehearsal, not a substitute for the conversation.
There’s something AI genuinely cannot do for you, no matter how good the prompt is: it cannot read your manager’s face, tone, or mood in a meeting. What it can help with is pattern recognition once you’ve done the noticing yourself. Does your boss want the headline or the full history? Do they prefer a written update before a call, or do they want to talk it through live? Do they push back hardest on numbers or on timelines? Once you’ve noticed the pattern, AI can help you build a short, private profile you refer back to before every interaction.
This sounds almost too simple to be useful, but it changes how your updates land more than any clever phrasing trick I know. A boss who wants the headline first and gets three paragraphs of buildup will tune out before you get to the point. The point usually wasn’t bad. It just arrived in a shape that doesn’t match how they process information.
Try this prompt: “Help me build a short communication profile for my manager based on what I’ve noticed: [describe what you’ve observed, e.g. ‘always asks about budget first,’ ‘prefers bullet points over paragraphs,’ ‘gets impatient with long context before the ask’]. Summarize this into 3 or 4 rules I should follow whenever I write to them, so I can paste it back in before I draft anything important.”
Keep that profile somewhere you’ll actually reuse it, and update it as you learn more. This is the single most useful habit in this entire guide, because it makes every other prompt here sharper the more accurate it gets.
If your manager has read a hundred of your messages and suddenly gets one that sounds like a press release, they’ll notice, even if they can’t say why it feels off. I’ve had people bring me a draft in a workshop that read beautifully and I still told them to rough it up, because it didn’t sound like the person sitting across from me. Use AI to structure and tighten, then read it out loud and put back in the phrases that sound like you. A status update that sounds like nobody wrote it is worse than a rougher one that sounds like a person.
Drafting the perfect message about a problem isn’t the same as sending it. And if the topic is sensitive enough to deserve a real conversation, sending a message isn’t a substitute for that either. AI is excellent at helping you prepare for a hard conversation. It’s a bad substitute for having it.
AI doesn’t know your office politics, your manager’s mood this week, or the history behind why a particular ask is more sensitive than it looks on paper. Use it to draft and structure. Keep the actual call, on tone, timing, and whether to send something at all, with yourself.
Not every Slack message needs a prompt. Save this for updates, asks, and feedback that actually carry weight. If you’re running everything through AI first, you’ll slow yourself down and start sounding stiff in situations that called for a quick, human reply.
The test I use: if I wouldn’t want my manager to know I drafted this with AI first, I’ve either used it wrong or I’m avoiding something I should just say directly.
Sometimes, and it matters less than people think if the content is accurate and sounds like you. What gives it away is over-polished, generic phrasing that doesn’t match how you normally write. Use AI to structure and tighten your message, then edit it back into your own voice before you send it. Treat the AI draft as a first pass, not a final answer.
Any of the mainstream chat tools handle this well, since the core task is language restructuring, not specialized reasoning. Claude and ChatGPT both do a strong job with tone and reframing when you give them specific context about your manager and situation. Use whichever tool you already have a paid or work-approved account for, and keep sensitive details about people or pay out of any tool your company hasn’t cleared for that kind of information.
I get asked this in almost every session I run on the topic. Not if you’re using it to communicate more clearly rather than to manufacture something untrue. There’s a real difference between using AI to structure an honest update and using it to spin a bad outcome into something it isn’t. The tools in this guide are about clarity, not deception, and a manager can generally tell the difference over time anyway.
Use AI to sort your thoughts into an agenda, not to write out exact sentences you’ll recite. Walk in with three bullet points, not a script. The goal is knowing what you want to cover and why it matters, which frees you up to actually listen and respond naturally instead of reading from a mental page.
AI can help you adapt your side of the communication, which is often the only side you can control. Build a profile of how your manager processes information and adjust your delivery to fit it, even if their side of the exchange stays inconsistent. It won’t fix a genuinely dysfunctional working relationship, and if that’s what you’re dealing with, that’s a bigger conversation than a better prompt can solve.
This guide draws on Gallup’s State of the Global Workplace 2026 report, Radical Candor’s May 2026 report on workplace psychological safety based on a survey of 600 US employees, and Slingshot’s January 2026 Digital Work Trends Report. I checked every figure against the original source myself this week, the same way I check any number before it goes in front of the people I train. Sources are linked below.
Weekly tips. Real examples. Practical help for busy professionals.
We care about your data, check out our privacy policy.