Don't Ask AI to Choose: Use This Prompt to Compare Two Options More Clearly

Choosing between two options often feels difficult not because the information is lacking, but because everything is mixed up in our heads. Whether to use an old laptop or buy a new one, choosing software A or B, accepting an additional project or declining it—each choice has different benefits, costs, risks, and consequences.

In situations like this, AI can help not as a 'judge' that directly makes the choice, but as a tool to organize our thinking. The key is not to just ask, 'Which is better?' That question is too broad and usually results in safe answers that end up with 'it depends on your needs.'

It is more useful if we ask AI to compare options based on clear criteria, highlight unknowns, and suggest ways to test assumptions before making a decision.

Why the question 'Which is the best?' is often unhelpful

A decision rarely has one best answer for everyone. A high-spec laptop may be suitable for design work, but too expensive for someone who only uses it for typing and online meetings. An application with comprehensive features can be useful for large teams, but can complicate things for small businesses.

If context is not provided, AI will fill that gap with assumptions. The problem is, those assumptions may not align with our conditions. Answers that sound reasonable can also be misleading if priorities, budgets, technical capabilities, or deadlines are not explained.

Therefore, a good prompt needs to force the comparison process to be more structured. AI should be asked to:

• identify relevant criteria;
• separate facts from assumptions;
• explain trade-offs or consequences between options;
• mention information that is still lacking;
• provide conditional recommendations, not absolute decisions.

Prompt template that can be used directly

Here is a template that can be used to compare two options. Replace the parts in square brackets as needed.

I am comparing two choices: [Option A] and [Option B].

My context:

- Main goal: [explain the outcome you want to achieve]

- Budget or constraints: [fill in if any]

- Time for use or decision: [fill in]

- My/team's technical capabilities: [beginner, intermediate, expert]

- What is most important to me: [for example cost, speed, ease, security, flexibility]

Please help me with the following steps:

1. Identify 5-7 most relevant criteria for comparing both options.

2. Create a comparison table based on those criteria.

3. Mark which are facts, assumptions, and information that still needs to be verified.

4. Explain the main trade-offs of each option.

5. Provide conditional recommendations: when Option A is more suitable and when Option B is more suitable.

6. Mention three important questions I need to answer before making a decision.

This structure prevents AI from jumping to conclusions immediately. It is asked to build a decision map first.

Example: choosing a task management app

For example, we are choosing between a simple task management app and a more comprehensive project management app. A prompt that is too short might say, 'Which app is better for organizing work?'

This prompt does not explain the number of team members, the type of work, the frequency of collaboration, reporting needs, or budget. As a result, the answer that comes up may sound general and hard to apply.

With the above template, the context can be written like this:

I work in a team of four. We manage weekly content projects, often share documents, and need deadline reminders. Some team members are not familiar with complex software. Our priorities are ease of use, task visibility, and low cost. Compare a simple task management app with a comprehensive project management app.

A good answer should not only mention features. AI needs to explain that a simple app may be quicker to use and easier for the team to adopt, while a comprehensive app may be stronger for task dependencies, progress tracking, and reporting—but requires more time to learn and more disciplined management.

Add weights so the comparison doesn't feel flat

Not all criteria have the same level of importance. Cost may be important, but ease of use could be far more important because the app will be used every day. To make priorities clearer, add weights.

Use the following weights:

- Ease of use: 30%

- Cost: 25%

- Collaboration features: 25%

- Flexibility: 10%

- Ease of migration: 10%

Score each option 1-5 on each criterion. Explain the reason for each score and calculate the weighted value. Do not treat scores as absolute facts; show the parts that are still estimates.

Weighted assessments are not a math that automatically makes decisions correct. Their function is to openly show our priorities. If the results change drastically just because the weight of cost is slightly increased, that is a sign that the decision is sensitive to the budget and needs to be thought through more carefully.

Don't forget to ask for 'conditions that invalidate the recommendation'

One of the most useful additions is to ask AI to mention conditions that could make its recommendations invalid. This helps us see the weaknesses in the analysis, not just the reasons supporting a particular choice.

After providing recommendations, explain:

- What assumptions most affect the outcome?

- What information, if changed, could reverse the recommendation?

- What is the biggest risk of following that recommendation?

- What is the cheapest and quickest way to test the recommendation before committing?

The last part is important because many decisions do not have to be permanent right away. We can use a trial period, create a small project, request a demo, or test one feature first.

What does this mean for us?

AI is most useful in decision-making when it helps clarify our thinking, not when we hand over the decision entirely. AI's answers can still be wrong, especially if the data used is incomplete or our context is too general.

Use the comparison results as a basis for discussion and initial review. For decisions involving large sums of money, security, health, legal matters, or many people's jobs, verify important information from relevant sources and experienced parties.

What you can do now

1. Choose one small decision that is pending.
2. Write down two options and the main goal you want to achieve.
3. Include real constraints such as budget, time, capabilities, and number of users.
4. Use the comparison template above.
5. Check if the recommendations change when priorities or assumptions are altered.
6. If possible, conduct a small trial before making a final decision.

In this way, AI is not used to find the answer that sounds most convincing. AI is used to sharpen our questions, make choices clearer, and make decisions more accountable.

– Rio Yotto @rioyotto