How to make a big decision when both options look right
The offer letter is sitting in your inbox. The deadline is Friday. And the worst part isn't that the choice is hard — it's that both options are genuinely good. The senior engineering role in Seattle pays more and puts you inside a fast-moving tech hub. Staying in Austin keeps you ten minutes from your parents, your roots, and the life you've built. Your gut says yes to both. That's not indecision. That's a real paradox, and generic advice about "following your heart" doesn't touch it.
What actually helps is structure. Not a pros-and-cons list you'd scribble on a napkin, but a repeatable sequence: write the dilemma as a single sentence, name every criterion you're actually weighing, stress-test each path with a pre-mortem, project forward two years instead of two weeks, then lock in a choice and schedule a review. Every step below comes with a copy-paste prompt template you can run right now. By the end you'll have a living decision log, not just an answer.
Write the decision as one sentence
Most people start with a cloud of feelings. The first job is to compress that cloud into one declarative sentence. Not a question. Not a list. One sentence that names both options and forces you to see the actual trade-off.
The difference in how the AI responds is significant. A vague prompt gets vague output.
Before (vague): "Help me decide between a job in Seattle and staying near my parents."
A structured prompt gives the model something to work with and gives you a sentence you can paste into every subsequent template.
# Decision Statement Generator Write a single declarative sentence that captures the core dilemma: - Option A: Accept the senior software engineer role in Seattle, WA. - Option B: Remain in Austin, TX to stay close to family. Output only the decision sentence.
Output:
Choose between moving to Seattle for a senior software engineer role or staying in Austin to remain near family.
That one sentence is now your anchor. Paste it at the top of every prompt you run. It keeps the AI oriented and keeps you from drifting into new sub-questions every time you reopen the conversation.
One sentence. Every prompt. Non-negotiable.
Name what you are actually weighing
Both options feel right because you're comparing them on different criteria depending on your mood. Monday you're thinking about salary. Tuesday you're thinking about your dad's health. Thursday you're thinking about career trajectory. The criteria keep shifting, so the decision keeps shifting.
Fix that by pulling every factor into one table.
# Criteria Extractor Based on the decision sentence below, list every measurable factor that matters for each option. Decision: "Choose between moving to Seattle for a senior software engineer role or staying in Austin to remain near family." Return a table with columns: Criterion | Option A (Seattle) | Option B (Austin)
Output:
Criterion Option A (Seattle) Option B (Austin) Base Salary Higher (research current market rates) Lower (current role) Career Growth Rate High (fast-growing tech hub) Moderate Family Time per Week ~2 hrs (distance, long commute) ~15 hrs (live nearby) Cost of Living Index Significantly above the national avg. Verify current rates — has shifted in recent years Relocation Stress High (cross-country move) Low
Notice what the table does immediately: a meaningful salary bump looks smaller when you set it against a cost-of-living gap that wide. Seattle consistently indexes well above the national average, which eats into compensation gains faster than most people model upfront. And Austin's cost of living has shifted considerably in recent years, so the financial spread between the two cities is often narrower than the raw salary numbers first suggest. The AI didn't make the decision — it made the trade-off visible. That's the whole point.
Save this table. It feeds every prompt that follows.
Run the pre-mortem on each choice
A pre-mortem is the single most underused tool in personal decision-making. The idea, popularized by psychologist Gary Klein, is simple: imagine it's 12 months from now and the choice failed. Not "what could go wrong" in the abstract — what specifically happened. Forcing that specificity surfaces risks your optimism was papering over.
Run it on both options separately, or the exercise gets contaminated by whichever option you're currently favoring.
# Pre-Mortem Analyzer For each option in the criteria table, imagine it has failed after 12 months. List three plausible failure scenarios and their impact on: - Career trajectory - Personal wellbeing - Financial stability Output results in two separate bullet lists labeled "Option A – Seattle" and "Option B – Austin".
Output:
Option A – SeattleOption B – Austin
- Company restructuring cuts the engineering team, leaving you without a clear career path. Impact: stalled promotion, reduced morale.
- High cost of living forces you to take on contract work to cover rent, eroding work-life balance. Impact: burnout, even less family contact.
- Relocation stress and social isolation compound into sustained mental fatigue. Impact: decreased wellbeing, possible early exit from the role.
- Current employer stalls salary growth due to budget freezes. Impact: slower financial progress than projected.
- Limited exposure to cutting-edge projects reduces skill development over time. Impact: long-term career plateau.
- Family obligations increase unexpectedly, limiting bandwidth for side projects or upskilling. Impact: reduced personal fulfillment.
Read those failure scenarios and notice which ones feel inevitable versus which feel manageable. That reaction is data. The pre-mortem doesn't decide for you — it shows you where your real anxiety lives. Document the three failure modes per option and carry them into the next step.
If you want to go deeper on building repeatable AI workflows for complex analysis like this, the post on building an AI stack you own is worth reading — it covers how to structure multi-step prompt sequences without getting locked into a single tool.
Ask what happens in year two, not week one
Short-term excitement distorts everything. The first weeks of a new job feel electric regardless of whether it's the right move. The first weeks of staying put feel safe regardless of whether that safety is good for you. Week one is almost useless as a signal.
Year two is where the real life shows up. The honeymoon is over. The patterns are set. The question is whether those patterns are ones you can sustain.
# Year-2 Outcome Generator Using the criteria table and pre-mortem insights, write a concise (150-word) story of life in year two for each option. Include: job performance, family relationships, cost-of-living adjustments, and personal fulfillment.
Output:
Seattle – Year Two
After a rocky first year, you secure a lead role on a flagship product. Compensation has grown, partially offsetting the high rent, but you're spending 30+ hours a week on core work and commuting 45 minutes daily. Family visits happen quarterly. Video calls fill the gaps but don't replicate Sunday dinners. You've built a strong professional network and your skills have sharpened, yet personal downtime is scarce enough that you've started scheduling weekly "unplug" evenings just to stay functional.
Austin – Year Two
You negotiate a raise and take on a mentorship role that expands your skill set without requiring relocation. Cost of living has climbed from where it was a few years ago, but your savings rate has room to grow now that you're earning more. Family gatherings are weekly rituals now, a real source of energy. The project scope isn't always cutting-edge, but you feel balanced — and you're six months into a side project that aligns with where you want to be in five years.
Neither story is obviously better. That's the point. The Year-2 Generator doesn't hand you a verdict; it hands you two specific futures to compare honestly. Most people, reading those two paragraphs, feel a pull toward one of them that they didn't feel when staring at a salary spreadsheet. That pull is worth listening to — it's informed now, not just impulsive.
Decide, then set the review date
Analysis paralysis doesn't come from having too much information. It comes from treating the decision as permanent. Nothing closes the loop like locking in a choice and immediately scheduling the moment you'll revisit it.
The review date does two things. First, it gives you permission to stop analyzing — you know there's a checkpoint coming. Second, it creates accountability: if reality diverges from the Year-2 story you wrote, you'll catch it at 90 days instead of two years later when the cost of pivoting is much higher.
# Decision Log Builder Create a markdown file that records: 1. Decision sentence 2. Criteria table 3. Pre-mortem findings 4. Year-2 narratives 5. Final choice and why 6. Review date (90 days from today) with a prompt to re-run the Year-2 Outcome Generator. Output the complete markdown ready for copy-paste into your notes app.
Output:
# Decision Log – Seattle vs. Austin
## 1. Decision Sentence
Choose between moving to Seattle for a senior software engineer role or staying in Austin to remain near family.
## 2. Criteria Table
| Criterion | Option A (Seattle) | Option B (Austin) |
|---|---|---|
| Base Salary | Higher (fill in your offer) | Lower (current role) |
| Career Growth Rate | High | Moderate |
| Family Time per Week | ~2 hrs | ~15 hrs |
| Cost of Living Index | Significantly above national avg. | Verify current rates — has shifted in recent years |
| Relocation Stress | High | Low |
## 3. Pre-Mortem Findings
Option A – Seattle
- Restructuring cuts team, stalls career path
- High COL forces contract work, accelerates burnout
- Social isolation leads to early exit
Option B – Austin
- Salary growth stalls in budget freeze
- Limited project exposure creates long-term plateau
- Unexpected family obligations reduce flexibility
## 4. Year-2 Narratives
Seattle: Strong network, lead role, salary growth — but sparse family contact and no unscheduled time.
Austin: Raise secured, stable savings rate, weekly family time, side project underway.
## 5. Final Choice
Chosen Path: Stay in Austin. Stable finances, strong family support, and manageable career growth outweigh the higher salary and higher stress of Seattle.
## 6. Review Date
Date: [90 days from today]
Action: Re-run the Year-2 Outcome Generator with updated real-world criteria. Compare to this log.
Copy that into Notion, Apple Notes, Obsidian — anywhere you'll actually open it in 90 days. The document is more useful than the decision. It shows future-you exactly what past-you was thinking and why, so the review is a real calibration, not a vague feeling that you made the wrong call.
If you're thinking about how these kinds of structured AI workflows fit into broader time management and productivity — especially when you're already stretched thin — the post on time management techniques for work maps out how prompt sequences can reduce the cognitive overhead of high-stakes tasks without handing the thinking off entirely.
Run it now
The recipe is free. Each link opens the finished prompt with the blanks already named, so you fill in your job and paste the result into the AI you use.
More like this, for whatever you are working on: open Ultra Prompt.