Airfare Is a Decision Problem, Not a Shopping Problem
WEEK 94 :: POST 2 :: CHATGPT
Directions Given To The A.I. This Week+
Instructions Given to each A.I. — Please provide 3 prompt variations that share this objective:
Each A.I. also received two static attachments: the blog post template (structure) and the authoring instructions (voice and standards). The text below is the week-specific assignment as sent — reflowed for the web; wording unchanged.
I'd like you to write this week's Ketelsen.ai post. Two files are attached: the blog post template (the structure to follow) and the authoring instructions (context, voice, and standards). Please read both before you begin, then produce the complete post in a single response.
This week's theme: "Getting the Flights Right" — Airfare Strategy.
This is Week 3 of an eight-week series on planning a vacation with AI. Week 1 established the reader's real constraints — a validated budget ceiling and a constraint profile. Week 2 turned that into a chosen destination, or a short ranked list of finalists. This week they buy the hardest part of the trip.
Flights are where vacation budgets are won and lost, and where most travellers feel least in control. Prices move daily, the rules are opaque, and the internet is full of confident advice that is either outdated or was never true. The job this week is to give the reader a defensible booking decision — knowing what a fair price looks like for their route, when to buy, what to trade, and when to stop optimising and just book.
The three prompts should help a reader work through:
- What "a good price" actually means for their specific route — a fare is only cheap relative to that route's own history and season, and a reader with no baseline cannot tell a deal from a markup.
- Timing and the cost of waiting — how to decide whether to book now or hold, and how to put a number on the risk of waiting rather than guessing.
- The real trade space — connections, nearby airports, off-day departures, red-eyes, basic-economy restrictions, and baggage. Each saves money and spends something else; the reader should see the exchange rate, not just the headline fare.
- Total cost, not ticket price — seats, bags, changes, and the ground transport a cheaper outlying airport quietly adds back.
- When to stop — a stopping rule that prevents weeks of fare-watching for a saving that no longer justifies the attention.
The output a reader should walk away with is a booking decision they can defend: this fare, on this routing, bought now or held until a stated date, for these reasons.
A note on the strongest version of this week: at the advanced end, this is a fare decision framework — a baseline for the route, a target price, a walk-away price, a hold-or-book rule tied to a date, and an explicit list of the trade-offs the reader will and will not accept. That structure is worth reaching for.
A hard constraint, and the most important instruction in this brief. AI models cannot see live fares, and their price knowledge is stale by construction. No prompt in this post may ask the AI to state a current price, predict a specific future fare, or claim what a route "usually costs" right now. That is the single most damaging thing an AI can do to a traveller in this domain — it produces confident, checkable, wrong numbers, and the reader finds out at the checkout page.
Design the prompts so the AI does what it is genuinely good at: structuring the decision, naming the variables, building the comparison framework, and telling the reader what to go and look up. The reader supplies the live data from a fare search; the AI turns it into a decision. Prompts that make this division of labour explicit are the strongest possible answer to this week's theme, and posts that blur it should expect to be marked down on Practical Utility.
Series dependency chain, for the Metadata block: Week 3 consumes the destination (or final shortlist) chosen in Week 2 and the budget ceiling validated in Week 1 — the airfare decision is scored against both, and a fare that breaks the ceiling is a signal to revisit the destination, not to quietly raise the budget. Week 3 produces the confirmed routing and dates, which Week 4 (lodging) and Week 5 (itinerary) both assume. Locked flights are what turn a plan into a trip.
Because readers may arrive at this post without having read Weeks 1 and 2, the prompts should work for someone who knows roughly where they are going and what they can spend, while making clear they get far more from them with a real constraint profile and a chosen destination in hand.
Three difficulty tiers as always — Beginner, Intermediate, Advanced — each a genuinely different approach to the same problem, not the same prompt at three lengths.
On examples: this is a consumer travel topic. The template lists tech startup / retail / freelance as suggested industry examples — those are marked MAY, and this week you should almost certainly adapt them. Families coordinating school holidays, couples with mismatched leave, solo travellers with flexible dates, and people flying to a fixed-date event are the right contexts here. Choosing them over the suggested business examples is correct behaviour and will not be scored against you.
A note on supplied figures. Anything marked `[SUPPLIED — use as given]` above came from Ketelsen.ai's own research brief. Use it freely — you are not fabricating by repeating it, and you will not be marked down for leaving it uncited. Do not attach an invented source to it. (No supplied figures this week. Given the live-pricing constraint above, this week is a particularly bad one to invent any — if you find yourself reaching for a number, that is the signal to restructure the prompt so the reader supplies it instead.)
## BEFORE YOU SUBMIT — STRUCTURAL CHECK
(This block is identical every week. It exists because these specific items are the ones posts drop, and a dropped structural item costs compliance points for something that takes one minute to add.)
Your post is parsed by a script before any human reads it. Confirm all seven:
1. ☐ Response begins with `PLATFORM: <your name>` and `WEEK: 3` 2. ☐ `## Lead` present once, at the very top, before Variation 1 3. ☐ `## In one line` present in all three variations 4. ☐ `## What this prompt gives you` present in all three variations 5. ☐ `## The Prompt` present in all three variations, with the prompt in double quotes beneath it 6. ☐ `## Introductory Hook` and `## Current Use` present in all three variations (three of each — not one) 7. ☐ Every template heading written as `##`, none bolded instead; prompt breakdown is running text split on ` : `, with no `###` headings inside it
A complete post has 57 `##` headings. If your count is well short, a section is missing or was bolded instead of hashed.
One extra check this week: confirm no prompt asks the AI to state, predict, or recall a specific airfare. If one does, restructure it so the reader brings the fare and the AI brings the framework.
Airfare looks like a shopping problem, but it is really a decision problem: the cheapest ticket may be the wrong flight, waiting may save nothing, and an attractive fare can quietly break the vacation budget after bags, seats, and airport transfers are added. This week’s three prompts turn live search results into a defensible decision through a beginner-friendly Fare Reality Check, an intermediate Total-Cost Trade Space, and an advanced Fare Decision Policy with targets, deadlines, and a stopping rule. The AI supplies the structure; you supply the live fares.
The Fare Reality Check
Turn pasted live fares into one clear book-or-hold decision.
A fare-search screen can give you dozens of flights and still leave you unable to answer the only question that matters: should you buy one? The lowest number may hide baggage charges, a punishing connection, an outlying airport, or restrictions that make the ticket unusable for your family. The Fare Reality Check is for the traveler who wants a clear answer without building a spreadsheet or learning airline pricing theory. You bring several live options, your budget, and your non-negotiables; the AI turns them into an apples-to-apples comparison and tells you what to verify before committing.
Why this matters now
Modern fare-search tools can show date comparisons, alternative airports, route-specific price indicators, and price tracking, but those features still leave the traveler responsible for interpreting the trade-offs. Google Flights, for example, distinguishes between “Best” and “Cheapest” results because the lowest-priced itineraries can involve self-transfers, airport changes, or other compromises. Its tracking system can also notify users when fares change significantly, allowing the tool to watch prices while the traveler follows a defined decision rule. (Google Help)
This prompt provides that rule without pretending the AI knows today’s fare. It requires the model to work only from information you paste into the conversation, mark missing information as unknown, and separate the ticket price from the actual cost of taking the flight.
Act as a practical airfare decision coach. You do not have reliable access to live fares, so do not state, recall, estimate, or predict a current airfare. Use only the live information I provide.
Start by asking me for the following:
1. My origin, destination, travel dates, and number of travelers.
2. My total trip budget ceiling and the maximum amount currently allocated to airfare.
3. My booking deadline: the last date on which I am willing to leave the flights unbooked.
4. My non-negotiables, such as maximum stops, acceptable departure times, required baggage, mobility needs, airport limits, or fare-change flexibility.
5. Three to five live flight options copied from a fare-search tool or airline booking page.
6. For each option: ticket price for the entire party, airline, airports, departure and arrival times, total duration, number and length of connections, fare class, included baggage, seat fees, change or cancellation restrictions, and estimated ground-transport cost.
7. Any route-specific price information displayed by the search tool, such as low, typical, or high; a historical price range; or a booking-timing message. Treat this only as data I supplied, not as your own knowledge.
If information is missing, write Unknown and tell me exactly where to verify it. Never fill a missing fare, fee, restriction, or price history with an estimate.
Then:
* Calculate the all-in airfare cost for my party by adding the ticket price, required bags, required seats, airport transfers, and any unavoidable flight-related expense I supplied.
* Compare the all-in cost with my airfare allocation and total trip budget ceiling.
* Identify the main sacrifice attached to each option, such as extra travel time, an inconvenient airport, a restrictive fare, a red-eye, or a risky connection.
* Label each option Acceptable, Stretch, or Reject according to my budget and non-negotiables.
* Identify the strongest option, if one is defensible.
* Finish with exactly one recommendation: BOOK, HOLD, or SEARCH AGAIN.
Use BOOK only when a viable option fits my limits and the benefit of continued searching is not worth the remaining uncertainty or attention.
Use HOLD only when I still have time before my booking deadline and there is a specific condition worth waiting for. Give me a recheck date based on my deadline and a booking trigger based on information I supplied. Do not predict that the fare will fall.
Use SEARCH AGAIN when every option violates an important constraint, exceeds the airfare allocation, or depends on information that has not been verified.
End with a two-sentence decision record stating the selected flight or next action, the reason, and the condition that would change the decision.
How the AI reads this prompt
Practical examples from different industries
Illustrative example — A family traveling during a school break:
Two parents are planning a family vacation with dates constrained by the school calendar. They paste several live flight options, including a lower-priced itinerary with a long evening connection, a nonstop basic-economy option, and a more flexible fare from their preferred airport. The prompt asks them to verify baggage and seat-selection charges before comparing anything. The resulting decision rejects the connection because it arrives too late for the children, identifies the basic-economy option as a Stretch because the family needs adjacent seats, and recommends booking the flexible nonstop once its verified all-in cost is confirmed to fit the Week 1 airfare allocation.
Illustrative example — A traveler attending a fixed-date wedding:
A solo traveler must arrive before a wedding rehearsal and cannot move the outbound date, although the return is flexible. He pastes live options from two nearby departure airports and includes the cost and travel time required to reach each one. The AI discovers that the headline-cheapest ticket becomes less attractive after the longer airport transfer and restrictive return rules are included. It recommends the slightly higher ticket from the closer airport because it satisfies the arrival requirement, stays within the existing budget, and removes the risk of a self-transfer before a fixed event.
Illustrative example — A couple with mismatched leave schedules:
One partner can leave a day early, but the other cannot depart until after work. The couple provides separate live options for traveling together and for meeting at the destination. The prompt calculates the total party cost rather than evaluating each ticket in isolation, then exposes the inconvenience of coordinating two itineraries. The output recommends traveling together on the later flight because the separate-ticket saving is too small to justify duplicate airport transportation, a longer overall travel day, and the possibility that one partner reaches the destination alone after a disruption.
Creative use case ideas
- Compare an airline ticket with a rail-and-flight combination when the nearest major airport requires a long ground transfer.
- Evaluate whether using points for one traveler and cash for another creates useful savings or merely adds incompatible fare rules.
- Help a student group choose between one shared itinerary and several cheaper departures while preserving an acceptable arrival window.
- Recheck a booked fare during an applicable cancellation period using verified airline rules rather than assuming every booking can be changed.
- Create a decision record for relatives who keep forwarding new flight options after the family has already chosen a defensible itinerary.
Adaptability tips
For a quick decision, paste only the two most realistic flights and ask the AI to identify the missing facts that could reverse the choice. For a flexible traveler, add several date combinations and tell the model to treat each date pair as a separate option. For international travel, expand the intake to include passport-validity timing, overnight connections, terminal changes, separate-ticket baggage handling, and the cost of an unexpected hotel.
The same structure can also compare award bookings, but points and cash should remain separate until you supply a personal value for the points. Do not let the model invent that value. When coordinating a group, add a rule that an option is viable only if the required number of seats remains available at the quoted fare when the final airline checkout is opened.
Pro tips
- Paste a screenshot’s text or copy the final checkout summary, not only the first fare-search result. The checkout page is where fare restrictions and optional charges become easier to verify.
- Add a “decision reversal” request: ask the AI to name the single missing fact most likely to change its recommendation.
- Ask for a five-line verification checklist to complete immediately before payment.
- Save the two-sentence decision record with the date and time of the live search so later comparisons do not mix observations from different moments.
Prerequisites
Know your approximate destination, travel window, number of travelers, and maximum airfare allocation. Readers who completed Week 1 should bring their validated budget ceiling, and readers who completed Week 2 should bring the selected destination or final shortlist. You also need at least two realistic live flight options copied from a current fare-search tool or airline website.
Before using the recommendation, verify the final price, baggage allowance, fare class, airport, dates, traveler names, and change or cancellation terms on the actual booking page.
Required tools
Any current general-purpose conversational AI that can analyze text and perform basic arithmetic. You also need a live fare-search source and access to the airline or booking provider’s checkout page. A spreadsheet is not required.
Frequently asked questions
Q: Can I ask the AI whether the fare is historically cheap?
A: Only after you provide route-specific evidence from a live search tool, such as a displayed price label, historical graph, or dated observations you collected yourself. The AI can interpret that evidence, but it should not recall what the route supposedly costs or create a baseline from memory. A good price is specific to the route, dates, cabin, baggage assumptions, and restrictions you are actually considering.
Q: What happens if the lowest option has an unknown baggage fee?
A: The prompt should leave the fee marked Unknown and suspend any conclusion that depends on it. Check the airline’s current baggage page or proceed far enough through booking to see the passenger-specific charge. Google Flights also warns that displayed airfare may exclude additional services such as luggage, depending on the airline or booking option. (Google Help)
Q: Does booking a flight give me twenty-four hours to change my mind?
A: For flights covered by the United States rule, an airline must provide either a twenty-four-hour hold or a penalty-free cancellation and refund when the ticket is purchased at least seven days before departure. Airlines do not have to offer both choices, and the federal requirement does not automatically apply to tickets bought through an online travel agency or other third-party agent. Verify the seller’s policy before relying on the protection. (Department of Transportation)
Q: What if the prompt recommends HOLD and the fare rises?
A: HOLD is not a prediction that the fare will fall. It is a controlled decision to keep searching until a stated date while accepting the possibility that the current option disappears or becomes more expensive. The prompt should therefore name a maximum acceptable price, a booking deadline, and the condition that ends the hold.
Recommended follow-up prompts
- “Turn this flight decision into a final pre-purchase verification checklist covering names, dates, airports, baggage, seats, fare rules, and total payment.”
- “Using my confirmed routing and dates, create the lodging search boundaries Week 4 should follow without changing my total trip budget.”
- “Create a simple flight-price observation log I can update without asking you to predict future fares.”
Tags and categories
Tags:
airfare, flight booking, vacation planning, travel budget, total trip cost, fare comparison, beginner prompts, decision making
Categories:
Travel Planning, Personal Finance
Citations
- Google Travel Help, “Track flights & prices.” (Google Help)
- Google Travel Help, “How to find the best fares with Google Flights.” (Google Help)
- Google Travel Help, “Find plane tickets on Google Flights.” (Google Help)
- U.S. Department of Transportation, “Refunds.” (Department of Transportation)
The Total-Cost Trade Space
Price every compromise, then choose the trade-off you actually prefer.
Travelers often compare flights as though price were the only number changing. It is not. A cheaper itinerary may demand an extra connection, an earlier alarm, a longer airport transfer, a night without sleep, stricter baggage, or several hours that disappear from the vacation itself. The Total-Cost Trade Space converts those compromises into visible exchange rates: how much cash you save for each extra travel hour, each airport change, or each restriction you accept. Instead of declaring one universal winner, it produces three strategies—cash-minimizing, balanced, and time-protecting—so the traveler can see which bargain matches the trip they are actually taking.
Why this matters now
Fare-search tools increasingly expose the difference between a low price and a convenient itinerary. Google Flights says its “Best” results balance factors such as price, duration, stops, and airport changes, while its “Cheapest” results may include self-transfers or different airports that require additional work from the traveler. It also provides date, price-graph, and alternative-airport tools for people who can change parts of the trip. (Google Help)
Those features are useful inputs, but they do not know what an hour of vacation time, a red-eye, or a distant airport is worth to you. This prompt makes those preferences explicit and applies them consistently across the live options you supply.
Act as an airfare trade-off analyst. You do not have reliable live-fare access. Do not provide a current fare, route average, booking-window rule, or future price prediction. Use only the live prices, fees, route indicators, and preferences I supply.
Build the analysis in five stages.
STAGE 1 — INHERIT THE TRIP CONSTRAINTS
Ask for:
* Origin and destination.
* Required arrival and departure windows.
* Number and needs of travelers.
* Total vacation budget ceiling.
* Current airfare allocation.
* Booking deadline.
* Destination or shortlist selected in the previous planning stage.
* Any condition that would force me to revisit the destination rather than increase the budget.
STAGE 2 — DEFINE THE TRADE SPACE
Ask which variables I can trade and which are non-negotiable:
* Travel dates.
* Departure and arrival airports.
* Number of stops.
* Connection duration.
* Overnight or red-eye travel.
* Separate tickets or self-transfers.
* Basic-economy restrictions.
* Checked and carry-on baggage.
* Seat selection.
* Schedule-change flexibility.
* Total door-to-door travel time.
Ask me to assign either a Low, Medium, or High penalty to each acceptable compromise. If I can do so, also ask me for my personal cash value of one additional travel hour and any fixed penalty I want assigned to a red-eye, self-transfer, distant airport, or highly restrictive fare. Do not invent a value I do not provide.
STAGE 3 — COLLECT THE LIVE SNAPSHOT
Ask me to paste several live options. For each option, collect:
* Total ticket price for the party.
* Date and time of the observation.
* Airline and booking source.
* Fare class.
* Departure and arrival airports.
* Departure and arrival times.
* Stops, connection details, and total duration.
* Included bags and seats.
* Verified charges for required bags and seats.
* Change and cancellation restrictions.
* Ground-transport cost and travel time.
* Any overnight hotel, meal, or other unavoidable cost created by the itinerary.
* Any low, typical, high, historical-range, or timing insight displayed by the live search tool.
Mark missing values Unknown. Do not estimate them.
STAGE 4 — NORMALIZE AND COMPARE
For every option, calculate:
1. Ticket cost.
2. Required ancillary cost.
3. Ground-transport and unavoidable connection cost.
4. All-in cash cost for the party.
5. Door-to-door travel time.
6. Budget variance against the airfare allocation.
7. Dollars saved or spent compared with the lowest all-in viable option.
8. Extra travel time compared with the fastest viable option.
9. Dollars saved per additional travel hour.
10. A friction-adjusted cost only when I supplied the required cash values. Show the formula and keep the actual all-in cash cost visible beside it.
Reject any option that violates a non-negotiable. Do not allow a high score in another category to compensate for a true boundary violation.
Then produce three rankings:
* CASH-MINIMIZING: the lowest verified all-in cost among viable options.
* BALANCED: the strongest compromise between cost, schedule, and restrictions using my stated penalties.
* TIME-PROTECTING: the viable option that preserves the most usable vacation time.
For each ranking, explain the exact trade being made. State the exchange rate in plain language, such as paying more to remove travel hours or saving money by accepting a restriction. Use only values calculated from my data.
STAGE 5 — MAKE THE BOOK-OR-HOLD DECISION
Ask me for a target all-in airfare, a walk-away airfare, and the maximum increase I am willing to risk while waiting. If I do not have them, derive provisional thresholds only from my supplied budget, live route indicators, and current options. Label them Provisional.
Run scenario tests using increases or decreases I specify. Do not attach probabilities unless I provide them.
Recommend BOOK when a viable option meets my target or when the remaining possible improvement is not worth the search time and waiting risk I said I would accept.
Recommend HOLD when a viable option has not reached my target, the walk-away threshold has not been crossed, and time remains before my booking deadline. Provide the next review date and the exact booking trigger. Do not say the fare is likely to fall.
Recommend REVISE when no viable option fits the airfare allocation. Identify which assumption should be reconsidered first: dates, airport, comfort preference, trip duration, or the Week 2 destination. Do not quietly raise the Week 1 budget ceiling.
Finish with:
* Recommended strategy.
* Recommended live option or next search.
* Verified all-in party cost.
* Main sacrifice accepted.
* Main sacrifice rejected.
* Book, Hold, or Revise.
* Next action and deadline.
* A concise decision statement I can send to the other travelers.
How the AI reads this prompt
Practical examples from different industries
Illustrative example — A couple with different definitions of convenience:
One traveler dislikes early departures, while the other is willing to wake before dawn to protect the vacation budget. They paste several live itineraries and assign a High penalty to red-eyes, a Medium penalty to distant airports, and a Low penalty to one ordinary connection. The prompt calculates the all-in cost and extra door-to-door time of each option. Its Cash-Minimizing strategy favors the early connecting flight, its Time-Protecting strategy favors the nonstop, and its Balanced strategy shows precisely how much more the couple would pay to recover the lost travel time. The argument shifts from “this flight feels too expensive” to “is this amount worth these hours?”
Illustrative example — A family choosing between its home airport and a regional alternative:
The family can drive to a second airport, but doing so requires parking, fuel, and a longer return trip after vacation. They supply those verified costs and assign a High penalty to arriving home late before the children return to school. The headline fare at the regional airport is lower, yet the all-in comparison reveals that much of the saving disappears after ground transportation and seat requirements. The prompt’s Balanced ranking selects the home-airport itinerary, while the Cash-Minimizing ranking remains available for the family to consider if protecting the budget matters more than the longer travel day.
Illustrative example — A solo traveler with flexible departure dates:
A remote worker can leave on several weekdays and is willing to use nearby airports, but must retain the ability to change the return. She pastes live searches for multiple date pairs and fare classes, then places a High penalty on restrictive tickets and a Low penalty on afternoon connections. The analysis removes a cheaper basic-economy option because its limitations conflict with her stated requirement. It then calculates which date shift creates the best verified saving without consuming too much usable vacation time, producing a recommendation based on her flexibility rather than generic advice about which weekday is usually cheapest.
Creative use case ideas
- Compare traveling together with splitting a party across two flights when only some travelers have schedule flexibility.
- Evaluate whether adding a vacation day to use a better routing creates more value than purchasing the convenient peak-date flight.
- Compare a nonstop flight with an overnight connection that requires a hotel and consumes part of the first vacation day.
- Help a community organization select group travel without allowing a cheap itinerary to override accessibility requirements.
- Analyze whether a mixed cash-and-points strategy reduces the real trip cost after separate booking restrictions are considered.
Adaptability tips
Travelers who dislike assigning cash values can use only Low, Medium, and High penalties; the three strategy rankings will still reveal the major compromises. More analytical users can assign values to added hours, missed sleep, airport transfers, and restrictive fares, but those values should be personal and documented. Never import a generic value of time from the internet and treat it as the traveler’s own.
For groups, collect preferences from each traveler separately before creating a shared penalty profile. For fixed-date events, turn the arrival deadline into a non-negotiable rather than a weighted preference. For international itineraries, add border formalities, terminal transfers, separate-ticket baggage handling, minimum connection comfort, and overnight lodging to the trade space.
Pro tips
- Ask the AI to identify dominated options: itineraries that cost more and perform no better on any important criterion.
- Run the analysis twice with different penalty settings to see whether the recommendation is stable or preference-sensitive.
- Add a “break-even question” for every compromise: how much cheaper would the inconvenient option need to become before you would choose it?
- Keep separate fields for cash cost, time cost, and preference penalties so a constructed score never obscures the actual amount payable.
Prerequisites
Bring the Week 1 budget ceiling, the Week 2 destination or shortlist, the viable travel windows, and a current set of live flight options. Decide which travel requirements are genuine boundaries and which are preferences you are willing to exchange for savings. You will get a stronger result if required baggage, seating, ground-transport costs, and fare restrictions have already been verified.
Readers using personal cash values should agree on them before seeing the calculated outcome. Choosing the values after viewing the options invites the travelers to manipulate the model until it endorses the flight they already prefer.
Required tools
A general-purpose conversational AI capable of arithmetic and structured analysis, plus a current fare-search service and the relevant airline checkout pages. A calculator or spreadsheet can be used to independently verify the arithmetic, especially for a large party or a complex multi-city itinerary.
Frequently asked questions
Q: Why not compare only the prices shown in the search results?
A: Advertised airfare for travel sold in the United States generally includes mandatory government taxes and carrier-imposed surcharges, but optional services can still change what a particular traveler pays. Required baggage, seating, ground transportation, and itinerary-created costs therefore belong in the comparison when they apply to the trip. The purpose is not to inflate every option with speculative expenses; it is to add only verified costs the traveler expects to incur. (Department of Transportation)
Q: Is friction-adjusted cost a real price?
A: No. It is a decision metric created from the values you supplied for time and inconvenience. The airline will not charge that amount, which is why the prompt keeps verified all-in cash cost visible beside the adjusted figure. Its purpose is to stop substantial inconvenience from disappearing merely because it does not appear on the payment screen.
Q: What if different travelers assign different penalties?
A: Preserve the disagreement rather than averaging it immediately. Run separate scenarios for the travelers whose preferences differ, identify the options that remain viable in every scenario, and then discuss the remaining exchange. This often reveals that the real disagreement concerns one variable, such as a red-eye or distant airport, rather than the entire flight decision.
Q: Should a cheaper alternative airport always remain in the analysis?
A: Only when it is operationally realistic. Include the complete transfer cost, transfer time, parking or transit requirements, and the consequences of reaching that airport during the proposed hours. Google Flights can surface alternative-airport fares, but the traveler must still evaluate the ground journey and itinerary fit. (Google Help)
Recommended follow-up prompts
- “Run a sensitivity analysis showing which preference or cost assumption would have to change before another flight becomes the winner.”
- “Turn the selected flight into a final group-approval message that states the cost, schedule, accepted compromise, and booking deadline.”
- “Carry the confirmed airports, arrival time, and departure time into a Week 4 lodging search-radius decision.”
Tags and categories
Tags:
airfare strategy, flight comparison, opportunity cost, travel time, nearby airports, basic economy, total trip cost, intermediate prompts
Categories:
Travel Planning, Decision Analysis
Citations
- Google Travel Help, “How to find the best fares with Google Flights.” (Google Help)
- Google Travel Help, “Find plane tickets on Google Flights.” (Google Help)
- U.S. Department of Transportation, “Buying a Ticket.” (Department of Transportation)
The Fare Decision Policy
Build a reusable airfare policy that knows when to stop.
Advanced airfare planning is not about discovering the secret day when airlines release cheap tickets. It is about writing a policy before a changing price can manipulate your judgment. A strong policy defines the route baseline you will accept as evidence, the all-in target that earns an immediate booking, the walk-away threshold that forces a larger trip decision, the conditions under which waiting remains rational, and the moment research must stop. The Fare Decision Policy converts live observations into a repeatable control system. It is designed for travelers who want to reduce both financial regret and the endless attention cost of checking the same route again.
Why this matters now
Live fare tools can supply useful evidence, including date comparisons, alternative airports, historical context, and tracked-price alerts. They are not complete representations of the entire market, however: Google states that some airlines or options may not appear because participation depends on carrier and booking-partner relationships. A robust policy therefore records the sources searched, the time of each observation, and the limits of the available evidence rather than treating one screen as universal truth. (Google Help)
The regulatory environment can also change. In July 2026, the U.S. Department of Transportation formally restored the ancillary-fee disclosure regulations that existed before its vacated 2024 rule, reinforcing the practical need to verify current passenger-specific optional costs at the booking source rather than assuming every comparison interface exposes them identically. (Department of Transportation)
Act as a decision-science analyst helping me build and execute a reusable Fare Decision Policy. You do not have reliable access to live fares. Do not state a current route price, recall what this route usually costs, prescribe a universal booking window, or predict a specific future fare. Every numerical fare, fee, historical range, trend indicator, or probability must come from data I provide and must retain its source and observation date.
Your job is to transform my live evidence into a policy and a current decision.
PHASE 1 — INHERIT THE VACATION CONTROL LIMITS
Collect:
* Destination or final shortlist from the destination-selection stage.
* Validated total trip budget ceiling from the budget stage.
* Maximum airfare allocation for the entire party.
* Number and needs of travelers.
* Required arrival and departure boundaries.
* Acceptable airports, routings, cabin types, and fare restrictions.
* Conditions that would require revisiting the destination instead of raising the budget.
* Date after which lodging or itinerary planning can no longer wait for flight confirmation.
Treat these as inherited controls. Do not relax them unless I explicitly authorize reopening an earlier decision.
PHASE 2 — CREATE THE EVIDENCE REGISTER
Ask me to provide dated live observations. For every observation, record:
* Observation date and time.
* Search source.
* Route and travel dates.
* Number of travelers.
* Complete itinerary.
* Fare class.
* Ticket price.
* Required baggage and seat charges.
* Ground-transport and unavoidable connection costs.
* Verified all-in party cost.
* Refund, change, credit, and cancellation conditions.
* Availability notes.
* Any route-specific low, typical, high, historical-range, price-tracking, or timing information displayed by the source.
* Any known coverage limitation, such as an airline checked separately.
Never merge observations from different dates as though they were simultaneous. Never infer a missing historical price.
PHASE 3 — DEFINE THE DECISION POLICY BEFORE SCORING OPTIONS
Help me set these policy fields:
* TARGET: an all-in price at which I will book any compliant itinerary without further optimization.
* RESERVATION RANGE: the area above the target in which I may book after considering schedule quality and waiting risk.
* WALK-AWAY: the maximum all-in airfare permitted by the trip plan.
* BOOKING DEADLINE: the final date for an airfare decision.
* REVIEW CADENCE: how often I will inspect live results before the deadline.
* IMPROVEMENT THRESHOLD: the minimum saving required to justify another review cycle or a less convenient itinerary.
* ATTENTION COST: the value I assign to another search and comparison cycle.
* WAITING-RISK BUDGET: the maximum fare increase or loss of itinerary quality I am willing to accept by delaying.
* NON-NEGOTIABLES: conditions no saving may override.
* ACCEPTABLE CONCESSIONS: trade-offs I will accept, with explicit limits.
* REJECTED CONCESSIONS: trade-offs I will not accept.
* REGRET PRIORITY: whether I care more about overpaying modestly or losing a compliant itinerary.
If I cannot supply a target, derive a Provisional Target only from my airfare allocation, verified live options, and route-specific evidence I pasted. Explain the derivation. Do not use your remembered route knowledge.
PHASE 4 — NORMALIZE AND FILTER THE CURRENT OPTIONS
For each live option:
1. Calculate the verified all-in party cost.
2. Test every non-negotiable.
3. Measure budget headroom or overrun.
4. Identify the concessions used.
5. Identify any Unknown field that could alter viability.
6. Compare total travel time and usable vacation time.
7. Calculate the saving per additional travel hour against the fastest compliant option.
8. Calculate the premium per travel hour saved against the lowest-cost compliant option.
9. Classify the option as Dominated, Efficient, or Incomplete.
An option is Dominated when another verified option is no more expensive and is equal or better on every criterion I marked important.
An option is Efficient when no other verified option improves one important criterion without worsening another.
An option is Incomplete when missing information could change its classification.
Do not rank an Incomplete option above a verified option.
PHASE 5 — EXECUTE THE BOOK, HOLD, REVISE, OR STOP RULE
Recommend BOOK when any of these conditions is met:
* A compliant option is at or below the Target.
* A compliant option lies within the Reservation Range and the cost of waiting exceeds my Waiting-Risk Budget under at least one scenario I supplied.
* The current option is Efficient, the Booking Deadline is near, and no remaining search can plausibly meet my Improvement Threshold without violating a policy limit.
* My Regret Priority favors protecting the itinerary and the current option is within the authorized airfare allocation.
Recommend HOLD only when all of these conditions are met:
* Time remains before the Booking Deadline.
* At least one compliant current option remains below the Walk-Away threshold.
* I have defined a specific improvement worth waiting for.
* I accept the stated Waiting-Risk Budget.
* A next review date and automatic booking trigger can be written.
Do not attach a probability to a future fare unless I supplied that probability from an identified source. Use scenario analysis instead: show what happens if the all-in cost improves, remains unchanged, increases within my risk budget, or increases beyond it.
Recommend REVISE when no compliant itinerary fits below the Walk-Away threshold. Identify the least damaging upstream assumption to reconsider: travel date, airport, trip length, comfort constraint, destination shortlist, or another discretionary trip cost. Never raise the validated total trip budget without explicit authorization.
Recommend STOP OPTIMIZING when any of these conditions is met:
* A compliant option meets the Target.
* The maximum remaining saving I am willing to pursue is smaller than the Attention Cost plus the Waiting-Risk Budget.
* The Booking Deadline has arrived.
* Further searches are producing no new viable routing, date, airport, or fare-class information.
* The selected flight is already booked and no verified cancellation or rebooking path offers a meaningful net benefit after all costs and risks.
PHASE 6 — PRODUCE THE DECISION PACKAGE
Return:
1. Evidence Quality — what was verified, missing, stale, or limited.
2. Policy Summary — Target, Reservation Range, Walk-Away, Booking Deadline, Review Cadence, Improvement Threshold, Attention Cost, and Waiting-Risk Budget.
3. Trade-Off Boundaries — accepted and rejected concessions.
4. Efficient Options — only the non-dominated verified choices.
5. Scenario Analysis — no unsupported probabilities.
6. Decision — BOOK, HOLD, REVISE, or STOP OPTIMIZING.
7. Rationale — the three strongest reasons.
8. Reversal Conditions — the exact new facts that would justify changing the decision.
9. Next Action — one action, one owner, and one deadline.
10. Decision Record — a compact, machine-readable block with fields for observation date, selected option, verified all-in cost, decision, policy trigger, accepted trade-off, rejected trade-off, next review date, booking deadline, and evidence sources.
Before finishing, audit your own response for any fare, fee, probability, route history, or prediction that did not come from my supplied evidence. Remove or mark any unsupported item Unknown.
How the AI reads this prompt
Practical examples from different industries
Illustrative example — A family planning around school holidays:
The family has fixed travel dates, a validated total budget, and a destination chosen in Week 2\. They collect several dated fare snapshots and record the route-specific indicators displayed by their live search source. Before evaluating the flights, they set a Target, a higher Walk-Away boundary, a booking deadline tied to lodging availability, and a rule rejecting overnight self-transfers. The policy identifies two Efficient options: a lower-cost connecting itinerary and a more convenient nonstop. Because the nonstop enters the family’s precommitted Reservation Range as the deadline approaches, the framework recommends booking and stopping rather than spending another week attempting to capture a smaller hypothetical saving.
Illustrative example — A couple traveling to a fixed-date international event:
The travelers must arrive with enough buffer before the event and will check baggage. Their evidence register separates airline-direct results from metasearch observations and records that one carrier required a separate search. A low headline fare is marked Incomplete until baggage, terminal transfer, and separate-ticket protections are verified. After verification, the option becomes Dominated by a slightly different routing with a lower all-in cost and fewer operational risks. The policy recommends the Efficient alternative and records the exact evidence that would permit reconsideration, preventing the couple from reopening the decision whenever a new advertisement appears.
Illustrative example — A flexible solo traveler optimizing without a deadline:
A solo traveler can shift dates and airports, which creates an almost unlimited search space. He sets an Improvement Threshold representing the minimum saving that would matter, an Attention Cost for each complete comparison cycle, and a maximum review cadence. The policy initially recommends HOLD because no option has reached the Target and meaningful date variations remain unexplored. After later searches stop producing new viable structures and the gap between the best compliant fare and the Target becomes smaller than the traveler’s combined search cost and waiting-risk budget, the model recommends STOP OPTIMIZING and creates a final decision record.
Creative use case ideas
- Maintain a shared family fare policy so one person can monitor prices without renegotiating everyone’s preferences after every alert.
- Create separate policies for outbound and return flights when a mixed-carrier or open-jaw itinerary is under consideration.
- Audit an employer-reimbursed personal trip where the traveler may pay extra for convenience but needs the reimbursable baseline documented.
- Apply the evidence register to award availability, preserving points, cash surcharges, transfer requirements, and cancellation rules as separate fields.
- Use the stopping rule for a honeymoon, reunion, or graduation trip where protecting a scarce itinerary matters more than obtaining the theoretical lowest fare.
Adaptability tips
A lighter advanced version can omit the constructed Attention Cost while retaining the Target, Walk-Away, Booking Deadline, and stopping rule. A more technical user can add a structured sensitivity analysis that varies personal values for time and inconvenience, but the model should never manufacture fare probabilities. The policy can also be divided into two documents: a stable policy created before searching and a dated execution record generated from each live snapshot.
For a group, designate who may authorize a constraint change and who owns the next search. For multi-city travel, create a policy for the complete itinerary rather than optimizing each leg independently and accidentally producing incompatible tickets. For award travel, replace the single Target with separate limits for points, cash surcharges, and the opportunity cost value supplied by the traveler.
Pro tips
- Freeze the policy fields before pasting the current options, then ask the AI to flag any threshold that appears to have been chosen after the fact.
- Require every calculated value to point back to an evidence-register entry.
- Keep an explicit Stale status for observations that no longer represent a bookable option.
- After booking, permit reconsideration only when a verified rebooking path exceeds the Improvement Threshold after cancellation consequences, fare differences, and attention costs are included.
Prerequisites
Bring the validated Week 1 budget ceiling, Week 2 destination decision, complete traveler requirements, and enough live observations to distinguish an isolated fare from a repeatable pattern. At minimum, record where and when each observation was collected. Advanced users should decide their target, walk-away threshold, booking deadline, review cadence, and acceptable waiting risk before asking the model to score the current options.
This framework is most valuable when a traveler is genuinely prepared to follow the policy. A threshold that will be ignored as soon as an emotionally appealing flight appears is documentation, not control.
Required tools
A general-purpose conversational AI capable of structured analysis and arithmetic, a current fare-search service, and direct access to airline or booking-provider terms. A spreadsheet or note system is recommended for maintaining dated evidence, but the prompt can produce a reusable text-based decision record without additional software.
Frequently asked questions
Q: How many fare observations are enough to build a baseline?
A: There is no universal count that makes a route baseline reliable. The observations must be comparable in route, dates, cabin, party size, baggage assumptions, and itinerary quality. Treat a small evidence set as provisional, preserve the source and time of each observation, and rely more heavily on the budget and deadline controls when historical evidence is thin.
Q: Can the AI assign probabilities to fare increases and decreases?
A: Not unless you provide those probabilities from an identified source and understand their limitations. A route-specific tool may display a directional signal or confidence statement, which the evidence register can preserve, but the model should not convert that into an invented numerical probability. Scenario analysis is safer because it shows the consequence of each outcome without pretending to know which one will occur.
Q: Is one fare-search platform enough?
A: It can be enough for an initial snapshot, but it should not automatically be treated as complete coverage. Google states that not every airline or available flight necessarily appears in its results because inclusion depends on participating partners. For an important or constrained trip, record which major carriers or booking sources were checked separately. (Google Help)
Q: Why verify ancillary fees directly when comparison tools already display them?
A: Optional charges can depend on fare class, passenger status, loyalty benefits, credit cards, itinerary, and booking channel. The applicable U.S. disclosure framework also changed in July 2026 when DOT implemented the court’s vacatur of its 2024 ancillary-fee rule and restored the prior regulatory standards. The robust practice is therefore to record fees shown by the search tool but verify the traveler-specific amount before treating the option as complete. (Department of Transportation)
Q: Does STOP OPTIMIZING mean the selected fare was the lowest possible fare?
A: No. It means the selected option satisfies the policy and that further searching no longer earns enough expected decision value to justify its attention cost and waiting exposure. The lowest possible fare is unknowable in advance and unnecessary for a defensible decision. A successful policy aims for a good compliant outcome, not retrospective perfection.
Recommended follow-up prompts
- “Convert this Fare Decision Policy into a reusable blank template for future trips.”
- “Audit my completed decision record for unsupported assumptions, stale observations, and threshold changes made after seeing the fares.”
- “Pass the confirmed airports, routing, arrival window, departure window, and remaining trip budget into the Week 4 lodging framework.”
Tags and categories
Tags:
airfare policy, decision science, fare monitoring, stopping rule, route baseline, risk management, travel optimization, advanced prompts, evidence provenance
Categories:
Travel Planning, Decision Systems
Citations
- Google Travel Help, “Track flights & prices.” (Google Help)
- Google Travel Help, “Find plane tickets on Google Flights.” (Google Help)
- U.S. Department of Transportation, “Increasing Flexibility on Disclosure of Airline Ancillary Fees,” final rule issued July 2026\. (Department of Transportation)
Which of the three should you use?
The Fare Reality Check is the fastest route from confusion to action. It asks for a small set of live options, calculates verified all-in cost, applies the traveler’s non-negotiables, and finishes with BOOK, HOLD, or SEARCH AGAIN. It is the right choice when the traveler has already found realistic flights and mainly needs protection from hidden fees, incomplete comparisons, and indecision.
The Total-Cost Trade Space is for travelers who have several viable options but disagree about what “best” means. It exposes the exchange rate between money, travel time, schedule quality, airport convenience, and fare restrictions. Its three strategies prevent a subjective compromise from being presented as a universal winner: the traveler can deliberately choose to minimize cash, balance competing priorities, or protect vacation time.
The Fare Decision Policy is designed for repeated monitoring and higher-stakes trips. It preserves evidence provenance, inherits the budget and destination decisions from earlier weeks, sets target and walk-away boundaries before scoring flights, separates efficient options from dominated or incomplete ones, and makes STOP OPTIMIZING an explicit outcome. Choose it when fares will be watched across multiple sessions, several people must remain aligned, or the cost of reopening settled decisions has become part of the problem.
TAGS:
