Skip to content
Agentic Levels

Everything starts here.

GuestLocal progress only
PreferencesSign in
01Start with one taskBest first move for beginners.02Check your LevelMeasure where you are.03Score an AI resultFind the habit to practice first.04Return to Your WorkScores, links, and checkpoints.
Start here

Begin

HomeThe main entry point.New to AIStart with one useful task.
Know where you are

Measure

Check your LevelUse this after you have tried AI.Fluency ScoreScore an AI result you can review.
Build the habit

Learn

LevelsLessonsTracks
Find the reference

Library

PromptsReferenceResourcesCompare Tools
Turn it into work

Apply

Your Next MoveChoose what AI should change next.Tool SetupGet the tools ready.
Come back later

Return

Your WorkScores, links, and checkpoints.My PathContinue from your level.Updates
Site

Site

PricingAboutFAQ & FeedbackPreferences

© 2026 Fuentes Studio

Privacy·Terms
yourCouncil
Ready to help
✦

What do you want to understand?

Ask anything about what you're learning.

Research, synthesis, and literature reviews, done faster

Most researchers blame the model for generic answers when the real problem is the question. This track teaches you to brief before you ask, verify every claim, analyze real documents, connect live sources, and eventually run recurring research operations across Claude, ChatGPT, Gemini, Perplexity, NotebookLM, Elicit, Codex, and approved internal tools. No code required.

Start with the research mindset resetModule 1 · Lesson 1 · 12 min
47
Lessons
0–10
Levels
6
Modules
🔬

Start at Module 1

This track starts at Level 0 and assumes no research-AI experience. Module 1 resets the one habit behind every generic answer. If you are new to giving an AI tool context, the two core lessons below make Module 1 land faster, but they are optional.

Begin Module 1 →

Who this is for

Analyst or Consultant

I spend days on desk research that should take hours, and the output still reads generic.

Graduate Student

My literature review is drowning me. I cannot keep up with the papers, let alone synthesize them.

Market Researcher

Every competitive analysis starts from a blank page. I rebuild the same structure every time.

Journalist

I need current, citable sources fast, but AI keeps inventing references that do not exist.

Policy Researcher

I have to turn dense regulatory documents and academic papers into a briefing memo, again and again.

Product Manager

Competitive intelligence is a weekly grind. I want it monitored and summarized without doing it by hand.

6 modules · Build in order

Module 1
7 lessons

Levels 0–1 · 7 lessons · Start here

The Research Mindset Reset

Generic AI research answers come from generic questions, not from the model's limits.

Most researchers paste a half-formed question into an AI tool, get an encyclopedia entry back, and blame the model. This module fixes the one habit behind every generic answer: briefing before asking. You will run the same question two ways, learn to make the tool surface your blind spots, and build a verification contract so you never trust a fabricated summary again.

Worth knowing:The same research question produces completely different answers for different contexts. Two to three sentences of background changes output quality more than any other single move you can make.
First-time AI researchersAnalysts and consultantsGraduate students
  1. 1Why your AI research answers are genericL0
  2. 2The briefing-first habitL0
  3. 3Your first structured research promptL1
  4. 4Ask AI to find your blind spotsL1
  5. 5The verification contractL1
  6. 6Reading AI confidence signalsL1
  7. 7When to trust the summary and when to read the sourceL1
Start Module 1 →
Module 2
8 lessons

Levels 2–3 · 8 lessons

Research Templates and Persistent Workspaces

Stop rewriting the same prompts every session. Build reusable structures instead.

At Level 2, the work shifts from one-off prompts to durable systems. You will set up a saved research workspace in Claude Projects, ChatGPT Projects, NotebookLM, or the closest approved tool, write a Project brief that makes every future query sharper, and build reusable templates for competitive analysis, market mapping, and multi-source synthesis. By the end you have a devil's advocate pass and an annotated bibliography workflow you can run on any topic.

Worth knowing:A well-written Project brief is only five sentences, but it means every research query you run afterward starts with your full context already loaded. No re-explaining from scratch.
Strategy and market researchersFounders sizing a marketCompetitive intelligence leads
  1. 1Project spaces as a research workspaceL2
  2. 2Writing your Project brief: the 5 sentences that make a Project workL2
  3. 3Competitive analysis template: build a reusable structure in 20 minutesL2
  4. 4Market mapping with AI: players, segments, whitespaceL2
  5. 5Multi-source synthesis: loading 3-5 sources and finding the real tensionsL3
  6. 6The devil's advocate pass: always ask for the strongest objectionL3
  7. 7Building an annotated bibliography with AI as a thinking partnerL3
  8. 8Using Artifacts to capture structured outputs that survive the conversationL2
Start Module 2 →
Module 3
8 lessons

Levels 3–5 · 8 lessons

Document Analysis at Scale

Feed the AI tool what it needs to do serious analytical work, then verify every claim.

This module is about putting real documents in front of Claude, ChatGPT, Gemini, NotebookLM, or another source-grounded tool: PDFs, reports, interview transcripts, surveys, and full literature sets. You will learn what the tool can and cannot do with a document, use the quote-before-claiming instruction to eliminate fabricated summaries, extract structured tables of findings, compare contradictory reports, and build a PRISMA-style literature review workflow without writing any code.

Worth knowing:AI-generated citations are dangerous: a tool can produce a perfectly formatted reference to a paper that does not exist. One instruction, applied before analysis, removes the fabrication problem entirely.
Literature reviewersQualitative researchersPolicy and legal analysts
  1. 1How AI reads a document: what it can and cannot do with a PDFL3
  2. 2The quote-before-claiming instruction: eliminating fabricated summariesL3
  3. 3Uploading a report and extracting a structured table of findingsL4
  4. 4Analyzing an interview transcript or survey: surfacing missed themesL4
  5. 5Multi-document comparison: loading three reports and finding contradictionsL4
  6. 6Building a PRISMA-style literature review workflow (no code required)L5
  7. 7Citation hygiene: why AI-generated citations are dangerous and how to verify every oneL4
  8. 8When to use Projects (RAG) vs. active context, and how to tell the differenceL4
Start Module 3 →
Module 4
8 lessons

Levels 5–7 · 8 lessons

Live Research with External Sources

Combine synthesis with tools that have current, citable information.

Every model has limits, so this module connects synthesis to live, citable sources. You will learn the search-then-synthesize pattern: Perplexity, Gemini Deep Research, ChatGPT search, Elicit, PubMed, or approved connectors retrieve sources, then Claude, ChatGPT, Gemini, or NotebookLM does the comparison work. The module ends with delegating multi-step research tasks that run in the background, with review gates before use.

Worth knowing:Live-source work changes the job: the assistant can organize and compare, but every factual claim still needs a dated source you can open.
Biomedical and clinical researchersJournalists and fact-checkersLegal researchers
  1. 1Why model memory is not enough for current researchL5
  2. 2The search-then-synthesize pattern: search finds, AI comparesL5
  3. 3Grounded research queries without leaving your AI workspaceL6
  4. 4PubMed for biomedical research: source retrieval before synthesisL6
  5. 5Legal research: Thomson Reuters and Free Law Project MCP connectorsL6
  6. 6Building a research plan before executing itL5
  7. 7AI as a background research agent: delegating a multi-step taskL7
  8. 8Standing delegations: research that runs every week without you askingL7
Start Module 4 →
Module 5
8 lessons

Levels 7–10 · 8 lessons

Research Operations

Build systems that do research for you, not just with you.

This is where research becomes infrastructure. You will build a weekly competitor monitoring system in Claude Cowork, ChatGPT Tasks, Gemini, Codex, or another approved automation surface, set up literature monitoring so new papers find you, and design a multi-agent research team where four roles eliminate each other's blind spots. The module covers how Anthropic built their own multi-agent system, what 'done' means for autonomous research, and what solo researchers running at team throughput actually do differently.

Worth knowing:Anthropic's multi-agent research system reported a 90% improvement over a single agent on research tasks. This module breaks down what that number actually measures and how to design the same roles yourself.
Research leads building systemsSolo researchers scaling outputKnowledge-management practitioners
  1. 1How to build a weekly competitor monitoring systemL7
  2. 2Setting up automatic literature monitoring so new papers find youL7
  3. 3Designing a multi-agent research team: four roles that eliminate each other's blind spotsL8
  4. 4How Anthropic built their multi-agent research system (and what the 90% improvement actually means)L8
  5. 5Turning your Obsidian vault into a searchable research brainL7
  6. 6The self-writing research brief: from raw sources to structured output automaticallyL8
  7. 7Defining done for autonomous research: threshold events and signal vs. noiseL9
  8. 8The 10x research operation: what solo researchers running at team throughput do differentlyL10
Start Module 5 →
Module 6
8 lessons

Levels 2–7 · 8 lessons

Research Archetypes

The same principles, applied to your specific field.

Every prior module taught a general skill. This one shows the full workflow end to end for a specific role, so you can copy the pattern closest to your work. Consultant desk research, VC due diligence, journalist sourcing, the graduate literature review, policy briefing memos, product competitive intelligence, and the PKM practitioner's compounding knowledge system. The final lesson helps you assemble your own workflow from the templates in this track.

Worth knowing:These workflows are not different methods. They are the same briefing, verification, and synthesis habits from Modules 1 through 5, sequenced for one job. Find the closest archetype and adapt it.
Consultants and VC analystsJournalists and policy researchersProduct managers and PKM users
  1. 1The consultant's research workflow: desk research, synthesis, and deck-ready outputL2
  2. 2The VC analyst's due diligence workflow: ingesting the package and stress-testing the thesisL3
  3. 3The journalist's research workflow: live sourcing, then pattern analysisL3
  4. 4The graduate student's literature review: Elicit to find, AI to synthesizeL3
  5. 5The policy researcher's workflow: regulatory documents and academic literature into a briefing memoL4
  6. 6The product manager's competitive intelligence workflow: weekly monitoring into a positioning updateL5
  7. 7The PKM practitioner's workflow: Obsidian vault plus AI synthesisL6
  8. 8Building your own research workflow from the templates in this trackL5
Start Module 6 →

Optional: 2 core lessons that make Module 1 easier

The research track is a true entry point. You can start at Module 1 with zero background. If you have never briefed an AI tool with context before, these two short sprint lessons teach the habit Module 1 builds on.

Your First Productive AI Conversation (l0-02)

The research track assumes you can give an AI tool context and iterate on its output. This short sprint lesson establishes that foundation. Helpful, not required.

Go to lesson →
Role and Context Framing (l1-01)

Module 1 is built on briefing before asking. This lesson teaches you to write the role and context sentence that changes the quality of what the tool produces.

Go to lesson →

Fix your generic answers in 12 minutes

Lesson 1 shows you the single habit behind every generic AI research answer, then has you run the same question two ways so you see the gap for yourself. No research-AI background required. Just bring a question you have been putting off.

Start Module 1 →