AI Tools for University Students

AI didn’t make university easier for me; it made the work clearer. Back when the only widely used tools were ChatGPT (and briefly Gemini), I learnt that the best results didn’t come from asking a bot to “write it for me”, but from using AI to decode briefs, structure thinking, find credible sources, and pressure-test arguments. That shift alone helped me push for 75–80% work consistently.

Since then, the toolkit has expanded. We now have systems that traverse citation graphs, read hundreds of papers, map author networks, and interrogate your own notes—all while keeping you in control. Used well, these tools create the time and headspace for it. They move you out of mechanical tasks (skimming PDFs, re-formatting references, chasing links) and into the higher-order work markers and supervisors actually reward: analysis, synthesis, originality, and methodological transparency.

What follows is not a tool parade. It’s a workflow you can run end-to-end on any assignment, report, dissertation chapter, or research proposal:

1. Understand the task — surface the learning outcomes, define the scope, plan the structure.

2. Search with provenance — use cited answers, deep literature discovery, and synthesis tools.

3. Organise sources — build a reference library and visualise networks.

4. Draft and interrogate your argument — build stepwise reasoning and structured critique.

5. Polish — refine for clarity, accuracy, and proper citation without losing your voice.

Throughout, I show exactly how to prompt, when to switch models, and how to avoid fabricated citations.

Two principles anchor the whole approach:

• Provenance over convenience. If a claim matters, it needs a verifiable source you can cite and find again.

• Human judgement at the centre. AI accelerates exploration; you decide relevance, weight of evidence, and how competing findings fit your argument.

  
  

Used this way, AI isn’t a shortcut; it’s an accelerator for careful thinking. If these tools had been available earlier in my degree, I’m confident my 74% would have climbed—not because the AI wrote more words, but because it freed me to write better ones.

Stage 1 — Understanding the Assignment

ChatGPT: The Academic Interpreter

Free | Pro Version Available

When I first started using ChatGPT, I realised it wasn’t about getting it to write my assignments—it was about getting it to understand them with me. I used to paste entire assignment briefs and ask it to break things down. It helped me see what the question was really asking, where marks would be gained, and what structure would make the most sense.

ChatGPT is brilliant for turning confusion into clarity. It doesn’t know your exact module, but it can help you unpack complex instructions and generate structured thinking from them.

How I use it:

• Paste the entire brief and ask, “What are the core tasks here? What topics or theories might apply?”

• Ask it to suggest structures for 75–80% work. Then refine those structures yourself.

• Use it for revision — create mock quizzes, flashcards, or question banks.

• Summarise your lecture notes into one-page recaps.

• Have it simulate a tutor: “Ask me 10 questions on this topic to test my understanding.”

Why it works: ChatGPT shines when you guide it. Tell it your discipline, the tone you’re aiming for, or your word count. It adjusts surprisingly well.

How not to use it:Don’t copy its full answers. Instead, copy its questions—the ones that challenge your logic. Let it spark your ideas, not replace them. And never trust AI-generated references; check them manually.

Prompt idea

“Here’s my full assignment brief. Break it down for me: Explain the main divisions or sections I’ll need to include. Suggest how many words I should allocate to each section if the total is 2,500 words. Identify the key topics, theories, or debates I should focus on. Give me 10–12 keywords I can use to begin my research on Google Scholar or library databases.”

Claude: The Deep Thinker

Free | Pro & Max Versions Available

If ChatGPT helps you get started, Claude helps you think deeper. It’s brilliant for analytical writing and long documents—it can hold context better and remember what’s already been discussed. I often describe Claude as my “academic co-pilot.” If I’d had it at university, I would’ve used it to test my arguments before final submission.

Claude isn’t here to give you answers; it’s here to help you find them. I like to think of it as a Socratic tutor — one that asks, “Why?” and “What makes you think that?” until you begin questioning your own assumptions.

It’s trusted by universities such as LSE, Northumbria, Northeastern, and the University of San Francisco, where educators use it not only for teaching but also for research acceleration — analysing datasets, reviewing vast collections of papers, and even generating adaptive assignments through its API.

Why I See Claude as a ‘Thinking Partner’

When I use Claude, I don’t ask it to write my essay or solve a problem.I ask it to walk through my reasoning with me.For example, when I was developing arguments for an essay, I’d share my outline and ask:

“Does this reasoning make sense? Where could I strengthen my logic? Are there any unexamined counterarguments?”

Claude doesn’t just tell me what’s missing — it helps me understand why it’s missing. That’s where its real power lies.

Its questioning feels almost human. It doesn’t rush to provide solutions. Instead, it tries to develop your thinking — pushing you to reflect, reframe, and clarify.

This makes it particularly effective for:

• Drafting analytical essays or dissertations

• Evaluating the strength of an argument or hypothesis

• Exploring ethical questions or theoretical debates

• Planning research frameworks

• Summarising complex readings into principles rather than answers

How to Use Claude Effectively

Claude works best when your prompts give it structure and purpose. It thrives on clarity and context.

1. Start with your thinking, not a blank slate: Paste your outline, notes, or even a messy brainstorm and say:

“Let’s build on this together. Help me test my reasoning and highlight where evidence is weak or missing.”

2. Use Claude for guided discovery: You can literally ask it to teach through questioning:

“Instead of giving me the answer, guide me through questions I should be asking to reach it myself.”

3. Give it structure with tags: Claude responds well to XML-style prompts that clarify what you want:

<task>Critique my argument</task>
<focus>Logic, evidence, counterarguments, and ethical implications</focus>
<output>Improved version + key questions I should ask before finalising</output>

4. Use it for research preparation: Before diving into databases, ask Claude to create:

• Research question templates

• Comparison frameworks (“What factors could I use to compare these theories?”)

• Concept maps or study guides from your readings

  
  

5. Pair Claude with your own notes: If you upload excerpts or drafts, Claude can identify patterns in your writing — recurring assumptions, shifts in tone, or arguments that lack grounding.

Claude is much more than a chatbot. It’s part of a growing academic ecosystem — one where AI doesn’t replace critical thinking but scales it.

Prompt Idea

“I’m sharing my essay outline and draft introduction below. Please evaluate the logical flow and identify any weak reasoning. Ask me clarifying questions before suggesting revisions. Provide a list of key principles or concepts I should explore further. Summarise your feedback in 200 words and include 3 reflection questions I can use to improve my next draft.”

This type of prompt lets Claude challenge your thinking while still keeping you in control. It becomes your collaborator, not your ghost-writer.

Stage 2 — Researching and Building Context

Perplexity AI: The Research Search Engine

Free | Student Pro Access

Perplexity isn’t just another AI search engine — it’s what I call a researcher’s shortcut to understanding. It blends the power of a large language model with real-time web search, so instead of trawling through dozens of websites, you get summarised answers with verifiable sources right beside them.

When I first switched from Google to Perplexity, I was sceptical. Could an AI really give trustworthy answers? But Perplexity doesn’t just summarise — it shows exactly where it found the information. Each response comes with citations that link directly to the original papers, reports, or news articles. It’s like having a personal research assistant who reads everything for you, then hands you a curated digest with references attached.

The best part? It’s completely interactive. You can ask follow-up questions, refine your focus, and guide it to think like a subject specialist.

How Perplexity Works

Perplexity combines a natural-language model with a live search engine.Here’s what happens behind the scenes:

1. When you ask a question, it runs a real-time web search.

2. It reviews multiple credible sources (academic journals, databases, reports, and trusted media).

3. It summarises the findings while listing citations.

4. You can then dive deeper into any cited source with one click.

In other words, it doesn’t generate facts — it curates them. That distinction makes it ideal for academic work, where provenance and credibility matter.

The platform also lets you choose different models (Claude, Gemini, GPT, or Perplexity’s own Sonar). That flexibility means you can pick whichever model suits your goal — faster responses, deeper reasoning, or broader exploration.

How I Use It

Whenever I begin a new assignment, Perplexity is my starting point. Let’s say I’m researching “AI in public health systems.”

• I type the question into Pro Search mode:

  - “How are low- and middle-income countries integrating AI into primary healthcare systems, and what ethical concerns have emerged since 2020?”

  Within seconds, I get a well-structured summary divided into key themes — implementation challenges, data governance, ethical concerns — with citations for each.

• I scan through the references. If a source looks promising, I click through to the original article.

• I save my results in Spaces — Perplexity’s built-in folders where you can organise threads by project, topic, or module. I often name mine after assignment titles or dissertation chapters.

• When I’m working in a group project, I invite collaborators to the same Space, so we all see the same curated information.

• If I need depth rather than speed, I turn on Deep Research mode. This mode takes a few minutes — but in that time, Perplexity scans through hundreds of sources, comparing and validating information before responding. The final answer reads more like a mini literature review than a quick summary.

  
  

How to Use Perplexity Without Overreliance

Perplexity can make research feel too easy. You type a question, get answers, and think, “I’m done.” But the key is to see Perplexity as a starting point, not an endpoint.

Here’s how to stay in control:

• Always click the sources — never quote Perplexity directly. Quote the article it links to.

• Use its structure — headings and key themes — as a research plan for your own literature review.

• Compare multiple searches — slightly change the wording to see if it surfaces new angles.

• Add your analysis — the real mark of distinction is what you do with the findings.

Prompt idea:

“Pro Search: Summarise the most recent evidence (2020–2025) on AI ethics in education, highlighting major themes, controversies, and policy gaps. Provide sources I can verify.”

Always verify the sources given by AI

Tip: Combine Perplexity with Zotero. When you find good sources, capture them instantly.

Undermind: The Literature Review Assistant

Free (Abstracts) | Pro (Full-Text)

If Perplexity helps you scan the surface, Undermind dives deep. It’s not a search engine — it’s a true AI research analyst. While most tools retrieve a list of links, Undermind actually reads the papers. It analyses their abstracts, methodologies, citations, and conclusions, then ranks them by relevance to your topic.

When I first tried Undermind, I realised it wasn’t just fast — it was strategic. Instead of giving me fifty random papers, it returned ten that actually mattered. It recognised recurring authors, connected citations across studies, and even highlighted theoretical gaps.

How Undermind Works

Think of it as a tireless research assistant:

1. You input your topic or question.

2. The AI scans academic databases — open-access journals, preprints, repositories, and metadata archives.

3. It reads the abstracts (and full texts, if you’re on the Pro plan).

4. It cross-compares studies using citation graphs and keyword clustering.

5. It produces a summary explaining the dominant themes, trends, and research gaps.

Unlike search engines that match keywords, Undermind uses recursive retrieval — meaning it learns from its own searches. As it reads, it adjusts its query to find increasingly relevant papers. That’s what allows it to reach both breadth and depth without drowning you in noise.

How I Use It

Undermind has completely changed how I begin literature reviews. Here’s my typical workflow:

1. Define the topic clearly: I don’t just write “sustainable tourism.” I give detail: “Impact of digital platforms on sustainable tourism practices in Europe, 2018–2024.”

2. Choose what kind of research I want: I specify whether I need empirical studies, theoretical frameworks, or systematic reviews.

3. Let it process: Undermind reads through thousands of papers (yes, literally thousands) and delivers a concise report — often divided into themes, trends, key authors, and gaps in the literature.

4. Review and extract: I read the summaries of the top-ranked papers and decide which ones to pursue in full.

5. Export to Zotero or Research Rabbit: That’s where organisation begins. Undermind gives me the raw material, Zotero and Research Rabbit turn it into structure.

   
   

Why It’s So Powerful

• Depth over surface: It doesn’t just skim; it reads and compares content.

• Breadth with control: You can set filters (year range, region, study type).

• Citation mapping: It follows the “who cited whom” trail to show intellectual networks.

• Adaptivity: The more specific you are, the more precise its search becomes.

• Academic reliability: It’s trusted by institutions like MIT, Princeton, Georgetown, and Cambridge — which speaks volumes about its credibility.

How to Use Undermind Without Losing the Human Touch

AI can identify gaps in the literature, but you need to interpret why those gaps exist. That’s where your originality lies. So, once Undermind presents the data:

• Read the top 3–5 papers fully.

• Compare their methodologies and conclusions.

• Ask: What perspectives are missing? What populations or variables are underexplored?

• Build your argument around those insights.

  
  

Use Undermind to collect data about knowledge, not knowledge itself.

Prompt Idea

“Conduct a comprehensive literature scan on [your topic]. Focus on papers from 2018–2025. Prioritise peer-reviewed and open-access sources. Identify 10 recurring themes, key authors, and the most cited papers. Summarise major debates and highlight where research is lacking. Provide a short paragraph explaining how these gaps could shape new research questions.”

Follow up with:

“Now group those findings into three possible structures/outlines for a 3,000-word literature review.”

Using Undermind well requires the same discipline as any good literature review: clarity, curiosity, and critical distance. Let it do the reading; you do the reasoning.

It saves time, yes—but more importantly, it saves attention for the part that really matters: thinking.

Stage 3 — Organising Sources

Zotero: The Organised Mind of a Researcher

Free

If ChatGPT and Perplexity help you generate ideas, then Zotero is where those ideas finally find a home.It’s not just a reference manager — it’s your research brain, a place where every useful paper, quote, or idea gets captured, organised, and ready for later use.

When I started using Zotero, it changed how I approached research completely. Before that, I used to keep scattered PDFs on my laptop, saved bookmarks in my browser, and sticky notes with citation details I’d inevitably lose. Zotero ended that chaos.

Now, every time I find a useful source — whether from Perplexity, Undermind, or a random journal database — I save it straight to Zotero. It stores the full reference, metadata, tags, notes, and even a local copy of the PDF. With just one click, the citation is saved, searchable, and ready to insert into my writing.

How Zotero Works

At its core, Zotero is an open-source research management system that automatically detects scholarly content from your browser. When you visit a page with a journal article, book, or report, it captures the metadata — author, title, date, publisher, DOI, and URL — and stores it in your library.

From there, it does four powerful things:

1. Organises: You can create folders (called “Collections”) for each topic or module. Within them, you can tag entries with keywords like theory, methodology, or case study.

2. Annotates: You can write notes directly under each source — ideal for summarising key arguments or quotes you might use later.

3. Cites while you write: It integrates with Microsoft Word, Google Docs, and LibreOffice, letting you insert citations automatically in any format.

4. Syncs across devices: You can access your full library anywhere. It even lets you collaborate in group libraries with classmates or research partners.

   
   

Think of it as a personal, cloud-synced research archive that gets smarter as your projects grow.

How I Use It

Every essay, report, or assignment I now write passes through Zotero at some point. My process looks like this:

1. Capture every promising source: Whenever I find something useful — through Perplexity, Undermind, or even Google Scholar — I click the Zotero Connector icon on my browser. In seconds, it saves the full reference and PDF.

2. Organise immediately: I drag it into the right folder (e.g., Dissertation > Literature Review > Methodology).I then tag it with 2–3 keywords like qualitative, case study, healthcare.

3. Annotate for later: I add a short note under each source:“Explains link between primary care funding and service quality — might support argument in section 3.”

4. Write directly with Zotero: When I’m drafting in Word or Google Docs, I use the Zotero citation tool to insert references as I write. It automatically formats them — APA, Harvard, MLA, you name it.

5. Generate bibliography instantly: When I’m done, I click “Create Bibliography from Collection” and Zotero compiles every cited source in perfect formatting.

Why It’s So Valuable

• Zero formatting stress: You’ll never have to manually fix commas, italics, or brackets again.

• Total organisation: Keeps everything categorised by topic, making revision or re-submission far easier.

• Consistency: Every reference comes from the same reliable source — your Zotero database.

• Collaboration-ready: You can share a library with peers working on similar topics.

The beauty of Zotero is that it teaches good research habits. You start seeing patterns in what you’ve read — which authors dominate a field, how ideas have evolved, where the debates are centred.

How to Use Zotero Without Overreliance

Zotero doesn’t think for you — it organises your thinking. It’s tempting to treat it like a magic bibliography generator, but the real power comes when you use it actively: adding notes, grouping ideas, reflecting on the literature.

I often revisit my Zotero notes weeks later, and I’m reminded not just what I read, but why it mattered. That’s something no AI can replicate.

Prompt (Workflow) Idea

Step 1: Save all sources related to your assignment in Zotero.

Step 2: Create collections for Theory, Evidence, Methodology, and Critiques.

Step 3: Under each source, write a 2–3 sentence summary of its argument.

Step 4: When writing, use Zotero’s “Insert Citation” feature to build your bibliography automatically.

By the end of your project, Zotero becomes a snapshot of your entire research journey — every reading, reflection, and argument neatly stored in one place.

Research Rabbit: See the Bigger Picture

Free

Research Rabbit helps you visualise your reading. When I first discovered Research Rabbit, it felt like stepping inside a visual version of Google Scholar — except this time, I could see how knowledge flows. It’s an interactive graph that links papers by shared authors, citations, and themes. You start with one paper, and Rabbit shows you everything related — who cited it, who collaborated with those authors, and what newer work builds on the same ideas.

It’s like watching a living ecosystem of research unfold before your eyes.

How I use it:

• Start with one important paper in my field.

• Explore its “cited by” and “related works.”

• Follow author networks to find others working on the same issue.

• Connect it to Zotero to sync and store everything.

It’s especially useful when you feel like you’ve read everything but still suspect there’s something missing. Research Rabbit often finds that one paper that ties your argument together.

Stage 4 — Putting It All Together

Using these tools isn’t about automation—it’s about amplification. They help you shift from the repetitive parts of academic life to the parts that matter: analysis, creativity, and judgement.

Here’s how it all fits:

1. Start with ChatGPT to understand your task.

2. Move to Claude to test reasoning and logic.

3. Use Perplexity and Undermind for your research and literature review.

4. Organise with Zotero and Research Rabbit.

Final Thoughts

AI isn’t the enemy of learning—it’s the evolution of it. When used responsibly, it makes your work smarter, not shorter. The key is to stay in control: don’t let AI write for you, let it write with you.

If I had this entire toolkit back when I was studying, I know my writing would have been sharper, my references cleaner, and my arguments more balanced. But more importantly, I would’ve spent less time stuck in confusion and more time doing what really matters—thinking.

And even though I couldn’t use all these tools back then, I’ve started integrating them into my work now—and the difference is undeniable.I spend far less time buried in administrative or repetitive tasks and far more time on the parts that actually move ideas forward: designing better structures, refining arguments, and questioning assumptions.AI has shifted the balance of my work from collecting information to creating understanding.

It’s no longer about finishing faster; it’s about thinking deeper.That’s the real promise of AI—not to make learning easier, but to make it richer.