Skip to main content

deep-research

#DeepResearch

Deep Research capabilities, integrated into models such as OpenAI and Gemini, represent a qualitative leap over traditional web search in chatbots.

While a normal web search makes a query and summarizes the first results, a Deep Research agent works autonomously and iteratively:

  1. Plan a research strategy.
  2. Run multiple searches and read dozens of pages.
  3. Iterate looking for additional information if you find gaps in your initial research.
  4. Synthesizes all the information in a final, rigorously referenced report.

This tool is exceptional for teachers when they need to prepare new syllabi, research methodologies, or analyze complex regulatory frameworks.

Usage example​

Below is a prompt structure that takes advantage of these capabilities for educational research:

Asistente IA
👤

Draft an updated, comprehensive, and evidence-based report that explores strategies for implementing Artificial Intelligence (AI) in secondary education and vocational training. The primary audience will be vocational training teachers in marketing and commerce, extending also to areas like physical conditioning and English. The report should address the following points:

  1. Applications of AI in Teaching:

    • Explore all available AI tools that can facilitate teaching work, from language models to other technological solutions.
    • Present a list of possible uses and applications to improve teachers' tasks, including practical and applicable examples. It is not necessary to include step-by-step tutorials, but general indications and prompts that serve as a starting point for future queries and implementation.

  2. Technological Integration and Moodle:

    • Specifically describe how to generate importable and Moodle-compatible materials, offering specific examples of strategies and methods to make the most of this platform in the educational context.

  3. Practical Cases and Applicable Examples:

    • Include case studies or practical examples that illustrate how AI has been successfully implemented in similar institutions, highlighting the impact on teaching tasks and processes.
    • Focus on specific examples for the marketing and commerce areas, but also consider scenarios for physical conditioning and English.

  4. Additional Aspects and Recommendations:

    • Offer a brief reference to other aspects to consider, such as ethical challenges, continuous teacher training, and the adaptation of pedagogical methodologies given the incorporation of AI.
    • Include recommendations for training educators in the use of these technologies and suggest updated and relevant sources to delve deeper into the topic.

The goal is to provide reference material that not only informs but also serves as a practical guide so teachers can effectively identify, evaluate, and apply AI tools in their educational work.

🤖

Delivery proposal ("deep research" mode)

Rigor note: for an "evidence-based" report I need to consult sources (or have you provide documents). If there is no web access or sources are not attached, I can:

  • develop a practical framework,
  • propose a search and verification protocol,
  • and list what evidence to collect.

A) Clarifying questions (to adapt the report)

  1. Is the goal to train teachers (methodology) or implement tools (operations)?
  2. What school restrictions apply (privacy, institutional accounts, student age)?
  3. What platform will you use (Moodle/H5P/Google Workspace/M365) and with what resources?
  4. Do you prefer a cross-cutting approach (valid for any subject) or by professional families?

B) Recommended structure of the report (list of sections)

  1. Principles of responsible use (privacy, biases, transparency, traceability)
  2. Teaching use cases (preparation, classroom, assessment, attention to diversity)
  3. Prompt design for educational materials (reusable templates)
  4. Moodle: importable formats (GIFT, question banks, rubrics) and workflow
  5. Quality verification (rubrics, examples of failures, checklist)
  6. Teacher training plan by phases + security minimums

C) Evidence to collect (so the report is verifiable)

  • Official guides from providers (prompting, assessment, structured outputs, security)
  • Educational organizations (e.g., digital teaching competence frameworks) and AI recommendations
  • Studies/reviews (e.g., surveys and systematic reviews) on AI in education
  • School policies: data use, minimum age, consent, and traceability

D) Practical deliverables included

  • 5–8 reusable "mega-prompts" for: activity worksheets, rubrics, feedback, GIFT, and level adaptation
  • 2 checklists: (1) before using AI, (2) quality/security review of the material
  • 1 quick assessment template to detect hallucinations and biases

If you answer the questions in section A, I will generate the final index and a first cross-cutting version (valid for any subject), with specific examples adaptable to marketing/commerce, physical conditioning, and English.

Why it works:

  • Start by declaring limits of rigor: if there are no sources, propose a framework and a protocol.
  • Asks clarifying questions (to avoid “generic reports”).
  • Ask for evidence and define practical deliverables (templates, checklist, rubrics).
  • Take advantage of the agent's ability to navigate multiple sources before providing a final response.