“Which AI Model Should I Use?” Decision Prompt

Picking the right AI model shouldn’t require reading five comparison articles. This prompt turns the decision into a short back-and-forth: you describe the task, the prompt asks a few clarifying questions, then gives you a specific recommendation with reasoning — instead of a generic “it depends.”

What this prompt does

It structures the model you’re talking to as a brief decision consultant: gather constraints first (task, budget, priorities), then recommend. This works in any chat interface — paste it as your first message.

The prompt template

Copy this and fill in the bracketed variables, or leave them blank and let the model ask you directly:

You are helping me choose the right AI model/subscription for a specific task. Before recommending anything, ask me 2-4 clarifying questions about:
- The primary task: [TASK_DESCRIPTION]
- My monthly budget: [BUDGET, e.g. "$20/month" or "flexible"]
- What matters most to me: [PRIORITY, e.g. "accuracy over speed" or "lowest cost"]
- How often I'll use it: [FREQUENCY, e.g. "daily" or "occasionally"]

Once you have enough information, give me:
1. A specific recommendation (name the model/plan)
2. Two sentences explaining why it fits my stated priorities
3. One reasonable alternative and when I'd want it instead
4. Any caveat I should know before subscribing (usage limits, features NOT included, etc.)

Do not recommend based on general reputation alone — tie the recommendation directly back to what I told you about the task and priorities.

Variables explained

Example run-through

Input: “I write long-form blog content daily and need something that keeps a consistent voice across a 3,000-word draft. Budget is $20/month, and accuracy/tone matters more than speed.”

What a good response looks like: a named recommendation (e.g. Claude Pro or ChatGPT Plus), two sentences tying it to “long-form” and “consistent voice,” one alternative (e.g. “if you also need image generation, consider X instead”), and a caveat about usage limits at that price point.

Variations

Role-prompting version

Add this line to the top: “Act as an independent AI procurement advisor with no preference for any single provider.” This nudges the model away from defaulting to whichever assistant you’re currently talking to.

Few-shot version (for teams)

Before the main prompt, paste 1-2 examples of past task-to-tool decisions your team already made and liked. This anchors the recommendation style to your team’s actual preferences instead of generic advice.

Optimization tips

• Fill in as many variables as you can up front — fewer clarifying rounds means a faster, sharper answer.
• If the recommendation feels generic, ask a follow-up: “What would change your recommendation if my budget doubled?”
• Re-run this prompt every quarter — model line-ups shift often enough that a six-month-old answer may be stale.

Debugging: if the answer is too generic

Vague answers usually mean the task description was too broad. Replace abstract descriptions (“help with writing”) with a concrete example of the actual output you need (“draft a 500-word product update email in a friendly but professional tone”).

Best practices

• Treat the output as a starting point, not gospel — cross-check against our model selection guide.
• Run the same prompt in two different assistants and compare — disagreement between them is often more informative than either answer alone.
• Pair this with the Subscription ROI Calculator to sanity-check the cost side of any recommendation.