Personal Research Assistant

---
name: personal-research-assistant
description: Research anything well with an AI assistant: frame balanced questions, compare options in tables, demand sources with dates, and spot weak claims. Use when comparing products, checking news, or deciding on health, travel, or tech topics.
category: everyday-assistant
---

# Personal Research Assistant

## Overview

Researching with an AI assistant is a conversation, not a single question. Most people type one
vague question, get one confident-sounding answer, and act on it. That's backwards. Good research is
iterative: start broad, narrow down, compare options side by side, and always ask "says who?"

This skill turns your assistant into a research partner. It covers how to frame questions that get
balanced answers instead of confirmation, how to build comparison tables that make trade-offs
visible, and how to tell a solid claim from a shaky one. The goal isn't to become a professional
researcher — it's to stop making decisions on the first answer you hear.

## When to use

- Comparing products or services (phones, laptops, appliances, insurance plans, schools)
- Checking a news story or health claim you saw online
- Planning a big decision (a move, a course, a job switch) and wanting the full picture
- Fact-checking something a friend told you or a video claimed
- Gathering background before a conversation with a professional (doctor, lawyer, financial advisor)

## Core concepts

- **The assistant is a starting point, not a source.** It can summarize what's known, point you at
  sources, and organize findings — but the actual sources (studies, official docs, reviews) do the
  proving. Treat assistant output as a map, not the territory.
- **Specific questions get specific answers.** "Tell me about vitamin D" invites vague fluff. "What
  are the two main schools of thought on vitamin D supplementation for adults, and what does each side
  cite?" forces balance.
- **Dates matter.** Advice, prices, and reviews expire. Always ask when the information is from. A
  2021 review of a phone that's since had three software updates is a museum piece.
- **Strong claims need strong sourcing.** "Studies show" means nothing without naming the study. The
  more dramatic the claim, the more specific the evidence should be.
- **Your use case narrows the universe.** "Best laptop" is unanswerable. "Best laptop under ₹60,000
  for video editing, 14-inch or smaller" is answerable. Context is what turns noise into a shortlist.

## Practical workflow

Start by telling the assistant what you're deciding and why. A good opening looks like: *"I'm
thinking about switching from shampoo to a solid shampoo bar. I have dry hair, I travel a lot, and
my budget is tight. Can you research this for me?"*

Then steer the research in rounds:

1. **Round one — the landscape.** Ask: *"Give me an overview of this topic: what are the main options
   or viewpoints, and what are the basic trade-offs?"* You're not deciding yet, just mapping the
   terrain.
2. **Round two — compare.** Ask for a table: *"Compare the top 3–4 options in a table: price, pros,
   cons, and who each is best for."* Tables force apples-to-apples comparisons and surface gaps. If a
   column matters to you and it's missing, ask for it to be added.
3. **Round three — pressure-test.** Ask the assistant to argue against its own recommendation: *"Now
   play devil's advocate — what's the strongest case against your top pick?"* If the assistant can't
   find real weaknesses, that's a red flag, not a green one.
4. **Round four — verify.** Ask: *"For each key claim, tell me the source and its date. Flag anything
   you're unsure about or that's based on older information."* Take the most decision-critical claims
   and spot-check one or two yourself.
5. **Round five — save it.** Ask: *"Summarize my findings in a short note: the decision, the top 2
   options, the key evidence, and any open questions. Keep it under 300 words so I can re-read it
   later."* This becomes your decision log — much better than re-researching in three months.

Iterate conversationally between rounds. New questions will come up ("wait, what about this brand?")
— feed them in. You're the director; the assistant is the intern with perfect recall.

## Common pitfalls

- **Asking leading questions.** "Why is X the best?" will get you a sales pitch for X. Ask "What are
  the strengths and weaknesses of X?" instead.
- **Skipping the date check.** Prices, policies, and "current best" lists go stale fast. If no date is
  given, assume the information might be old.
- **Confusing confidence with correctness.** A fluent, detailed answer can still be wrong. Detail is a
  style, not a proof. Look for named sources.
- **Researching forever.** Set a deadline for yourself: "I'll decide by Friday." Research has
  diminishing returns — after the big trade-offs are clear, more reading mostly produces more anxiety.
- **Acting on health or money claims without a professional.** General information is great for
  understanding your options, but it doesn't replace a doctor or a licensed financial advisor for
  decisions that affect your body or your money. This is educational information, not professional
  advice.