We recently announced Deep Research, a powerful AI research assistant designed to help users dive deep into topics with structured, adaptive research. Unlike simple search engines or basic AI models, Deep Research is the best AI for research, it analyzes information, follows links, and refines its research approach to deliver truly insightful results. To help you get the most value from Deep Research we’re sharing best practices for crafting effective prompts that guide your AI research assistant to provide the most relevant and comprehensive insights.

Why Better Prompts Lead to Better Research

The key to unlocking the full power of Deep Research is providing as much detail as possible in your prompt. Unlike a simple search or a basic AI model, your AI research assistant doesn’t just retrieve quick answers—it builds a research plan, follows links, and refines its approach based on what it learns. The more context your prompt contains, the more effectively it can analyze sources, uncover insights, and generate meaningful results tailored to your needs.

A vague prompt like "Tell me about renewable energy" may return broad, high-level information, whereas a detailed request such as "Analyze the impact of offshore wind farms on coastal ecosystems, focusing on environmental, economic, and regulatory factors" enables the system to dig into specific aspects of the topic and deliver a much richer, more actionable response.

Sample Prompts for Effective Deep Research

Here are some examples of strong research prompts across different fields that illustrate how to use Deep Research:

Finance:

  • Weak Prompt: "Tell me about the impact of rising interest rates."
  • Strong Prompt: "Analyze the impact of rising interest rates on the U.S. stock market over the past five years. Include trends, expert opinions, and case studies of major companies affected by rate hikes."

Science & Technology:

  • Weak Prompt: "What’s new in quantum computing?"
  • Strong Prompt: "Investigate the latest advancements in quantum computing and their potential impact on cybersecurity. Summarize key breakthroughs, ongoing challenges, and expected developments in the next five years."

Legal & Compliance:

  • Weak Prompt: "Explain new data privacy laws."
  • Strong Prompt: "Analyze recent changes in U.S. data privacy laws and their implications for businesses. Compare GDPR, CCPA, and emerging regulations, highlighting compliance requirements and potential legal risks."

Marketing & eCommerce:

  • Weak Prompt: "How should I design website buttons?"
  • Strong Prompt: "Conduct an in-depth study on the best button design strategies for eCommerce websites to maximize conversion rates. Analyze factors like color, size, wording, placement, and A/B test results from leading retailers."

Enhancing Your Research with the Ninja Prompt Improver

To create better, detailed prompts try Ninja’s Prompt Improver. This simple tool refines your prompt by automatically adding key details and expanding the context using AI. This tool ensures your Deep Research query is structured for maximum depth, helping you get the best possible results for your AI research assistant.

Simply enter a basic research prompt, and the Prompt Improver will expand it with additional clarifications, subtopics, and focus areas to provide a more precise and well-rounded request. For example:

Basic Prompt: "I'd like to buy a telescope for astrophotography"

Improved Prompt: "I'm looking for a telescope specifically designed for astrophotography and I need help choosing the right one. I'd like to know the key features to consider, such as aperture size, mount type, and camera compatibility. Could you provide me with a list of top recommendations for telescopes suitable for astrophotography, including their pros and cons, and a brief guide on how to select the best one based on my budget and skill level? Please include information on any additional equipment I might need to get started with astrophotography."

Reviewing and Refining Your Enhanced Prompt

While the Prompt Improver is a powerful tool, it’s always good practice to review the refined prompt before running Deep Research. Sometimes, the additional details it suggests may not fully align with your intent. Take a moment to check that the improved prompt includes the right level of specificity and that all extra context is relevant to your research goals.

For example, in the previous prompt, I may have some of the information specified in the updated prompt. Perhaps I know that I want a refractor telescope, with an aperture greater than 70 mm. I could update the improved prompt to better reflect my needs:

Update  Prompt: "I'm looking for a refractor telescope specifically designed for an intermediate astrophotographer and I need help choosing the right one. I'd like an aperture size greater than 70 MM and compatibility with DSLR cameras. Could you provide me with a list of top refractor telescopes suitable for astrophotography, including their pros and cons, and a brief guide on how to select the best one based on my budget of $1500? Please include information on any additional equipment I might need to get started with astrophotography."

By using a well structured prompt—or refining your own with the Prompt Improver—you can guide your AI research assistant toward delivering comprehensive, high-quality Deep Research tailored to your needs.

‍