Most career advice tells you to do great work, keep your head down, and trust that the system will reward merit. This is the version of the world taught to people who do not move up, by people who do not understand why they themselves did not move up. Great work is necessary. Great work is also nowhere near sufficient. The promotion goes to the person who did great work and was seen doing it by the people who matter, who built the alliances that protected the work, who picked battles strategically, who knew which rooms to enter and which to skip, who understood how the actual decision got made versus how it was supposed to get made.
The standard label for all of this is office politics, and the label carries a moral charge that gets in the way of using the skill. Engaging with politics is treated as the opposite of being a real professional, when in fact it is part of being a real professional. The honest reframe: politics is the structured study of how decisions actually get made. Choosing not to engage with it means accepting whatever outcomes the people who do engage decide for you.
This is where structured AI rehearsal earns its keep in a way that most career content cannot. Not because AI replaces the relationships. It cannot. Not because AI tells you who to befriend or who to cut. It will not. What AI does is decode the dynamics, surface the alliances you have not yet mapped, predict the move the room is about to make, draft the message that lands without becoming the thing you regret in two months. It is the cheapest political advisor you will ever access, available at 11pm on a Sunday before a Monday meeting where the actual decision will be made before the agenda is opened.
The POWER Framework is fifty prompts for the political work that determines whether your great work compounds or stalls. Five pillars across the full arc of navigating organisational dynamics: positioning yourself in the structure, observing the room, weaving coalitions, executing your moves, and recovering from the missteps that everyone eventually makes.
This guide is LLM agnostic. The prompts work in ChatGPT, Claude, Gemini, and any approved enterprise model. Anonymise identifying details before pasting sensitive context. Specificity in your inputs is the difference between generic political advice and a real plan for your specific environment.
What is the POWER Framework
A method for using AI as a structured thinking partner across the five categories of work that determine whether you navigate your organisation skilfully or get navigated by it. Each pillar maps to one phase of political work.
- P: Position. Map where you sit in the dynamics, formal and informal. Identify your sponsors, allies, sceptics, and blockers. Audit your political capital and what it is being spent on right now.
- O: Observe. Read the room. Decode meetings. Surface unwritten rules. Map coalitions. Catch the second meaning underneath the words people use.
- W: Weave. Build the relationships that compound. Cultivate sponsors before you need them. Identify whose trajectory aligns with yours. Repair what is broken before it becomes structural.
- E: Execute. Time your moves. Pick your battles. Use leverage without burning it. Know when to act with permission and when to act first and ask later.
- R: Recover. When you misstep, audit honestly, rebuild trust, manage the narrative without collapsing into apology. Re enter the rooms you got excluded from.
Five pillars. The POWER Framework runs in rough sequence but loops in practice. By Execute you are almost always back in Observe mode for new signals. By Recover you may need to start over in Position because the missed move changed where you actually sit. The point is structure: when the dynamics get murky, you can find the pillar that fits and run the right prompt rather than guessing.
How to use this guide
Three principles before you run any of the prompts below.
First, the framework is descriptive, not prescriptive. The prompts surface what is happening and predict what is likely to happen. They do not tell you what to do with that information. The choices about who to ally with, which battles to pick, and how to use the intelligence remain yours. The framework gives you better inputs to those choices, not the choices themselves.
Second, the model can only help with what you give it. Paste actual names (or consistent pseudonyms if data sensitivity matters), actual quotes from meetings, actual email threads, actual recent decisions. Generic input produces generic readings. The model is reading patterns. Patterns require real signal.
Third, the writing the model produces is a draft of your thinking, not a guide to action. Anything political the model suggests should pass your own ethical filter before you act on it. Where the prompt says [paste context], be specific. Where it suggests a move, ask yourself whether you would defend it in public if it became visible.
P: Position
P1: The Position Audit
Below is my role, my reporting line, my recent track record, and what
I know about how I am perceived in the organisation. Audit my
position. Map my formal authority, my informal influence, my visible
sponsors, my known sceptics. Tell me where my position is stronger
than my title implies, where it is weaker, and the one move this
quarter that would most shift the weak side. Context: [paste]
P2: The Sponsor Map
Below are the senior people in my organisation who have, at some
point, advocated for me, attended my meetings, opened doors, or
publicly backed my work. For each, score the strength of the current
sponsorship, the cost to them of continuing it, and what they need
from me to keep it active. End with the sponsor I am underweighting.
List: [paste]
P3: The Sceptic Identifier
Below is the broader stakeholder set around my work. Identify who is
sceptical of me, my function, or my approach, based on signals like
meeting behaviour, email cadence, public comments, and absence from
my work. For each sceptic, tell me whether the scepticism is about me
specifically, my role, or my function. Recommend the one to engage
this month. Context: [paste]
P4: The Political Capital Audit
List the political capital I have spent in the last 90 days: every
favour asked, every controversial position taken, every senior
person's attention I used. For each, tell me what I bought with it
and whether the trade was favourable. End with the verdict on whether
I am compounding capital or burning it, and the one behavioural
change for the next 90 days. Capital log: [paste]
P5: The Slot Reader
Based on the conversations, meeting invites, project assignments,
and informal references in my recent context, the organisation is
slotting me into a particular archetype (rising star, safe pair of
hands, future executive, technical specialist, troublemaker, etc).
Tell me which slot, the evidence for that read, whether the slot
serves my actual ambitions, and the moves that would shift the
classification. Context: [paste]
P6: The Unfair Advantage Locator
Given my background, my network, my known reputation, and what I
have observed about the political landscape of this organisation,
locate my unfair advantage in this specific environment. Where do I
read situations more accurately than my peers, who do I have access
to that others do not, what am I uniquely credible on. End with the
play that uses the advantage that no one else in my seat would run.
P7: The Decision Map for My Work
On the next significant decision affecting my work (paste decision),
who actually decides, who has formal authority that says they decide,
who can quietly veto, and who needs to feel included even if they
have no real vote? Map the decision graph. Tell me where my reading
diverges from the org chart and what the divergence implies for how
I work the decision.
P8: The Visibility Audit
Visibility is non linear in organisations. Audit mine. Based on my
recent calendar, my email distribution patterns, the meetings I am
invited to versus those I am not, and the work being credited to me
publicly, score my visibility to (a) my manager, (b) two levels up,
(c) cross functional peers, (d) the board or executive sponsors.
Recommend the one move this month to most increase the lowest score.
Inputs: [paste]
P9: The Threat Assessment
Based on the recent shifts in my organisation (reorgs, new hires,
strategic pivots, departed allies), tell me where my position is
most exposed in the next two quarters. Who might quietly route
around me. What initiative I lead might get reassigned. Which
relationship has degraded that I have not addressed. End with the
specific defensive move this month. Context: [paste]
P10: The Reposition Move
Given the audit above, what is the single repositioning move that
would most materially improve my standing in the next six months
without making me look like I am politicking? Be specific. Tell me
the action, the rationale, who would notice, and what could go
wrong. Recommend the version that is defensible if it became
visible. Context: [paste]
O: Observe
O1: The Meeting Decoder
Below are my notes or transcript from a recent meeting. Decode the
political dynamics underneath the content. Who spoke and who stayed
silent, in what order, who deferred to whom, who pushed back without
seeming to, where alliances showed and where tensions surfaced. End
with the three things that happened in the meeting that nobody named
out loud. Notes: [paste]
O2: The Silence Reader
Below are recent moments where someone I expected to speak up did
not, or someone I expected to engage in a thread did not. For each
silence, generate the three most likely interpretations: tactical
withholding, genuine indifference, signalling disagreement, waiting
for more information. Tell me which is most likely for each, and the
move that would distinguish them. Context: [paste]
O3: The Unwritten Rules Surfacer
Below is what I have observed about how meetings get scheduled,
disagreement gets expressed, credit gets allocated, and decisions
get communicated in this organisation. Surface the unwritten rules.
For each, score it as harmless local custom, load bearing norm, or
indicator of deeper dynamic. End with the one rule a newcomer would
not see that I should respect for the next 90 days. Observations:
[paste]
O4: The Coalition Mapper
Based on the meeting dynamics, lunch patterns, project assignments,
hiring decisions, and informal references I have observed, map the
coalitions in my division or company. Who hires people that look like
whom, whose names come up together, who covers for whom in difficult
moments. Tell me which coalition I am being slotted into and whether
that slot serves my ambitions. Inputs: [paste]
O5: The Rising Versus Declining Read
Based on the recent signals (project assignments, public mentions,
budget allocations, hiring decisions in their areas, board access),
identify who in the broader organisation is rising and who is
declining over the next two quarters. For each top three rising and
top three declining, tell me what evidence supports the read and the
implication for my own positioning. Inputs: [paste]
O6: The Subtext Translator
Below is the recent statement, email, or comment from [person and
their role]. Translate the surface words into the underlying signal.
What are they actually saying, what are they signalling without
saying it directly, what would they have said if they could speak
freely. End with the most likely thing they want from this exchange
that they did not name. Statement: [paste]
O7: The Information Asymmetry Audit
Map the information flow around my work. Who knows things about my
function, my projects, or my standing that I do not know. Who knows
things that I do that nobody else knows. Where am I dependent on
filtered intelligence, and where do I have leverage from holding
unique signal. End with the one source of intelligence I should
cultivate this quarter. Context: [paste]
O8: The Departure Pattern Reader
Below are the people who have left, been pushed out, or quietly
demoted in the organisation over the last 12 months. For each,
generate the most likely real reason (versus the stated reason) based
on context. Identify the pattern: what kind of person, doing what
kind of work, supporting whom, gets removed from this place. End with
the specific behaviour the pattern suggests I should avoid. List:
[paste]
O9: The Sponsor Behaviour Pattern
Below is recent behaviour from my primary sponsor or executive
backer: when they have engaged with my work, when they have stayed
silent, what they have promoted me on publicly, what they have not.
Audit the pattern. Is their support strengthening, holding, or
quietly eroding? Tell me what evidence supports the read, and the
move this month that would test it without seeming insecure. Pattern:
[paste]
O10: The Hidden Agenda Surfacer
For [paste upcoming meeting, initiative, or decision], list the
agendas that are not on the printed agenda. For each stakeholder
involved, what they publicly want from this meeting, what they
actually want, and where those diverge. End with the one agenda I
should accommodate without naming and the one I should actively
counter. Context: [paste]
W: Weave
W1: The Sponsor Cultivation Plan
For my top three sponsors (paste names and roles), draft the
cultivation plan for the next quarter. For each: the cadence of
contact, the form (formal updates, casual check ins, public
backing), the value I provide them that makes the relationship two
sided, and the moment in the next 90 days where I will most need
their air cover. End with the one sponsor I am undervaluing.
W2: The Cross Functional Ally Builder
Across the functions I work with (paste functions), identify the
person in each who is most worth cultivating as a long term ally
based on tenure, trajectory, working style, and stated priorities.
For each, tell me the natural collaboration that would build the
relationship without it feeling forced, and the first concrete move
to make it happen. Inputs: [paste]
W3: The Weak Tie Audit
Map my weak ties: people I know lightly across the organisation who
could become significant if I invested. For each, score on potential
value, current relationship strength, and the cost to upgrade the
relationship. End with the three weak ties most worth a quarterly
check in and the framing for the first message that does not feel
transactional. Context: [paste]
W4: The Relationship Repair Drafter
The relationship with [paste person] has degraded over [paste
context]. Audit what likely went wrong from their perspective. Draft
the move that opens a repair conversation: the message, the
in person ask, or the visible action that signals a reset. For each
option, predict their response. End with the version that fits this
specific person and history.
W5: The Coalition Joining Strategy
Based on the political map, there is a coalition I would benefit
from joining (paste). Tell me how the coalition forms in practice,
the credible entry move for someone in my role, who would sponsor
my entry, and the price of admission. End with whether the cost is
worth the access and the timing of the move.
W6: The Coalition Avoidance Strategy
There is a coalition that is currently slotting me in by association
that I should distance from (paste). Tell me how I distance without
making it personal, the cost of distancing visibly versus quietly,
and the specific actions over the next 60 days that would shift the
perception. End with the version that is defensible if questioned.
W7: The Manager Of Manager Strategy
My manager's manager is [paste]. Currently they see me as [paste
likely read]. I want them to see me as [paste target read]. Draft
the campaign to shift the perception over the next two quarters
without going around my manager in a way that creates a problem.
Specific moves, specific frequency, specific evidence to surface.
End with the visible behaviour that would unwind the campaign.
W8: The Mentor Refresh Move
My current set of mentors no longer fits the next stage of my
career. Identify the kind of mentor I need now based on my actual
trajectory and the gaps in my own judgement. Tell me where to find
them (former bosses, board members, external operators, industry
peers), the right ask for an initial conversation, and what would
make me a worthwhile mentee to them. Context: [paste]
W9: The Influence Without Authority Move
I need [paste outcome] from [paste person or function] over whom I
have no formal authority. Map the influence levers available:
shared interest, reciprocity, social proof, scarcity, expert status,
liking, commitment. For each, tell me whether it credibly applies
here and the specific move that activates it. End with the recommended
combination of two levers and the order.
W10: The Network Maintenance Audit
Audit my network maintenance. Who in my professional network have I
not spoken to in over 12 months who is worth a check in. Who am I
checking in with too often (signal of dependence). Who has reached
out to me that I have not yet returned. Recommend the next ten
outreaches in priority order and the message stem for the top three.
Network log: [paste]
E: Execute
E1: The Timing Decoder
I am about to make a move on [paste: raise ask, scope change, new
initiative, hire request, departure announcement]. Tell me the right
timing window based on the organisational cycle (quarter end, board
meeting cadence, leadership offsite, fiscal year). For each window,
predict the receptivity and the cost of waiting. End with the
recommended window and the trigger that activates the move.
E2: The Battle Picker
Below is the list of fights I could pick this quarter (process
changes I disagree with, decisions I want to revisit, behaviour I
want to call out, scope I want to claim). For each, score on
importance to my actual goals, probability of winning, cost of
losing, and the relationship cost of fighting. Recommend the two
fights to actively pick and the four to deliberately drop. Context:
[paste]
E3: The Momentum Read
Based on the recent signals in my organisation (announcements,
hiring patterns, budget changes, leadership statements), tell me
whether the organisation is currently in a phase of opening (open to
new initiatives, willing to take risks, allocating capital) or
closing (preserving capital, consolidating, risk averse). End with
the specific move that fits the current phase and the move that
should wait. Inputs: [paste]
E4: The Permission Decoder
I am about to do [paste move]. Tell me whether this requires
explicit permission, implicit permission, or no permission at all.
For the permission scenarios, who is the lowest credible person to
ask, what is the right framing of the ask, and what is the cost of
asking versus the cost of acting first. End with the recommended
approach and the message.
E5: The Leverage Use Audit
I have leverage from [paste: alternative offer, board sponsorship,
critical project, scarce skill, key relationship]. Audit how to use
it. For each leverage source, what is the move that compounds it,
what move spends it, what move burns it entirely. Recommend how to
use the leverage on the specific situation in [paste context] without
exhausting it.
E6: The Announcement Sequencing
I need to communicate [paste announcement] across multiple
stakeholders. Sequence the conversations: who hears it first
(privately), who second, who in the broadcast. For each tier, tell
me the framing, the questions to be ready for, and the cost of
getting the sequence wrong. End with the timeline that minimises
political damage.
E7: The Pre Meeting Lobby
Before [paste upcoming meeting], who needs to be pre lobbied so the
meeting itself becomes a confirmation rather than a debate. For
each, the framing of the pre lobby conversation, the concession or
information I should bring, and the signal that the pre lobby
worked. End with the meeting move that follows the lobby. Context:
[paste]
E8: The Quiet Move
Sometimes the right move is invisible. Generate the version of
[paste objective] that I could execute without anyone noticing for
30 days. For each phase, the specific actions, what would tip people
off, what would be lost if the move became visible early. End with
the moment to make the move public and the framing for that
announcement.
E9: The Public Move
The opposite. Sometimes the right move requires being seen doing it.
Generate the version of [paste objective] designed for maximum
visibility to the right audience. The setting, the framing, the
witnesses, the documentation that survives the moment. End with the
specific risk of the public version and the contingency if it
backfires.
E10: The Strategic Pause
I am tempted to make a move on [paste]. Audit whether the right
choice is to pause for [paste timeframe] instead. What would change
between now and then that improves my position. What is the cost of
moving early. What is the cost of moving late. End with the verdict
and the trigger that would convert the pause into action.
R: Recover
R1: The Misstep Audit
I made a misstep on [paste situation]. Audit it honestly. What
specifically went wrong from each stakeholder's perspective. What
damage is real versus imagined. Who knows, who suspects, who is
unaware. End with the verdict on whether to address it directly,
let it dissipate, or actively counter the narrative. Context: [paste]
R2: The Public Apology Drafter
I need to publicly acknowledge [paste mistake]. Draft three versions:
maximally accountable, balanced, minimally accountable. For each,
predict the reaction from my three most important stakeholders, the
durability of the reset, and the cost to my standing. End with the
version that protects what matters most without sliding into
self abasement.
R3: The Trust Rebuild Plan
[Specific person or group] no longer trusts me on [paste topic].
Draft the rebuild plan over the next three months. The specific
visible behaviours, the conversations to have, the work product to
deliver, and the moments to deliberately under promise. End with the
signal that would tell me the rebuild is working and the signal that
it is not. Context: [paste]
R4: The Narrative Counter
A narrative is forming about me or my work that is not accurate
(paste). Audit the narrative: who is repeating it, what evidence
they are using, where it originated. Draft the counter narrative,
the specific moments to deploy it, and the audience to convince
first. End with the move that is more effective than direct
contradiction.
R5: The Excluded Room Re Entry
I have been quietly excluded from [paste meeting, project, decision
group]. Audit why. Draft the move to re enter without making the
exclusion explicit. For each candidate move (asking my manager,
asking a peer who is in the room, providing unsolicited value to the
group), predict the success rate. End with the recommended path and
the timing.
R6: The "Take The Hit" Move
The situation calls for me to take responsibility for [paste]. Draft
the take the hit move. Two versions: full ownership without naming
others, contextual ownership that surfaces shared responsibility.
For each, predict who reads it as integrity and who reads it as
weakness. End with the version that fits the room and the framing
of the sentence that does the heaviest lifting.
R7: The Recover Without Apology Move
The situation requires recovery but not apology because the
underlying decision was correct even though the outcome was bad.
Draft the response. How to acknowledge the outcome without
disclaiming the decision, how to reassure stakeholders without
caving, how to maintain authority through the rough patch. End with
the sentence that closes the loop. Context: [paste]
R8: The Quiet Exit From A Failing Initiative
I am leading or associated with [paste initiative] that is failing.
Plan the quiet exit that preserves my standing. The handover, the
narrative, the timing, the language for the closing memo. For each
phase, what could tip the exit into looking like an abandonment.
End with the sequence and the sentence I should never say.
R9: The Public Failure Recovery
I or my team had a visible failure on [paste]. Draft the 90 day
recovery arc: the immediate response, the 30 day stabilisation, the
60 day rebuild, the 90 day "where we are now" communication. For
each stage, the audience, the evidence, and the tone. End with the
single visible move that would most accelerate the recovery.
R10: The Reputation Reset Audit
My reputation in this organisation is not where I want it. Audit
the current reputation specifically: what people say about my work,
my judgement, my collaboration, my political instincts. For each
dimension, the cheapest move this quarter to begin shifting the
read, the realistic timeline to see the shift, and the one
behaviour that would unwind the entire reset. Context: [paste]
The POWER Framework in one image
P O W E R
Position Observe Weave Execute Recover
Five pillars. Sequential in rough terms but the loop is constant. By Execute you are always back in Observe mode for new signals. By Recover you may be repositioning entirely. The point is structure: when the dynamics get murky, you can find the pillar that fits and run the right prompt rather than guessing in real time.
How to combine POWER with your model of choice
ChatGPT: 4 series models are strong for the generative pillars (Weave, Recover) where the work is drafting messages and reading specific people. o series models are stronger for the analytical pillars (Position, Observe) where the work is mapping dynamics, scoring stakeholders, and surfacing patterns.
Claude: works well across all five pillars, especially when you load meeting notes, organisational charts, and recent context into a Project. Claude tends to be more honest about the political dynamics in a way the other models hedge around. Tell it explicitly to give you the version a senior peer would deliver in private rather than the public review.
Gemini: deep research mode is useful for the Observe pillar when you are triangulating public signals about your industry, recent executive moves at competitors, or external context that frames your internal politics. Watch for context drift on the longest stakeholder mapping prompts.
The cross model habit that pays compounding dividends is to run the meeting decoder (O1) and the subtext translator (O6) in two models. Where they agree on the read, trust it. Where they diverge, the divergence reveals where your inputs were ambiguous and where you should look harder before making a move.
Frequently Asked Questions
Is using AI for office politics ethical?
The ethical question is what you do with better reads, not whether you have them. AI that helps you communicate clearly, anticipate problems, and rebuild trust is straightforwardly useful. AI that helps you manipulate or damage colleagues is not, and the framework does not provide that. The prompts surface dynamics. The choices remain yours.
Will my colleagues know I am using AI to navigate politics?
Not if you use AI for the thinking upstream and your own voice for everything that touches another person. AI generated prose reads as inauthentic in political contexts because the substance is relational. Use the prompts for analysis. Write messages in your own voice.
What if my company has a "no politics" culture?
Every culture has politics. The "no politics" framing is itself a political move that benefits the people who are good at politics without naming it. In those environments POWER matters more, not less, because the dynamics are still real but harder to read.
How is the POWER Framework different from generic career advice?
Generic career advice tells you to "build your network" and "find a sponsor" without telling you how, in which order, or how to know it is working. POWER gives you the decoding, mapping, drafting, and auditing work. Career advice is the goal. The framework is the operating system.
Can I use these prompts with confidential information about colleagues?
Check your model tier's data retention policy. Enterprise tiers of ChatGPT, Claude, and Gemini typically allow internal personnel context. Consumer tiers usually do not. For sensitive information, use consistent pseudonyms before pasting. The framework reads patterns, not identities.
Is there a paid version of these prompts?
Yes. The fifty prompts above are the free version. The POWER Pro Pack includes 150 expanded prompts with example outputs, ready to load Claude Projects and custom GPT configurations for each pillar, and role specific sets for new managers, ICs aiming for promotion, executives navigating board dynamics, and operators inside founder led companies. The Pro Pack is on the PromptLeadz Pro Collection at $29.
Where to go next
POWER sits across the entire career arc. Pair POWER with the role anchors (7P, FOUNDER, OPS7) to make the politics easier. Pair with LAUNCH when entering a new organisation. Pair with HARDER for the conversations the political work produces. Pair with MONEY for the money conversations that often determine your standing. Pair with CRITIC for the thinking that needs sharpening before any move.
The thing to internalise is that the people who claim not to do politics are the ones who lose most often to the people who do. Engaging deliberately, with honesty about what you are doing, beats being a victim of politics you refused to learn to read. The framework is the operating system. The integrity is yours.
PromptLeadz publishes battle tested AI prompt packs for founders, product, sales, marketing, operations, HR, finance, customer success, adversarial thinking, hard conversations, new role launches, job searches, money conversations, and office politics. All prompts are LLM agnostic. Pricing is in USD.
댓글 남기기: