You ask AI for help with your business and receive ten suggestions. Several sound useful. You can imagine trying them. Then you look at your calendar and realise you have time for one.
At that point, another list of ideas may add to the work. A more useful next step is to compare two realistic options against the result you want and the limits you are working within.
First, remove ideas that break your firm limits. Then shortlist two that directly address the problem.
AI can help you organise that comparison. Your role is to supply the business facts, challenge the assumptions and decide what you are willing to try. Start with a small decision you can review and change.
Name the decision before comparing the ideas
Write one sentence that describes the choice. For example: “I have one hour this week to make booking enquiries easier to handle. Should I create a short explanation video or a simple frequently asked questions page?”
This is an illustrative decision for a fictional service business. The owner already has an audience asking about available services and how to book. Both ideas could help, but the immediate goal is clearer enquiries, and the owner has limited time.
Notice how much the goal matters. A video might be an appealing idea, but “it could reach more people” does not settle a decision about helping existing customers understand the booking process. Keep the comparison tied to the job you actually need done.
Give both options the same test
Choose a few criteria that could change your decision. Four useful starting points are:
- Fit: How directly could this option help with the stated problem?
- Effort: What must I prepare, produce and check before it is usable?
- Missing information: What do I still need to know?
- Upkeep: What will need updating after the first version?
Add a firm limit if you have one, such as using only tools you already own or staying within a set amount of time. Separate those limits from preferences. An option that depends on unavailable equipment is not ready to start, however attractive the idea sounds.
For the service business, the video requires recording, editing and accurate spoken details. The FAQ page requires clear writing, somewhere suitable to publish it and a way for customers to find it. Both need checking. Neither becomes useful merely because a draft exists.
Ask AI to show its assumptions
A polished comparison can make guesses look settled. Watch for statements such as “customers prefer video”, “this will take fifteen minutes” or “this will generate more bookings”. Without relevant evidence, those remain assumptions or estimates.
Ask the assistant to identify what each conclusion rests on. Does it come from information you supplied, a source you can inspect or an assumption it made? If it uses outside research, check that the source actually supports the claim and applies to your situation.
You can also ask: “What missing fact would be most likely to change your recommendation?” In the example, that might be whether customers already struggle to find an existing FAQ, or whether the owner has a useful video ready to edit. One answer could change the practical choice.
Use a comparison prompt you can challenge
Try this with a low-risk business decision:
“Help me compare two options: [option A] and [option B]. My goal is [specific result]. My confirmed constraints are [time, budget, tools and other limits]. The facts I can supply are [relevant information]. Compare both options using the same criteria: fit, effort, missing information and upkeep. Separate facts from estimates and assumptions. Ask about any missing fact that could materially change the choice. Do not invent customer preferences or predict sales. Suggest one small test, and explain what evidence would make you revise the recommendation. Prepare advice only; do not publish, spend money or contact anyone.”
Replace the brackets with facts you know. Leave uncertain information marked as uncertain. Keep customer identities, confidential records and unnecessary personal details out of the exercise.
Read the answer as something to examine. Asking AI to challenge its first recommendation may reveal another angle, but that second answer still needs your judgment and any relevant independent checks.
Choose a small test with an observable result
If the fictional owner already has a booking page they can edit, adding answers to the three questions customers ask most often could make a manageable first test. They could draft the answers, then review them with a colleague before publishing. The colleague could check whether each answer is easy to understand and matches the actual booking process.
Before changing anything, note the recurring questions in recent enquiries. Choose a review point, such as the next five relevant enquiries after publication, and record whether people saw the information and what still confused them. A small number of enquiries will not prove that the page increases sales. It can still reveal a broken link, an unclear instruction or a missing answer worth fixing.
If the owner chooses the video instead, the same discipline applies. Make a manageable first version, check its accuracy and choose a practical way to assess whether people understand it. Avoid turning the first attempt into a larger production than the original problem needs.
Keep the decision yours
A comparison is useful when it makes the trade-offs visible. You may accept the suggested option, choose the other one or decide that a missing fact needs checking first.
For decisions involving significant money, legal obligations, health, employment or other serious consequences, a quick comparison exercise is not enough. Use appropriate professional advice and stronger evidence before acting.
For today's practice, choose one ordinary business decision and finish this sentence: “I am choosing between ___ and ___. The result I need is ___. My main constraint is ___.”
Ask for the comparison, inspect the assumptions and choose one small next step. The useful outcome is a decision you can explain and a test you can learn from.