The Google Espionage Playbook: Using Gemini for OSINT Reporting

This guide explores how to leverage Google Gemini to streamline the production of Open-Source Intelligence (OSINT) reports. While the human analyst remains the final authority, Large Language Models (LLMs) serve as powerful force multipliers for data processing and preliminary analysis.

Executive Summary: Gemini as a Force Multiplier

Sophisticated LLMs like Google Gemini can automate the most time-consuming phases of the OSINT cycle: collection, processing, and preliminary analysis (Business Agility Institute, 2025; U.S. Army, 2024; ResearchGate, 2020). By utilizing structured prompt engineering, analysts can transform Gemini into a high-speed production engine for professional, citable intelligence reports (ResearchGate, 2024; IMSL, 2023; Packetlabs, 2025).

Chapter I: The Intelligence Cycle

The traditional OSINT lifecycle consists of five stages: Planning, Collection, Processing, Analysis, and Dissemination. Gemini excels at the middle three, allowing the analyst to focus on mission definition and final strategic delivery (Recorded Future, 2024; Lerner, 2024; SANS, 2023).

A. Gemini's Role

OSINT Stage	Function in Gemini-Augmented OSINT	Human Analyst Role
1. Planning & Direction	Defined by the specificity of the initial prompt (the mission directive) (BitSight, 2025).	Define the Key Intelligence Question (KIQ), scope, and ethical boundaries (Reddit, 2024).
2. Collection	Gemini's integrated search executes comprehensive, simultaneous searches across the public web, handling massive data volume (FedTech, 2025; Stanford Law, 2025; Neotas, 2025).	Validate source reputation and adjust the search scope if initial findings are lacking (Kamesh, 2024).
3. Processing & Analysis	Gemini filters unstructured text, de-duplicates information, identifies key entities (people, dates, organizations), and detects patterns and anomalies (Blackdot, 2025; IMSL, 2023; VEEXH, 2025).	Apply critical thinking, mitigate bias, and identify strategic implications that the AI cannot perceive (Lumenalta, 2024; Tableau, 2024).
4. Dissemination	Gemini generates the final, structured report in the specified format (e.g., LaTeX, Executive Summary), complete with citations and a bibliography (SANS, 2023; Packetlabs, 2025).	Deliver the final judgment and present findings with clear, actionable recommendations (Reddit, 2024).

This synergy leverages the LLM's computational speed alongside the analyst's expertise in strategic judgment and ethical oversight (Business Agility Institute, 2025; i2 Group, 2024).

Chapter II: Prompt Engineering

The quality of an OSINT report depends on the precision of the prompt. A professional mandate must define the role, scope, depth, and output format (ResearchGate, 2024).

A. The OSINT Triad

Every effective OSINT prompt should incorporate these three core elements:

The Action Mandate: Instruct Gemini to adopt a professional persona to ensure a structured, analytical output. Use: Act as an OSINT analyst, search, analyze, and then generate... (SANS, 2023).
The Scope Constraint: Define clear boundaries to prevent data overload and maintain focus. Use: ...a full OSINT report about [Subject X] with a focus on [Specific Timeframe]... (BitSight, 2025; ITU Online, 2024).
The Verification and Output Enforcer: This is essential for creating an auditable, credible intelligence product. Use: ...with at least [N] references. Ensure every key fact is supported by 3 citation links. Output in [Format]. (WebAsha, 2025; Kamesh, 2024; Penlink, 2025).

B. Indispensable Keywords

Integrate specific variables into your prompts to dictate the depth and structure of the final report:

Variable Category	Keyword to Use in Prompt	Purpose of the Keyword
Role & Tone	`Act as a geopolitical threat intelligence analyst...`	Enforces objective, professional tone and analytical framework (SANS, 2023).
Data Scope	`...using current and historical open-source intelligence...`	Ensures comprehensive data collection across multiple time periods and source types (Lerner, 2024; Neotas, 2025).
Analytical Depth	`...correlate and analyze the findings to identify key entities, relationships, and potential threats...`	Prompts advanced analysis beyond simple summarization, seeking patterns and links between data points (IMSL, 2023; VEEXH, 2025; ITU Online, 2024).
Structural Mandate	`...include a Methodology section, an Executive Summary (BLUF), and a Conclusion with Future Implications...`	Enforces the structure of a formal intelligence brief, ensuring the product is ready for senior-level dissemination (BitSight, 2025; Reddit, 2024; SANS, 2023).
Verifiability	`...citing 3 sources for every key fact...` / `...minimum 20 unique references...`	The most critical component. Directly addresses the risk of LLM hallucination and creates an auditable trail, guaranteeing credibility (WebAsha, 2025; Penlink, 2025; Lumenalta, 2024).
Final Output	`Output in LaTeX` / `Output in Markdown` / `Generate as a detailed executive briefing`	Specifies the final presentation format for seamless integration into your workflow and professional archives (Packetlabs, 2025; Homeland Security, 2024).

Chapter III: High-Value Prompt Templates

The following examples demonstrate how to generate complex, cited, and professionally structured OSINT reports.

Example 1: Geopolitical Deep Dive

Act as a geopolitical threat analyst. Search, analyze, and then generate a full OSINT report about PNAC (Project for the New American Century) with a focus on Donald Rumsfeld between 1997 and 2007. The report must correlate their stated goals with policy outcomes and identify key funding sources. Include a Methodology section, an Executive Summary, and a Conclusion with Future Implications and Recommendations. Provide at least 20 unique references, citing 3 sources for every key fact. Output in LaTeX. Author: Marie-Soleil Seshat Landry with AI (Google Gemini). Date: [Today's Date]. Publisher: marielandryspyshop.com

Example 2: Threat Actor Analysis

Act as a counter-terrorism analyst. Give me an OSINT report on ISIS covering their evolution from historical foundation to current activities in 2025. The analysis must identify key leadership changes, evolving TTPs (Tactics, Techniques, and Procedures), and digital recruitment strategies. Output in 2000 words with all facts cited from a minimum of 15 reputable sources, ensuring 3 citations for all key facts. Include a detailed timeline of significant events.

Example 3: Humanitarian Risk Assessment

Act as a risk assessment consultant. Generate a 5000-word OSINT report on the state of women's rights in Afghanistan in 2025. The analysis must cover legal frameworks, socio-cultural barriers, education, and economic impact. The report should structure findings by province where possible and include a risk assessment for NGOs operating in the region. Ensure every key finding is supported by 3 citations. Output in a detailed, professional PDF-ready format.

Chapter IV: The White Hat Mandate

Responsible OSINT requires strict adherence to ethical and legal boundaries. Gemini is subject to policies that prohibit illegal activity (ITU Online, 2024; Bolen, 2024).

A. Non-Negotiable Boundaries

No Access to Private Data (PII): Do not request non-public Personal Identifiable Information (PII). OSINT is limited to information accessible without intrusion (ITU Online, 2024; SANS, 2023; Bolen, 2024).
No Hacking or Intrusion: Do not request instructions for unauthorized access, bypassing paywalls, or exploiting system vulnerabilities (ITU Online, 2024; Kamesh, 2024).
No Malicious Intent: Gathered data must not be used for harassment, doxxing, or causing harm (SANS, 2023; Bolen, 2024).

The three-citation rule serves as a vital audit trail, ensuring that intelligence is derived from reputable, publicly accessible sources (WebAsha, 2025; Kamesh, 2024; Penlink, 2025).

Bibliographic Entry Key

All in-text citations correspond to the following references:

(BitSight, 2025) - BitSight Technologies. (2025). How to Use the OSINT Framework: Sources, Tools, & Steps.
(Blackdot, 2025) - Blackdot Solutions. (2025). The Language of Investigation: How LLMs and GenAI can support OSINT.
(Bolen, 2024) - Bolen, S. (2024). The Ethical Considerations of OSINT: Privacy vs. Information Gathering. Medium.
(Business Agility Institute, 2025) - Business Agility Institute. (2025). From Constraints to Capabilities: AI as a Force Multiplier.
(FedTech, 2025) - FedTech Magazine. (2025). AI Serves as a Force Multiplier in Fed Use Cases.
(Homeland Security, 2024) - Homeland Security. (2024). The Impact of Artificial Intelligence on Traditional Human Analysis.
(i2 Group, 2024) - i2 Group. (2024). Augmenting Human Intelligence: An Overview for the Application of AI in the Intelligence Cycle....
(IMSL, 2023) - IMSL. (2023). The Impact of Language Learning Models (LLMs) on OSINT Investigations: A Deep Dive into ChatGPT.
(ITU Online, 2024) - ITU Online IT Training. (2024). Mastering Open Source Intelligence: A Guide To Ethical OSINT Techniques And Practices.
(Kamesh, 2024) - Kamesh. (2024). Open-Source Intelligence in Legal Investigations: Ethical Boundaries and Best Practices. OSINT Team.
(Lerner, 2024) - Lerner, E. (2024). What Are The 5 Steps Of OSINT 🕵️‍♂️. Medium.
(Lumenalta, 2024) - Lumenalta. (2024). AI's limitations: 5 things artificial intelligence can't do.
(Neotas, 2025) - Neotas. (2025). What Is The OSINT Framework? - OSINT Tools & Techniques 2025.
(Packetlabs, 2025) - Packetlabs. (2025). Using LLMs Like ChatGPT To Support OSINT Campaigns.
(Penlink, 2025) - Penlink. (2025). Best Practices for Integrating Open-Source Intelligence (OSINT) into Investigations.
(Recorded Future, 2024) - Recorded Future. (2024). What is the OSINT Framework? How can you use it.
(Reddit, 2024) - Reddit. (2024). OSINT Report Examples / Templates.
(ResearchGate, 2020) - ResearchGate. (2020). Can AI be a force-multiplier for Intelligence Analysis?.
(ResearchGate, 2024) - ResearchGate. (2024). Implications of Large Language Models for OSINT....
(SANS, 2023) - SANS Institute. (2023). What is OSINT (Open-Source Intelligence?).
(Stanford Law, 2025) - Stanford Law School. (2025). Leveraging Artificial Intelligence to Empower Intelligence Analysis in the Space Domain.
(Tableau, 2024) - Tableau. (2024). What are the advantages and disadvantages of artificial intelligence (AI)?.
(U.S. Army, 2024) - U.S. Army. (2024). Artificial Intelligence as a Combat Multiplier: Using AI to Unburden Army Staffs.
(VEEXH, 2025) - VEEXH. (2025). How to Use AI for OSINT. Learn how to apply AI to open source.... OSINT Team.
(WebAsha, 2025) - WebAsha. (2025). AI-Generated OSINT Reports| Enhancing Intelligence or Risking Misinformation?.

Sources and further reading

Added October 2026: independent sources on points made in this post.

Sources

On “Gemini's integrated search executes comprehensive, simultaneous searches across the public web, handling massive data volume”: OSINT search engine techniques, A Practical Approach to Open Source Intelligence (OSINT) - Volume 1, 2025.
On “The traditional OSINT lifecycle consists of five stages: Planning, Collection, Processing, Analysis, and Dissemination.”: Building OSINT skills and workflow process, A Practical Approach to Open Source Intelligence (OSINT) - Volume 1, 2025.

Cite or link to this article

Free to cite, link and share. A link back helps others find this research.

Suggested citation

Landry, M.-S. S. (2025). The Google Espionage Playbook: Using Gemini for OSINT Reporting. Marie Landry Spy Shop. https://www.marielandryspyshop.com/2025/10/the-google-espionage-playbook.html

Link HTML

<a href="https://www.marielandryspyshop.com/2025/10/the-google-espionage-playbook.html">The Google Espionage Playbook: Using Gemini for OSINT Reporting</a>

More free OSINT resources: OSINT Research Hub · Tools directory · Glossary

Search This Blog