How I'd Use ChatGPT Before Buying a Laptop, Grill, or Drone
Buying a laptop should not require a second laptop just to hold all the comparison tabs.
The same thing happens with grills, drones, cameras, appliances, and almost anything expensive enough to make me nervous. I start with one question, open twelve reviews, discover six model numbers that look nearly identical, and eventually forget which problem I was trying to solve.
This is one of the everyday jobs where AI might actually earn its keep.
OpenAI's shopping-research guide describes an interactive process that asks about needs, preferences, trade-offs, and budget, then builds a buyer's guide with a smaller group of products to compare.
That sounds useful. It is not the same as handing ChatGPT my credit card and hoping for the best.
I have not run these exact laptop, grill, and drone tests as completed purchases, so this is the workflow I would use—not a story about money I supposedly saved. My rule would be simple:
Trust AI to narrow the list and explain the trade-offs. Verify the final buying details yourself.
TLDR
- Use shopping research when the purchase has several reasonable options, competing trade-offs, or a firm budget.
- Describe the job the product needs to do before asking for a winner.
- Answer the follow-up questions; they are where a vague search becomes a useful comparison.
- Ask for a small shortlist with a clear reason, downside, and “best for” label for each item.
- Challenge the recommendations before accepting them.
- On the retailer's page, verify the exact model, configuration, final price, stock, return policy, and warranty.
- If the AI gives you three better tabs instead of fifty confusing ones, it did something useful.
What Shopping Research Is Good At
OpenAI says shopping research is designed for decisions involving comparisons, trade-offs, or multiple constraints.
That matters because most buying decisions are not really “Which product is best?”
They are more like:
- Which laptop is light enough for school but still has enough memory to last several years?
- Which grill fits a small patio without making dinner for six feel like a camping emergency?
- Which drone is friendly enough for a beginner but not something I will outgrow in two weekends?
The official guide says ChatGPT can ask follow-up questions about things such as brand, size, performance, comfort, style, and price. It can use public product information and retail sources, show products as it finds them, and let the shopper remove options or change the constraints.
After a few minutes, the result is usually a buyer's guide. OpenAI says that guide can include:
- a plain-English explanation of what matters
- a small group of top picks
- the reason each pick fits
- strengths and trade-offs
- side-by-side comparisons
- links to merchants
- more options that matched
That is the part I want AI to do: turn the product pile into a readable decision.
Start With the Job, Not the Product
The fastest way to get a generic answer is to ask a generic question.
“What is the best laptop?” leaves out the student, the classes, the software, the weight, the budget, and how long the machine needs to last. The best laptop for a video editor is not automatically the best one for a student carrying it across campus all day.
Before I start shopping research, I would write down five things:
- The job: What does this product need to do in real life?
- The budget ceiling: What is the most I am willing to pay before tax?
- The priorities: Which two or three qualities matter most?
- The deal-breakers: What would make an otherwise good option useless?
- The time horizon: Is this a two-year solution or something I expect to keep?
That five-line brief is more valuable than ten clever prompt tricks.
My School-Laptop Starting Prompt
Here is the first prompt I would try:
I am shopping for a Windows laptop for a student who will carry it to school every day. My maximum budget is $800 before tax. The priorities are good real-world battery life, a weight under 3.5 pounds, 16 GB of memory, and enough performance for schoolwork, many browser tabs, video calls, and light photo editing. Gaming is not important. I want it to last about four years. Ask me follow-up questions before recommending anything. Then use shopping research to compare a small group of options, explain the important trade-offs in plain language, and tell me what I should verify on the retailer's page.
The numbers in that prompt are examples, not universal laptop advice. The useful part is the structure.
It identifies:
- who will use the laptop
- the hard budget
- the daily situation
- the important specifications
- the work it needs to handle
- what does not matter
- how long it should remain useful
I would answer the follow-up questions instead of treating them like a speed bump. If ChatGPT asks whether screen size or battery life matters more, that is not wasted time. That is the purchase becoming clearer.
My Five-Step AI-Before-You-Buy Workflow
1. Narrow
I would let shopping research gather options, but I would ask it to finish with no more than three main finalists and one budget alternative.
Too many recommendations put me right back where I started. A shortlist should be short enough that I can open every source and inspect it.
For each finalist, I would ask for:
- why it fits my brief
- its biggest strength
- its most important compromise
- who should not buy it
- the exact model or configuration being compared
2. Explain
Specifications are only useful when they connect to real life.
I would ask:
Explain which differences I will actually notice during a normal week. Translate technical specifications into battery, weight, speed, cleanup, setup, portability, noise, or repair consequences. Skip differences that are unlikely to matter for my use.
For a school laptop, I care less about winning a processor-name argument and more about whether the machine stays responsive during a video call with twelve browser tabs open.
For a grill, the practical question may be whether the cooking area fits the family and whether cleanup will make me avoid using it.
For a drone, it may be whether the full kit includes the batteries and charger I assumed were in the box.
3. Challenge
The first shortlist should not be the final answer.
I would ask ChatGPT to argue against its own recommendations:
What is the strongest reason not to buy each finalist? Which recommendation depends on information that may be old, incomplete, or hard to verify? What change in my priorities would produce a different winner?
That question matters because a confident explanation can still rest on the wrong configuration, an outdated listing, or a trade-off I never considered.
I would also ask it to separate:
- facts found on a product or retailer page
- review-based observations
- assumptions or estimates
Those are not the same kind of evidence.
4. Verify
This is where I stop treating the buyer's guide as the final word and start opening the source links.
OpenAI's own help page says shopping research can make mistakes. It also says prices, stock, and discounts change frequently and may not match the retailer's page.
For every finalist, I would verify:
- exact model number, year, configuration, and included accessories
- current price
- taxes, fees, and shipping
- actual availability
- the correct size or color when that matters
- seller identity
- return window and return conditions
- warranty terms
I would take a screenshot or make a note of the retailer page on the day I planned to buy. Yesterday's sale price is not a binding contract with today's shopping cart.
For safety information, recalls, certifications, or current drone rules, I would also check the manufacturer or the appropriate official authority. A shopping summary is not where I want the final word on a safety or legal question.
5. Buy—or Walk Away
The purpose of research is not to create emotional momentum toward checkout.
If the finalists are all over budget, if the configuration I wanted is unavailable, or if the return policy is terrible, walking away is a valid result.
I would rather have AI help me discover that before I buy than help me rationalize the wrong purchase faster.
The Finalist Card I Would Use
For each product still in the running, I would make one simple card:
- Product: Exact name and model number
- Best for: The job it handles best
- Why it fits: Two sentences maximum
- Main compromise: The downside I am accepting
- Current price: Verified on the retailer page
- Configuration: Memory, storage, size, bundle, fuel type, or kit contents
- Availability: In stock where I can actually buy it
- Return policy: Window, condition, fees, and exceptions
- Warranty: Length and who provides it
- Checked: Date and source links
If I cannot fill in a line, the card is not finished.
This is also a good way to catch a common shopping problem: the review covers one model, the comparison table describes another, and the sale page quietly lists a cheaper configuration with the same family name.
How I Would Change the Brief for a Grill
I would not ask, “What is the best grill?”
I would give ChatGPT the actual yard and cooking job:
I need a grill for a small patio and usually cook for four people, with occasional meals for eight. My budget is $600. I care more about even heat, easy cleanup, and replacement-part availability than smart features. Compare fuel type, cooking area, footprint, setup, maintenance, and warranty. Ask about any constraint I missed before researching.
The follow-up questions might reveal that storage space, apartment or neighborhood rules, fuel availability, or cleanup matters more than a feature I saw in an advertisement.
The final retailer check would still include the exact model, dimensions, included parts, delivery or assembly fees, return conditions, and warranty.
How I Would Change the Brief for a Drone
A drone comparison should start with the kind of flying, not the coolest camera specification.
I would write:
I am comparing beginner-friendly drones for outdoor photography and casual video. I care about a manageable learning curve, a complete starter kit, battery cost, replacement parts, portability, and image quality in normal daylight. My total starting budget is $1,000 including the batteries and charger I need. Ask where and how I plan to fly before making a shortlist.
Then I would verify the package contents, aircraft and controller compatibility, battery type, current firmware or app requirements, seller authorization, warranty, and the current rules that apply where I plan to fly.
The exciting object in the photo is rarely the whole cost of the hobby.
Where I Would Trust It—and Where I Would Slow Down
Green light: use AI freely
- Turn my needs into useful comparison criteria.
- Explain technical language in normal English.
- Keep the same criteria across several products.
- Reduce a long list to a few plausible options.
- Suggest alternatives when a finalist misses a constraint.
- Point me toward source pages worth opening.
Yellow light: treat as a lead
- Review summaries.
- Real-world battery or performance claims.
- Comfort, fit, noise, cleanup, or durability.
- Claims that depend on a specific model year or bundle.
- “Best for most people” conclusions.
These can still help, but I want to know where the claim came from and whether it applies to the exact product I am considering.
Red light: verify before paying
- Final price, tax, fees, and shipping.
- Stock and delivery timing.
- Exact configuration and included accessories.
- Size, color, or seller.
- Return and warranty policies.
- Safety, recall, certification, or regulatory information.
The retailer, manufacturer, or official authority gets the final word on those details.
A Note About Personalization
OpenAI says shopping research may use ChatGPT memory when memory is turned on. That could help if ChatGPT already knows I care about drones, travel weight, or a certain kind of computer work.
It could also carry an old preference into a new decision.
I would scan the starting assumptions and correct anything that no longer fits. If I did not want memory involved, OpenAI says it can be turned off or cleared in personalization settings.
The help page also says chats are not shared with retailers and that ads are separate from shopping research. That is useful to know, but I would still keep unnecessary personal, financial, medical, or account information out of a product brief.
ChatGPT needs my budget. It does not need my bank statement.
The One-Purchase Test
The next time I am considering a purchase expensive enough to research, I would score the process from 0 to 1 on five questions:
- Did the follow-up questions improve my brief?
- Did the shortlist respect my actual budget and deal-breakers?
- Could I explain the trade-offs without rereading the whole guide?
- Did the source links help me verify the finalists?
- Did the retailer-page details match the comparison?
A score of 4 or 5 means shopping research reduced real friction.
A score of 2 or 3 means the idea may be useful, but the prompt or constraints need work.
A score of 0 or 1 means I should close ChatGPT and go back to a normal checklist.
AI does not earn a participation trophy just because it produced a polished table.
My Bottom Line
I do not want ChatGPT to choose what I buy.
I want it to help me understand what I am choosing.
If shopping research can turn fifty confusing tabs into three reasonable finalists, translate the trade-offs, and tell me what still needs verification, that is a practical win.
The last step stays human.
I decide which compromises I can live with. I open the retailer page. I check the exact item. I read the return policy. I make sure the total still fits the budget.
AI can help me reach the decision with less clutter.
It should not make me careless at the finish line.
Source and Review Note
This draft is based on OpenAI's Using shopping research in ChatGPT help page, accessed July 25, 2026. During verification, the page displayed the relative label Updated: 10 days ago.
Shopping research and its interface can change. Product information, prices, discounts, inventory, merchant availability, and retailer policies change even faster. Recheck the OpenAI help page before publication and verify the exact retailer or manufacturer page before buying.