The Smart Research Stack
An Enhanced Workflow Powered By AI Tool Stack for Students, Young Professionals, Consultants, and Analysts
As the new adage goes…
"AI isn't going to take your job, someone who knows how to use AI, will"
With hundreds of apps and AI tools out there, many designed for very specific purposes, people who succeed will be those who know how to leverage these tools and upgrade their workflow. They won’t just outshine others—they’ll future-proof their careers.
These will be people with a whole lot of curiosity and a knowledge of how to stack the right tools and skills together.
A tool stack tailored to your workflow allows for work to be better, faster, cheaper, and more fulfilling. And the time you save? That’s bandwidth you can use to tackle more complex, creative tasks that only humans can do. That's more time for family, friends, and hobbies.
When I recently worked on an article, I experimented to combine a few tools and refined my research method.
The process was faster, more efficient, and the results were far superior.
In this post, I’ll explain my workflow and share my preferred tool stack and a step-by-step guide to help you do the same.
Why Many People Struggle To Adopt New Tools
AI has become a household buzzword over the past few years. Most of us have used at least one AI tool, like ChatGPT or CoPilot. But relying on a single tool for every task is like using a Swiss Army knife when what you really need is a power drill.
People struggle with AI tools for the following reasons:
Awareness: They simply don’t know about whats's available.
Overwhelm: With so many options, it’s hard to know where to start.
Workflow Integration: New tools often require rethinking how a task is completed. Trying to shoehorn new technology into old processes is like adding an engine to a horse cart—it doesn’t work.
Access: Some companies block AI tools over privacy and security concerns on corporate devices and networks.
Mindset: People don’t think about how to combine AI’s abilities with their personal strengths: critical thinking, creativity, and emotional intelligence.
That’s where this guide comes in. It’s for students, young professionals, and analysts who want to:
Find credible data sources.
Form and validate hypotheses tied to specific objectives.
Synthesize inputs effectively.
Create polished presentations or articles.
Academics can extend it to match the rigor required for their research.
Let’s Begin!
But First, Read This
Its Very Important
Don’t outsource the entire research process to AI.
At best, it will result in an outcome that sounds nothing like you and at worst, will have factual errors. Use this method to accelerate the process but evaluate the relevance and credibility of each source yourself, stay engaged in the critical analysis and ensure you read the entire draft twice before you publish.
This article is not an tutorial for the tools mentioned here. You will find plenty of resources available online. I have covered my process in detail and described how each tool fits in there.
Objective
Always start with a a clear research question or objective. Sometimes it's handed to you and sometimes you come up with it yourself. What question are you answering, what problem are you trying to solve?
With or without AI, a clear and well defined objective drives the success of the research.
Step 1: Identify Sources
First step in this research workflow is to identify reliable and relevant sources of information. For this, I recommend two tools: Elicit and Perplexity AI.
Both tools use AI to accelerate source discovery and provide a structured way to analyze data.
Elicit helps researchers systematically gather and evaluate evidence from published research papers. It supports structured searches and extracts key insights, making it easier to find high-quality sources. If you use Zotero for citations, Elicit integrates seamlessly with it.
Perplexity AI provides answers and insights from a wide range of sources. Unlike ChatGPT, Perplexity includes citations for every insight it generates, helping you trace the information back to its origin.
Example: Let’s say you’re researching the impact of AI on education. Elicit can help you find academic papers discussing this topic, while Perplexity can summarize articles or highlight debates.
Step 2: Synthesize Information
Once you’ve gathered your sources, the next step is to synthesize the information. This is where Notebook LM by Google shines.
Notebook LM: Think of this as an AI model trained on the sources you provide. You can upload up to 50 URLs and documents in various formats (PDFs, YouTube videos, Google Docs, Audio files etc.). Once uploaded, Notebook LM’s correlates information from each source and expedites the analysis and insight derivation process.
It's chat feature allows you to extract highlights, compare arguments, and identify patterns across all sources.
While you can use the likes of ChatGPT for this step, Notebook LM allows you to focus your analysis on the sources you selected versus the entire web.
Example: If one source supports your hypothesis and another contradicts it, use Notebook LM to pull highlights from both. Start with the “Briefing Document” feature, which summarizes key points from all sources.
This step is crucial because your ability to critically analyze and synthesize information will determine the quality of your output.
I like to spend the most amount of time on this step.
Step 3: Draft Content
With your synthesized information in hand, it’s time to draft your content. For this, I recommend ChatGPT or Claude.
ChatGPT/Claude: These tools can help you draft, proofread, or even completely generate content. While I prefer to draft content myself based on output of Step 2 and then use ChatGPT for editing, you can prompt these tools to generate entire sections.
If you are going to generate your text using these tools, make sure you get an outline of the key points from Notebook LM and then feed it in the prompt. Just remember, the output won’t match your unique voice or style.
To get the output to match your desired style, use ChatGPT’s “Custom Instructions” to guide it toward the tone and style you want. Lookup how to use custom instructions or how to engineer your prompts to do that for you.
Step 4: Visualize
When you need to explain processes or compare data points, visualization can make your work more accessible. Sometimes, even a visual summary makes it appealing.
For this, I recommend using Napkin.ai.
Napkin.ai creates visual representations of complex information, making it more accessible and easier to comprehend.
With customizable options, it’ll save you time while making your document more visually appealing.
Other Tools in my Toolkit
In addition to this research focussed stack, here is my Core Tool Stack:
Obsidian: For mind mapping.
Notion: For tracking and planning.
Kortex: For drafting content.
Readwise: For note-taking and building a knowledge catalog.
In conclusion
This method reduces the manual effort involved in source discovery, drafting, proofreading while widening the scope of your research.
By shifting much of the grunt work to AI, you can focus on what really matters: critical thinking and creativity.
And that, is the ultimate skill for any knowledge worker.
Most AI platforms are still in their early days, getting familiar with these systems now will give you a head start in the new era of human-machine collaboration. It will save you from a lot Purpose Anxiety in the future. (I will cover Purpose Anxiety in future posts)
So, what will you build with your AI tool stack? I am curious to hear about your experiments and discoveries!