Now that we have cleared all doubts, we can start. But where?
The setup phase is the phase where we are all feeling the most lost during the whole process. We know kind of the topic we want to write about, but we don’t know where the f we should start looking for information.
Looking at my topic, I have been stuck for a long time in this space between art, design, urban studies, photography, sociology and “I swear this matters, please trust me.”
Before I start going deep into keywords and literature, I wanted to get an overview of something else first:
AI tools.Because I knew ChatGPT exists, but I did not realize how many other tools are out there that can actually help during thesis research.
(BIG SHORT DISCLAIMER)
I’m very aware AI can make people lazy. So my rule is:
AI can help me find, compare and brainstorm, but it shall not do my thinking for me. If it gives me a claim, I verify it. If it summarizes a paper, I still read the paper.
Okay, let’s start.
I think the easiest way to think about AI tools is in categories:
- search
- map
- read
- write
- cite
- analyze
- create
This is the toolbox I’m building right now:
A) Literature discovery (search & find papers)
Elicit
Good for: jumping into a topic and quickly finding relevant papers, plus helping with screening/extraction.
How I’d use it: Start with queries like “context effects perception art public space” and let it propose other related papers.
Consensus
Good for: getting a research-backed overview quickly, with citations tied to papers (so you can trace everything).
How I’d use it: When I need a fast sense-check like: “Is there literature on how labels change perception?” Then I click through to the actual papers.
scite
Good for: checking how a paper is cited (supportive vs contrasting vs neutral).
How I’d use it: When I find a “key theory” paper, I can quickly see whether later work supports it or fights it.
B) Literature mapping
Connected Papers
Good for: visual graphs of related papers starting from one seed paper. Great when you have one perfect source.
ResearchRabbit
Good for: interactive maps + recommendations + tracking authors over time. It learns what you’re collecting and helps you expand it.
Litmaps
Good for: building and monitoring a literature map so you can see how research clusters connect and keep getting updates as new papers appear.
How I’d use these: Pick 2–3 “seed papers” (Duchamp/context theory, public space perception, photography framing), then map outward until patterns emerge. Once I see the clusters, I know what my literature chapter needs to cover.
C) Reference management + annotation
Zotero (+ plugins)
This was an obvious one, right?
Zotero is the base layer. Plugins can extend it (better workflows, translating, export setups, etc.).
How I’d use it:
- store everything
- tag papers by theme cluster
- export citations cleanly when writing
D) Reading support (summaries)
This is where general AI chat tools can help carefully.
ChatGPT (file upload + analysis features depending on plan)
Helpful for: summarizing a paper you upload, extracting key arguments, turning messy notes into structure, generating keyword variations, comparing two theories.
My rule: I still read the original PDF, make my own notes and then compare with what it summarized. AI can be confidently wrong in a very convincing voice lol.
E) Transcription + interview workflow
Since I’m planning interviews, I need tools for:
- recording
- transcribing
- coding themes
Even if I don’t choose a specific transcription tool yet, I know the pipeline:
audio → transcript → coding → quotes in thesis
AI becomes useful when it speeds up transcription and helps me find repeated themes but I still decide what counts as meaningful.
F) Writing + language polish
For writing support, I want tools that help with:
- clarity
- grammar
In case anyone of my student colleagues has another amazing AI tool please help a girl out and comment under my blogpost.
Lots of love and bussis
-Fiona
