Page
AI Icebreakers for Physicians
AI Icebreakers
Fifteen free prompts for physicians who want to turn one money question into a useful card, map, ledger, or decision memo. No email. Use a paid, top-tier model only.
How to use these
Copy one prompt. Paste it into a paid, top-tier model with extended thinking on and web access on. Answer its questions in one batch, with round numbers. Leave out names, account numbers, addresses, and employer names; a made-up first name on the artifact is fine. Act on the walk-away line. Verify before you act on anything with a number attached.
15 free icebreakers
1. Vital Signs Check
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/financial-vital-signs. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/financial-vital-signs/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute a similarly named framework from somewhere else.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented limits, tax figures, rates, or return assumptions: cite a source or leave it blank for me to fill. No product, carrier, lender, fund, ticker, or firm recommendations. Documents beat my memory; my memory beats your guess. "Can't read yet" is a legitimate answer. When a step needs a licensed human, say so. Artifact first, no preamble, no essay.
Task: run the Financial Vital Signs on my household, all six, exactly as the page defines them and in the page's order. I want a reading per vital, not an overview of personal finance. If the page defines a score or scoring key, apply that key exactly as written and label the total a reading, not a plan. If the page does not define a key, produce the six readings only and do not invent a score or a point system. Pain is self-reported; do not overrule it with the other five.
Ask me everything you need in one batch, no more than 8 questions (career stage and specialty, household, monthly take-home, monthly needs, cash on hand, debts with rates, what I save each month and where, what coverage I carry). Anything I skip, mark "placeholder" and keep going.
Deliver: my Vital Signs Card as a table with columns: vital / my reading / green, yellow, red, or can't read yet / one line why. If and only if the page defines a score, add the page's total and say "this score is a reading, not a plan." Under the table: the ONE vital to treat first and why it goes first; the three unknowns most likely to change a reading, each with the document that would settle it; and the single Fellowship page from financialfellowship.com/ai to open next. If a vital can't be read from what I gave you, write "can't read yet" and say what it would take.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.2. Cash Map Sketch
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/cash-map. There is no live /frameworks/ twin for the cash-map method page. If you can't open those pages, stop and ask me to paste them in. Do not guess what the method says, and do not substitute a budgeting system from somewhere else.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures. No product, carrier, lender, fund, ticker, or firm recommendations. Do not shame spending and do not label Wants as bad; the job is to make cash visible. Artifact first, no preamble, no essay.
Task: draw my cash map using the Fellowship's Needs / Goals / Wants method exactly as the page teaches it. This is a method for giving monthly dollars a job. It is not the Fellowship's 3 Bucket framework. Don't mix the two and don't use the word bucket in this answer.
Ask me everything you need in one batch, no more than 8 questions (monthly take-home from all sources, housing, debt payments, insurance premiums, childcare or school, groceries and household, automatic savings toward named goals, and roughly everything else: travel, dining, help, subscriptions). Anything I skip, mark "placeholder" and keep going.
Deliver: a three-column map, Needs / Goals / Wants, with dollars and percent of take-home in each column; an explicit "unmapped" line for the dollars I couldn't account for, even if it's zero, never hidden in "misc"; the three lines most likely mislabeled (Wants hiding in Needs, or Goals that have no name attached); no more than five questions that would explain the biggest unmapped dollars; the one number this map is asking me to decide; and a one-line note that this is a sketch from memory, not an audit. The audit version uses real statements.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.3. 3 Bucket Sort
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/3-bucket. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/3-bucket/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute a similarly named framework from somewhere else.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures or return assumptions. No product, carrier, lender, fund, ticker, or firm recommendations. Artifact first, no preamble, no essay.
Task: sort every account and dollar I have into the Fellowship's 3 Bucket framework: Safety Net, Wealth Engine, Accelerator, exactly as the page defines them. These are time horizons only. They are not account types, not the Needs / Goals / Wants cash-map method, and not feelings. Page rule to enforce: Accelerator holds only named, dated middle goals that need to arrive on time. Anything without a name and a date goes to "unassigned," not to Accelerator. If nothing has a name and a date yet, Accelerator is empty; empty is a result, not a failure. If a dollar has no defensible job yet, mark it unassigned rather than forcing it into a bucket.
Ask me everything you need in one batch, no more than 8 questions (accounts and balances including checking, savings, brokerage, retirement, HSA, 529, cash; when I'd realistically touch each; goals that have a name and a date; monthly contributions by account; debts for context; partner accounts if shared). Anything I skip, mark "placeholder" and keep going.
Deliver: a sort table with columns: account / balance / bucket / horizon or named date / one line why; bucket totals and percent of the whole; every dollar whose assigned bucket conflicts with its stated horizon or purpose; dollars doing two incompatible jobs at once; and the unassigned list with the question "name it and date it, or it stays out of Accelerator." Do not call a bucket starving, overfed, too small, or too large unless the Fellowship page itself defines the standard being applied. Do not prescribe a transfer amount. End with the single fact or household decision that most needs to be settled before new dollars are redirected, stated as an amount and a direction for me to verify, never a product and never an instruction.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.4. Accelerator Ledger
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/3-bucket. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/3-bucket/ and follow that. If you use the cash-map method's word Want, also open financialfellowship.com/ai/cash-map and use that page's definition; if you can't open it, do not use the word Want. If you can't open the required pages, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute a similarly named framework from somewhere else.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures. No return assumption unless the horizon justifies one, and then say exactly what you assumed and show a low and a high case. No product, carrier, lender, fund, ticker, or firm recommendations. Artifact first, no preamble, no essay.
Task: build my Accelerator ledger: the named, dated middle goals between now and retirement, using the 3 Bucket framework's Accelerator definition and its arrive-on-time standard exactly as the page states them. Goals I might list: a home, a practice buy-in, kids' school, a sabbatical, a car, family help, a wedding.
Ask me everything you need in one batch, no more than 8 questions (each goal with a target date and today's cost; what's already set aside per goal; monthly free cash after current savings; which dates are fixed and which are flexible; which amounts are fixed and which are flexible; anything that's still "someday"). Anything I skip, mark "placeholder" and keep going.
Deliver: a ledger table with columns: goal / date / cost today / set aside / arrive-on-time monthly / on track or at risk. Show the arithmetic. Then: for the goal most at risk, the single assumption that matters most and what happens to the monthly number if the date, the amount, or the contribution moves; the goal that is a Want wearing a goal's clothing (use the cash-map method's word Want, not bucket language, and only if the cash-map page loaded); the collision, meaning two goals fighting for the same dollars, and the honest sequence; and what each "someday" item needs before it counts as an Accelerator goal.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.5. House Decision
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers (if a partner shares this decision, we'll run it together). Before you answer, open and follow these pages: financialfellowship.com/ai/rules, financialfellowship.com/ai/decision-filter, and financialfellowship.com/ai/3-bucket. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/decision-filter/ or financialfellowship.com/frameworks/3-bucket/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the frameworks say, and do not run a generic rent-versus-buy calculator in their place.
Ground rules: education, not mortgage, tax, legal, real-estate, or personalized financial advice. Use my numbers and label every assumption. No invented rates, appreciation, or tax figures; a range with the assumption shown, or leave it blank. No lender, loan product, or agent recommendations. Artifact first, no preamble, no essay.
Task: run the Decision Filter, the five questions exactly as the page defines them, on "buy this house now" versus "not yet" (rent, wait, or buy later). Two defensible answers; I want the one that fits my facts and the assumption that flips it. Use 3 Bucket only to test whether closing would consume Safety Net money or displace a named, dated Accelerator goal. Treat the down payment and closing cash as an Accelerator goal: a name, a date, and an arrive-on-time monthly amount, per the 3 Bucket page. The Safety Net is not closing cash; if this purchase would spend it, say so in plain words. Do not run a full rent-versus-buy model, choose a mortgage structure, estimate appreciation, calculate a maximum house price, or perform the document-backed home analysis here.
Ask me everything you need in one batch, no more than 8 questions (likely years in the home; price and the total cash to close and move; the all-in monthly cost of owning, including taxes, insurance, dues, and maintenance, against what I pay now; cash reserve today and what would be left after closing; my other named, dated goals; career and location certainty: contract term, partnership track, restrictive covenants; household income stability; the date pressure, and whether it's real or manufactured). Anything I skip, mark "placeholder" and keep going.
Deliver: my House Decision Card: the five questions answered with my facts; the strongest buy case and the strongest not-yet case side by side; the Safety Net after closing, in months of Needs; known facts versus material unknowns; which Accelerator goals would lose funding if I close now, named only; the assumption that flips the answer; and the observable conditions under which I walk away from this house even if I love it, each tied to a number or a date. If the answer is not yet, the number and the date that make it yes, written as an Accelerator goal. Stop there. What the purchase does to my other dated goals' arrival math and to my lifestyle floor is a second pass with real numbers, not a sketch from memory.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.6. Transition Window
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/attending-transition-blueprint. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/attending-transition-blueprint/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute a generic new-attending checklist from somewhere else.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures. No product, carrier, lender, fund, ticker, or firm recommendations. Contract terms go to a contract specialist and coverage goes to an insurance specialist; say so where it applies. Artifact first, no preamble, no essay.
Task: I'm in or approaching the training-to-attending jump, or I made it already and I'm not sure the systems caught up. Build my Attending Transition Blueprint checklist for the window the page defines, dated to my actual start. If I'm already past my start date, don't pretend I'm at day one: mark each item done / partly done / never built / not applicable, and say what is late, what still fits, and what waits. Do not run a full student-loan analysis here; if loans are the loud decision, point me to the Loan Decision icebreaker after this card exists.
Ask me everything you need in one batch, no more than 10 questions (training level and start date, or months since start; specialty and practice type; offer or contract status; state move; loans: balance, plan, PSLF payment count, employer type; disability coverage in place; family and dependents; expected or actual change in take-home; housing plan; anything already signed). Anything I skip, mark "placeholder" and keep going.
Deliver: a dated checklist across the window phases the page names, dated to my actual start. If the page does not name phases, do not invent a calendar: list the page's items (or the items it tells you to take from my dates) ordered by deadline against my start date, and say in one line that the page didn't name phases. Each item carries what, why, deadline, status, and who does it (me, contract specialist, insurance specialist, planner, employer benefits office). If I'm already past my start date, mark each item done / partly done / never built / not applicable, and say what is late, what still fits, and what waits. Then: the two items that are hardest to undo if missed, and the single decision that has to be made before the first attending paycheck arrives, applied to my dates.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.7. Guardrails Check
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/lifestyle-guardrails. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/lifestyle-guardrails/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute a "live like a resident" rule or any other rule from somewhere else.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures. No product, carrier, lender, fund, ticker, or firm recommendations. Do not moralize about spending, travel, or childcare; the job is to make lifestyle growth explicit. Artifact first, no preamble, no essay.
Task: my income went up, or is about to. Apply the Lifestyle Guardrails, specifically the Double Up Rule exactly as the page states it, to my numbers. Crucially, distinguish a recurring raise from a one-time bonus as input types. Apply the page exactly as it states the rule to each type. Do not invent a second rule and do not restate the page's arithmetic. If this is a one-time bonus, do not treat it as monthly floor growth unless the page says to. Tell me what may grow, by how much, and what stays frozen.
Ask me everything you need in one batch, no more than 8 questions (monthly take-home before and after; my major spending lines today; the upgrades on my mind such as home, car, school, travel, help, a second property; current monthly savings and where it goes; anything already committed; whether this is a permanent recurring raise or a one-time bonus). Anything I skip, mark "placeholder" and keep going.
Deliver: my Guardrails Card: the income change in dollars and percent; the Double Up Rule applied line by line, each marked may grow to / frozen / shrink, with the math; what the raise does to my saving before it touches lifestyle, per the page; the upgrades split into recurring commitments that raise the floor and are hard to reverse versus one-time upgrades that don't; the one upgrade that passes and the one that fails, and what would make the failing one pass; a "before we add another fixed cost" test in one sentence; and a one-line guardrail I can say out loud the next time an upgrade comes up.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.8. Protection Gap Card
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/protection-stack. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/protection-stack/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute an insurance checklist from somewhere else.
Ground rules: education, not personalized insurance advice. Use my numbers and label every assumption. No invented figures, premium estimates, or coverage amounts. Never name a carrier or a product, and never tell me how much to buy. Policy language controls, not your summary of it. When a layer needs a licensed human, say so. Artifact first, no preamble, no essay.
Task: walk my Protection Stack from the ground up, all six layers exactly as the page defines them, starting at health. Give me a gap card, not a lecture and not a quote. A gap in a lower layer exposes everything above it, so read them in the page's order.
Ask me everything you need in one batch, no more than 10 questions (dependents and ages; income; employer benefits: health, group disability, group life; individual disability policy: monthly benefit, elimination period, benefit period, own-occupation definition yes / no / unsure and whether it's specialty-specific, taxable or not, riders I know of, and whether any training-window offer is still open; term life: amount and years left; umbrella amount; malpractice: claims-made or occurrence and who pays the tail; home, auto, renters; estate documents: will, guardianship, powers of attorney, beneficiaries current; debts and rough net worth). Anything I skip, mark "placeholder" and keep going.
Deliver: a six-layer card with columns: layer / what I have / what the layer is for, per the page / fine, thin, missing, or unknown / the question to ask about it. For every thin, missing, or unknown layer: the risk I'm retaining, the fact you still need, and the document page I should inspect. Then: the one layer to fix first and why it goes first; a "policy language to look up" list (definitions, elimination periods, exclusions, offsets, portability), never carriers or products; the document to photograph or upload next (the declarations page, not the whole booklet); and which layers go to the insurance specialist, with what to bring.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.9. Income Stop Test
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules, financialfellowship.com/ai/income-machine, financialfellowship.com/ai/protection-stack, and financialfellowship.com/ai/3-bucket. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/income-machine/, financialfellowship.com/frameworks/protection-stack/, or financialfellowship.com/frameworks/3-bucket/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the frameworks say.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures. Never name a carrier or a product. Artifact first, no preamble, no essay.
Task: run the Income Machine's "what if it stops" question on my household. My income stops tomorrow, once for illness or injury and once for job loss. Show the runway month by month. Use the Protection Stack page for how disability coverage enters the math. Use the 3 Bucket page for what the test says about the size of my Safety Net bucket. Show the system before anyone talks about buying anything.
Ask me everything you need in one batch, no more than 10 questions (my monthly take-home; partner take-home and how stable it is; monthly Needs, from my cash map if I have one; cash reserve; individual disability policy: benefit, elimination period, taxable or not; group disability; employer sick leave or PTO in weeks; benefits attached to my job: health, retirement match, group life; debts and minimum payments; liquid assets outside retirement; dependents). Anything I skip, mark "placeholder" and keep going.
Deliver: two runway tables, one for disability and one for job loss, at months 1, 3, 6, 12, and 24, with columns: income sources / Needs / gap / what covers the gap. Then: what disappears with the job besides the paycheck, and when; the month the plan breaks and what breaks first; the weakest link in my income machine; the single change that adds the most months (a coverage definition, a reserve target, or a cost, never a product); and what the test says about the size of my Safety Net bucket.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.10. Contract Lens
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/income-machine. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/income-machine/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says.
Ground rules: education, not legal, employment, or personalized advice. A physician contract specialist reviews before I sign anything, and an attorney where the wording is legal; you say so. Read only what I paste; if I haven't pasted the contract, output the blank card and the list of sections to paste, and stop. Keep three things separate everywhere: what the document says, what you infer, and what stays ambiguous. Use my terms and label every assumption. No invented market figures. On state non-compete rules or anything contested, verify it against a current source or leave it out; never assert a bright line. Artifact first, no preamble, no essay.
Task: read my employment offer or contract through the Income Machine: exactly how this job produces my income, and what could turn it off.
Ask me everything you need in one batch, no more than 10 questions (specialty and setting; comp structure: base, wRVU rate and threshold, bonus terms, call pay, partnership track; term and termination, including notice without cause; non-compete radius and duration; malpractice type and who pays the tail; sign-on and relocation with clawback terms; PTO and CME; retirement plan and match; my expected productivity against the threshold; and, redacted, the compensation and termination sections pasted in). Anything I skip, mark "placeholder" and keep going.
Deliver: my Contract Card: how the income is produced, written as a plain-English formula that marks what I control and what I don't, with each input tagged says / inferred / ambiguous; the off switches, ranked by how quickly each could interrupt or reduce income; the material terms whose economics or wording deserve discussion before signing; and a contract-specialist question list. For each question include: clause or topic / what the document says / economic consequence visible from the text / what remains ambiguous / question to ask the contract specialist. If I supplied a proposed negotiation number, organize it and show what term it changes. Do not invent market compensation numbers, tell me what I should demand, determine enforceability, or tell me a clause should or should not be signed.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.11. Work-Less Test
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules, financialfellowship.com/ai/income-machine, and financialfellowship.com/ai/cash-map. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/income-machine/ and follow that. There is no live /frameworks/ twin for the cash-map method page. If you can't open the rules page or at least one required page, stop and ask me to paste them in. Do not guess what they say.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures; where tax effects are unknown, show direction and a range, not false precision. No product, carrier, lender, fund, ticker, or firm recommendations. Artifact first, no preamble, no essay.
Task: run the Work-Less Test: what happens to my plan if I go from my current FTE to a lower one. Use the Income Machine page for how my income is actually produced (does pay fall in a straight line with hours, and what happens to benefits and retirement match) and the cash-map method, Needs / Goals / Wants, for what absorbs the cut. Needs / Goals / Wants is a cash-map method. It is not the 3 Bucket framework. Keep them separate. Income is not the only variable: count the time I get back and the costs that change with it.
Ask me everything you need in one batch, no more than 8 questions (current and target FTE, with take-home today; comp structure: salary, production, or shift; how benefits and match change at the target; my monthly Needs, Goals, and Wants; named, dated goals and a rough retirement target date; partner income; costs that would disappear or appear with the change, such as childcare, commuting, or call coverage; how long the change would last). Anything I skip, mark "placeholder" and keep going.
Deliver: my Work-Less Card: income at the target FTE with benefit and production effects flagged; the new Needs / Goals / Wants map and the gap, with the expenses that disappear and the expenses that appear; which column absorbs the cut, and in what order; which dated goals the cut touches, named, without re-dating them; a "yes if / no unless" line; the three facts I need before deciding; and the version of this that's a trial rather than a decision. Each goal's new arrival date, the reversibility of each consequence, and the guardrail to set before I cut are a second pass with real numbers, not a sketch from memory.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.12. Funding Order
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/tax-efficiency-ladder. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/tax-efficiency-ladder/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute a generic "max everything" order from somewhere else.
Ground rules: education, not personalized tax advice. Use my numbers and label every assumption. Tax rules and contribution limits change: never invent a limit or rely on a remembered one. Point me to the current IRS page or leave the number blank for me to fill. No product, fund, or firm recommendations. Artifact first, no preamble, no essay.
Task: put my accounts on the Tax Efficiency Ladder: rungs in the order the page defines them, not a menu, with my dollars on each rung for this year.
Ask me everything you need in one batch, no more than 10 questions (W-2, 1099, or both; plans available: 401(k) or 403(b) and the match, 457(b) governmental or not, an HSA-eligible health plan, whether a backdoor Roth is done this year and whether any pre-tax IRA balance exists, whether after-tax contributions are allowed, a solo 401(k) if I have 1099 income; contributions so far this year by account; income and filing status; state; loans and PSLF status if pre-tax contributions affect my payment; kids and any 529 state benefit; months left in the year). Anything I skip, mark "placeholder" and keep going.
Deliver: a ladder table with columns: rung / my account / this year so far / what filling the rung takes / gap / one mechanics note. Then: the first rung I'm out of order on; the rung I'm skipping, and whether the skip might still be rational for me (liquidity, a dated goal, debt, plan quality, employer restrictions) or is just a skip; what skipping it costs in rough terms with the assumption shown; the verify list for my plan document or my own CPA (current limits, pro-rata, non-governmental 457 risk, HSA eligibility); and the one paycheck election this ladder is asking me to look at next, for my plan document or my own CPA to confirm before anything changes.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.13. Wealth Equation Scorecard
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/physician-wealth-equation. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/physician-wealth-equation/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says, and do not substitute a net-worth benchmark or "you should have X by age Y" from somewhere else.
Ground rules: education, not personalized investment advice. Use my numbers and label every assumption. No invented return figures or peer comparisons; if you must assume a return to make a point, state it and show the sensitivity. No product, fund, ticker, or firm recommendations. Artifact first, no preamble, no essay.
Task: score me on the Physician Wealth Equation, all five components exactly as the page defines them. The page claims income is rarely the component doing the work. Test that against my numbers. Don't assume it's true or false for me.
Ask me everything you need in one batch, no more than 8 questions (age and years to my target; gross and take-home; savings rate, or dollars saved per month; invested assets and rough mix; fees: fund expense ratios and any advisory fee; debt balances and rates; spending trend: up, flat, or down; big changes coming). Anything I skip, mark "placeholder" and keep going.
Deliver: a scorecard with columns: component / my reading / working, idle, or working against me / one line why. Then: the component carrying my wealth-building right now; the one I'm leaning on too hard; the single lever with the biggest effect for me, with the arithmetic that justifies calling it biggest; and what to re-score in a year.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.14. Loan Decision
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules, financialfellowship.com/ai/decision-filter, and financialfellowship.com/ai/3-bucket. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/decision-filter/ or financialfellowship.com/frameworks/3-bucket/ and follow that. If I tell you I am still in training or inside the attending window, also open financialfellowship.com/ai/attending-transition-blueprint (or its live twin at financialfellowship.com/frameworks/attending-transition-blueprint/) and apply its timing only: which decisions can wait for attending income and which paperwork cannot. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the frameworks say.
Ground rules: education, not personalized advice. Use my numbers and label every assumption. No invented figures, forgiveness odds, or current refinance rates. Student-loan program rules change and some points are contested: verify against the current official source or leave them out; never assert a bright line. No product, lender, servicer, or firm recommendations. Artifact first, no preamble, no essay.
Task: run the Decision Filter, the five questions exactly as the page defines them, on my student-loan fork. Compare only paths that are actually available from my verified facts. Three paths only: pursue forgiveness (PSLF or another program I'm eligible for), refinance, or stay and pay it off. Usually two are defensible. I want the path that currently appears to fit my stated priorities best, why the alternatives currently fit less well, and the assumption that would flip that provisional conclusion. This is an educational comparison, not authorization to refinance, consolidate, change repayment plans, waive federal protections, or make another materially irreversible loan move. If a required program fact is unverified, the conclusion stays provisional. Run the Filter on the path, not on a lender. If I'm still in training, say which decisions can wait for attending income and which paperwork cannot. If there aren't enough verified facts to pick, say so and give me the missing list instead of a guess.
Ask me everything you need in one batch, no more than 10 questions (federal and private balances, rates, loan types; repayment plan; qualifying forgiveness payments so far; employer type now and likely next; income trajectory; household and filing status; years to attending, or years as an attending; my tolerance for program or rule changes; monthly cash-flow room; when I want to be done; what I've been toggling between). Anything I skip, mark "placeholder" and keep going.
Deliver: my Decision Card: the five questions answered with my facts; the paths side by side (years to done, total cost as a range with the assumption shown, risk, flexibility, and what I'd have to believe for each); a PROVISIONAL FIT or NOT ENOUGH VERIFIED FACTS line, and why the other paths lost in my facts; the assumption that flips the answer; what the fitting path does to my Safety Net bucket; and a verify list for current program rules, my qualifying payment count, and plan eligibility, each with the official source to check. No payment-by-payment schedule and no full scenario table here; that takes my servicer export, not my memory. Don't give me an action list; the verify list is what I do next. Do not tell me to refinance, consolidate, switch plans, or submit an irreversible election.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.15. Advisor Vetting
[Run this in a paid, top-tier model with extended thinking and web access turned ON. Free or cheap tiers give confident wrong answers here. Skip them.]
I'm a physician using the Financial Fellowship's free AI Icebreakers. Before you answer, open and follow these pages: financialfellowship.com/ai/rules and financialfellowship.com/ai/decision-filter. If an /ai/ framework page will not load, open the live twin at financialfellowship.com/frameworks/decision-filter/ and follow that. If you can't open the rules page or at least one framework page, stop and ask me to paste them in. Do not guess what the framework says.
Ground rules: education, not personalized advice and not a referral. Use my numbers and label every assumption. No market-return theater: if fee math needs a growth assumption, state it and show a low and a high case. Never name a firm or an advisor. Artifact first, no preamble, no essay.
Task: run the Decision Filter, the five questions exactly as the page defines them, on "keep DIY versus hire a planner this year." Then, if the answer leans hire, give me the vetting script. Anchor it on the pay test: fee-only means I pay them and nobody else pays them, which is not the same as fee-based.
Ask me everything you need in one batch, no more than 8 questions (rough net worth and complexity: 1099 income, practice ownership, real estate, kids; hours per month I'll actually spend on this; the decisions I've been avoiding; current help: CPA, advisor, and how they're paid, including any percent-of-assets fee; what a real mistake would cost me; what I'd pay per year for good advice; what made me consider it now). Anything I skip, mark "placeholder" and keep going.
Deliver: my Decision Card (five questions answered; both answers side by side; the assumption that flips it); fee math on my assets over 10 and 20 years for a percent-of-assets fee versus a flat fee, assumptions shown; a 10-question vetting script covering fiduciary at all times, fee-only and how that's verified (the disclosure filing and the dollar fee in writing), exactly how they're paid, who holds custody, credentials, physician experience, planning versus investments only, and what they won't do; the red flags that end the meeting; the walk-away line I can say to a pitch; and what to bring to a first meeting if I hire.
Close with three lines: VERIFY (the one thing to check before I act on this), WALK-AWAY (the one line I take with me), HUMAN (planning, own occupation disability for physicians, or stay DIY for now). Do not name a person, a firm, a logo, or a link. If a lane is not in play, say DIY.