{"course":{"slug":"ai-careers","title":"Working in AI","url":"https://addaly.com/learn/ai-careers","level":"beginner","track":"work"},"lesson":{"slug":"promotion-on-evidence","title":"Getting promoted, and the case somebody has to make","url":"https://addaly.com/learn/ai-careers/promotion-on-evidence","position":85,"of":85,"minutes":9,"answer":"Promotion is argued for you in a room you are not in, using evidence you supplied.","keyIdea":"Promotion is argued for you in a room you are not in, using evidence you supplied, and it is granted for work already being done at the next level rather than as a reward for the current one."},"body":{"format":"markdown","text":"## The room you are not in\n\nAt most companies of any size, promotion is not decided by your manager. It is decided in a **calibration meeting** where several managers argue for their people against a limited number of slots, using a written case each has prepared.\n\nYou are not there. Your manager is, holding whatever they can remember or find. Everything below follows from that single fact: your job is to make their document easy to write and hard to argue with.\n\n## The thing almost everybody gets wrong\n\nPromotion is generally granted for **work you are already doing at the next level**, not as a reward for doing your current level well.\n\nThis is the most commonly misunderstood mechanism in a career, and it explains the frustration of the person who has been excellent for three years and never moved. They have been excellent at their level. The case being made in the room has to say \"she is already operating as a senior engineer and we should recognise it\", and if the evidence only shows outstanding mid-level work, there is nothing to say.\n\nThe practical consequence is a sequence, not a request:\n\n1. **Find out what the next level actually means here.** Many companies publish a levelling guide; ask for it. Where none exists, ask your manager to describe the difference in behaviours, and write down their answer.\n2. **Do that work visibly for two or three quarters.** Take the ambiguous problem, run the project across teams, review other people's work, be the person who wrote the plan.\n3. **Then ask**, with the evidence assembled.\n\n## The packet\n\nKeep a running document from your first month. One dated entry per meaningful thing:\n\n- What the situation was and what you did.\n- What changed, with a number where one exists.\n- Who benefited and who said so: a quoted sentence from a stakeholder is worth a paragraph of your own description.\n- A link to the artefact: the pull request, the document, the dashboard, the incident review.\n\nTen minutes a fortnight. At review time you hand your manager a page and they write the case from it in twenty minutes instead of reconstructing your year from memory, which is what they would otherwise do and which systematically favours whoever was most recently visible.\n\n## Ask the two questions that get real answers\n\n**\"What would I need to be doing consistently for you to put me forward, and what would the case say?\"** This forces a specific answer. \"Keep doing what you are doing\" is not one, and it is worth saying so politely: *\"Could you name one or two things a senior person here does that I am not doing yet?\"*\n\n**\"When are promotion decisions made, and when should my case be ready?\"** Cycles usually close one to three months before anything is announced. Making your case a fortnight after the calibration meeting is a year lost to timing.\n\n## Visibility is not self-promotion\n\nMany people, particularly those raised to believe that good work is noticed, find this section uncomfortable. It is worth separating two things: boasting, which nobody likes, and **making work legible**, which is part of the job.\n\nLegibility looks like this:\n\n- A short weekly or fortnightly written update to your manager: shipped, in progress, blocked. Three bullets.\n- Presenting your own work at the team demo rather than letting somebody summarise it.\n- Writing the document instead of leaving it to whoever volunteers.\n- Naming other people's contributions explicitly, which costs you nothing, is honest, and makes you the person others want credit from.\n\nThe credit problem is structural: work done by people who do not talk about it gets attributed to people who do, and this falls hardest on those least comfortable with claiming it. The remedy is not to become louder. It is to write things down, with dates and names, so that the record does the claiming for you.\n\n## When it does not happen\n\nAfter two cycles of \"next time\", ask directly and specifically: *\"Is the constraint my performance, the level's requirements, or the budget?\"* A good manager answers honestly, and the three answers imply completely different actions.\n\nIf it is performance, you now have something to work on. If it is the level's requirements, ask what evidence is missing. If it is budget or headcount (which it very often is) then no amount of further excellence changes it, and the honest options are to wait deliberately with an agreed date, or to move, using the arithmetic from the lesson on how pay moves.\n\n## The limitation\n\nPromotion is partly a function of company growth. A company that is not adding scope has few senior slots, and they open when somebody leaves. In a flat or shrinking organisation, outstanding work and a perfect packet can both be true and produce nothing for two years.\n\nThat is not a reason to stop building the evidence (it travels with you and makes the case at the next employer) but it is a reason to distinguish between a problem you can solve and a queue you are standing in. Waiting for a slot is a legitimate choice. Waiting for one without knowing that is what you are doing is not a choice at all."},"check":{"question":"An engineer has had three years of strong performance reviews at mid level and has been passed over twice. Their manager says \"keep doing what you are doing\". What is the most useful reading?","options":[{"text":"The manager is not advocating for them and they should escalate to skip-level management.","correct":false,"tempting":"Assumes a failure of advocacy before establishing whether there is a case to advocate with."},{"text":"Consistent strong reviews at the current level are evidence of doing the current job well, which is not what the promotion case has to claim, so the next step is to establish what next-level scope looks like here and do it visibly before the next cycle.","correct":true},{"text":"Three years without promotion means the company undervalues them and they should leave immediately.","correct":false,"tempting":"May end up being right, but acts before asking the one question (performance, requirements or budget) that distinguishes the three very different causes."},{"text":"They should ask for a pay rise instead, since the level is clearly blocked.","correct":false,"tempting":"Substitutes the smaller mechanism for the larger one, and a merit-pool raise does not address being at a level ceiling."}]},"prev":{"slug":"writing-for-people-who-decide","title":"The one-pager that gets a decision","url":"https://addaly.com/learn/ai-careers/writing-for-people-who-decide"},"next":null,"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'Getting promoted, and the case somebody has to make', in Working in AI. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-careers/promotion-on-evidence (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Getting promoted, and the case somebody has to make","courseTitle":"Working in AI","module":{"index":9,"slug":"staying","title":"Staying: the first ninety days and the ten years after"},"answer":null,"keyIdea":"Promotion is argued for you in a room you are not in, using evidence you supplied, and it is granted for work already being done at the next level rather than as a reward for the current one.","hero":{"kind":"plate","image":null,"alt":null,"caption":null},"blocks":[{"t":"h2","id":"the-room-you-are-not-in","runs":[{"text":"The room you are not in"}]},{"t":"p","runs":[{"text":"At most companies of any size, promotion is not decided by your manager. It is decided in a "},{"text":"calibration meeting","b":true},{"text":" where several managers argue for their people against a limited number of slots, using a written case each has prepared."}]},{"t":"p","runs":[{"text":"You are not there. Your manager is, holding whatever they can remember or find. Everything below follows from that single fact: your job is to make their document easy to write and hard to argue with."}]},{"t":"h2","id":"the-thing-almost-everybody-gets-wrong","runs":[{"text":"The thing almost everybody gets wrong"}]},{"t":"p","runs":[{"text":"Promotion is generally granted for "},{"text":"work you are already doing at the next level","b":true},{"text":", not as a reward for doing your current level well."}]},{"t":"p","runs":[{"text":"This is the most commonly misunderstood mechanism in a career, and it explains the frustration of the person who has been excellent for three years and never moved. They have been excellent at their level. The case being made in the room has to say “she is already operating as a senior engineer and we should recognise it”, and if the evidence only shows outstanding mid-level work, there is nothing to say."}]},{"t":"p","runs":[{"text":"The practical consequence is a sequence, not a request:"}]},{"t":"ol","start":1,"items":[[{"text":"Find out what the next level actually means here.","b":true},{"text":" Many companies publish a levelling guide; ask for it. Where none exists, ask your manager to describe the difference in behaviours, and write down their answer."}],[{"text":"Do that work visibly for two or three quarters.","b":true},{"text":" Take the ambiguous problem, run the project across teams, review other people’s work, be the person who wrote the plan."}],[{"text":"Then ask","b":true},{"text":", with the evidence assembled."}]]},{"t":"h2","id":"the-packet","runs":[{"text":"The packet"}]},{"t":"p","runs":[{"text":"Keep a running document from your first month. One dated entry per meaningful thing:"}]},{"t":"ul","items":[[{"text":"What the situation was and what you did."}],[{"text":"What changed, with a number where one exists."}],[{"text":"Who benefited and who said so: a quoted sentence from a stakeholder is worth a paragraph of your own description."}],[{"text":"A link to the artefact: the pull request, the document, the dashboard, the incident review."}]]},{"t":"p","runs":[{"text":"Ten minutes a fortnight. At review time you hand your manager a page and they write the case from it in twenty minutes instead of reconstructing your year from memory, which is what they would otherwise do and which systematically favours whoever was most recently visible."}]},{"t":"h2","id":"ask-the-two-questions-that-get-real-answers","runs":[{"text":"Ask the two questions that get real answers"}]},{"t":"p","runs":[{"text":"”What would I need to be doing consistently for you to put me forward, and what would the case say?”","b":true},{"text":" This forces a specific answer. “Keep doing what you are doing” is not one, and it is worth saying so politely: "},{"text":"”Could you name one or two things a senior person here does that I am not doing yet?”","i":true}]},{"t":"p","runs":[{"text":"”When are promotion decisions made, and when should my case be ready?”","b":true},{"text":" Cycles usually close one to three months before anything is announced. Making your case a fortnight after the calibration meeting is a year lost to timing."}]},{"t":"h2","id":"visibility-is-not-self-promotion","runs":[{"text":"Visibility is not self-promotion"}]},{"t":"p","runs":[{"text":"Many people, particularly those raised to believe that good work is noticed, find this section uncomfortable. It is worth separating two things: boasting, which nobody likes, and "},{"text":"making work legible","b":true},{"text":", which is part of the job."}]},{"t":"p","runs":[{"text":"Legibility looks like this:"}]},{"t":"ul","items":[[{"text":"A short weekly or fortnightly written update to your manager: shipped, in progress, blocked. Three bullets."}],[{"text":"Presenting your own work at the team demo rather than letting somebody summarise it."}],[{"text":"Writing the document instead of leaving it to whoever volunteers."}],[{"text":"Naming other people’s contributions explicitly, which costs you nothing, is honest, and makes you the person others want credit from."}]]},{"t":"p","runs":[{"text":"The credit problem is structural: work done by people who do not talk about it gets attributed to people who do, and this falls hardest on those least comfortable with claiming it. The remedy is not to become louder. It is to write things down, with dates and names, so that the record does the claiming for you."}]},{"t":"h2","id":"when-it-does-not-happen","runs":[{"text":"When it does not happen"}]},{"t":"p","runs":[{"text":"After two cycles of “next time”, ask directly and specifically: "},{"text":"”Is the constraint my performance, the level’s requirements, or the budget?”","i":true},{"text":" A good manager answers honestly, and the three answers imply completely different actions."}]},{"t":"p","runs":[{"text":"If it is performance, you now have something to work on. If it is the level’s requirements, ask what evidence is missing. If it is budget or headcount (which it very often is) then no amount of further excellence changes it, and the honest options are to wait deliberately with an agreed date, or to move, using the arithmetic from the lesson on how pay moves."}]},{"t":"h2","id":"the-limitation","runs":[{"text":"The limitation"}],"role":"limit"},{"t":"p","runs":[{"text":"Promotion is partly a function of company growth. A company that is not adding scope has few senior slots, and they open when somebody leaves. In a flat or shrinking organisation, outstanding work and a perfect packet can both be true and produce nothing for two years."}]},{"t":"p","runs":[{"text":"That is not a reason to stop building the evidence (it travels with you and makes the case at the next employer) but it is a reason to distinguish between a problem you can solve and a queue you are standing in. Waiting for a slot is a legitimate choice. Waiting for one without knowing that is what you are doing is not a choice at all."}]}],"check":{"question":[{"text":"An engineer has had three years of strong performance reviews at mid level and has been passed over twice. Their manager says “keep doing what you are doing”. What is the most useful reading?"}],"options":[{"text":[{"text":"The manager is not advocating for them and they should escalate to skip-level management."}],"correct":false,"tempting":[{"text":"Assumes a failure of advocacy before establishing whether there is a case to advocate with."}]},{"text":[{"text":"Consistent strong reviews at the current level are evidence of doing the current job well, which is not what the promotion case has to claim, so the next step is to establish what next-level scope looks like here and do it visibly before the next cycle."}],"correct":true,"tempting":null},{"text":[{"text":"Three years without promotion means the company undervalues them and they should leave immediately."}],"correct":false,"tempting":[{"text":"May end up being right, but acts before asking the one question (performance, requirements or budget) that distinguishes the three very different causes."}]},{"text":[{"text":"They should ask for a pay rise instead, since the level is clearly blocked."}],"correct":false,"tempting":[{"text":"Substitutes the smaller mechanism for the larger one, and a merit-pool raise does not address being at a level ceiling."}]}],"order":[2,1,0,3],"why":null},"prev":{"slug":"writing-for-people-who-decide","title":"The one-pager that gets a decision","minutes":9},"next":null}}