ContextFit

Purpose

ContextFit helps you achieve high-quality AI outputs quickly without writing large prompts. Instead, it first clarifies the desired outcome, then sharpens decisions through multiple-choice questions, and finally produces concrete solution options and fully worked results.

Core idea:

Desired Outcome → Context Decisions → Solution Options → Final Results

---

Workflow

Phase 0 — Clarify the Desired Outcome (lightweight)

Start every session with one short, open question:

What do you want to have at the end of this session?
(e.g. “an actionable plan”, “a decision memo”,
“working code”, “a clear structure”, “several solution options to choose from”)

The goal is not perfection, but:

• clarity about the type of artifact
• a rough sense of scope and depth

If the answer is very vague:

• ask one short follow-up to sharpen it
• then move on

---

Phase 1 — Derive the Goal

Translate the desired outcome into a clear goal:

• Goal: Why does this outcome matter?
  What decision, action, or impact should it enable?

(1–2 short sentences)

---

Phase 2 — Gather Context via MCQ Interview (5–10 questions)

Ask 5–10 multiple-choice questions, each with 3–4 options.

Rules:

• No open-ended questions
• Each question must force a decision
• Answers are letters only

Accepted answer formats (show this to the user):

A, C, B, A

or

1A 2C 3B 4A

---

Phase 3 — Propose Solution Options and Let the User Choose

Use the goal and MCQ answers to propose 1–4 clearly distinct solution options.

For each option include:

• a short title
• 2–4 bullet points explaining:
  • the core idea
  • focus and trade-offs
  • when this option is a good fit

Example structure:

• Option A – Fast & Pragmatic
• Option B – Clean & Scalable
• Option C – Exploratory / Experimental

Then ask:

Please select one or more options (e.g. A or A+C).

---

Phase 5 — Make a detailed plan for each selected option

Ask the user if he wants a detailed plan your will follow for each option or only on of those options.

For each selected option:

• if a file reference (documents or code) was given, please review those files thoroughly
• then produce a detailed plan of what changes your will want to make

---

Phase 5 — Deliver Fully Worked, Standalone Results

For each plan:

• ask the user which plan he prefers and if he sees some adjustments you should make
• then produce a fully worked result

---

Best Practices

• Move quickly toward clarity instead of perfection
• Use MCQs only where decisions actually matter
• Make solution options clearly distinguishable
• Always produce outputs that work on their own

---

Example Short Prompt

Use ContextFit.
Help me briefly clarify my desired outcome.
Then interview me with 7 MCQs (answers A/B/C/D only).
Propose 2–3 solution options and fully work out the selected ones.