In a recent LinkedIn post, Andrew Bolis discusses the critical need for structured systems when utilizing AI tools, particularly large language models like Gemini, to extract meaningful insights from data. Bolis emphasizes that simply uploading files to an AI will not yield the desired results unless a deliberate workflow is established around the technology.
Andrew Bolis highlights the capabilities of NotebookLM, powered by Gemini, as a tool that transforms disparate files into a cohesive, cited research repository. He outlines a comprehensive workflow that power users can adopt to maximize the AI’s potential.
“Most people upload a file and hope AI “figures it out.” It won’t … unless you build a system around it.”
Bolis explains that the core engine, Gemini Flash, is designed for fast, accurate synthesis across large document sets. For more complex tasks requiring deeper reasoning, the Gemini Pro (Plus Tier) is recommended, especially for enterprise workflows. He provides sample prompts for both, illustrating how to leverage these engines for distinct purposes.
Leveraging NotebookLM’s Features for Enhanced Research
Andrew Bolis details a multi-step process within NotebookLM, starting with the addition of various source types, including PDFs, documents, transcripts, and webpages, noting that formatting and images are preserved. This forms the foundation for subsequent AI-driven actions.
Source-Based Chat and Structured Output
A key feature Bolis points out is the ‘Source-Based Chat’ functionality, which restricts responses to the uploaded content, making it ideal for validation and review processes. He also emphasizes the utility of ‘Structured Study Guides,’ which can generate summaries, timelines, and briefs to help organize large volumes of material.
“It turns scattered files into a connected, cited research brain.”
Further elaborating on the advanced features, Bolis discusses ‘Visual Topic Mapping,’ which creates visual representations of topic relationships, and ‘Audio Overview,’ which converts notebooks into spoken recaps for convenient on-the-go review. The platform also offers ‘Video Summaries’ for creating short, narrated video updates or training materials.
Advanced Data Extraction and Collaboration
Andrew Bolis underscores the power of ‘Deep Research’ to find credible references and expand upon a given topic. He also highlights the ‘Structured Data Tables’ feature, which extracts key data points into organized, sortable rows, transforming unstructured information into usable data.
“Build the system once and turn raw files into insights fast.”
Collaboration is also a significant aspect, with ‘Collaborative Notebooks’ allowing users to share their research with teams or clients while maintaining structure and citations. Bolis provides a set of ‘Power Prompts’ designed for specific tasks such as summarizing content, comparing sources, identifying decisions, generating briefs, and creating scripts for audio or video content.
Practical Workflow Examples
To illustrate the practical application of these features, Andrew Bolis outlines several workflow examples. These include ‘Research Review,’ where papers are added, briefings are produced, and targeted questions are asked; ‘Shared Knowledge Base,’ for centralizing project documents and creating onboarding guides; and ‘Content Analysis & Creation,’ for analyzing competitor materials and generating presentation slides.
“Respond only using your uploaded content. Ideal for reviews, validation, and checks.”
In conclusion, Andrew Bolis advocates for building a robust system around AI tools to unlock their full potential, transforming raw files into actionable insights efficiently. He encourages readers to save his guide and test these workflows on their own deep-work tasks.
📝 About This Content
This article is based on insights shared by Andrew Bolis on LinkedIn.
📅 Originally posted on January 25, 2026 | View original post on LinkedIn →