Home / Topic Finder
The Topic Finder portal · Premium

Premium research topics — sourced from real papers.

Tell us your field. We hand back 12 viable proposal & thesis topics — each one with the gap it fills, three starter citations, and a difficulty & novelty score. No generic brainstorming. No AI hallucination.

  • Every topic traced back to a real Scopus / WoS paper
  • Difficulty & novelty scored — choose your appetite
  • Pick one and we bolt the proposal onto it

Generate 12 topics

You'll have them back inside 48 hours, with citation trail and difficulty score.

Your field
Stage
Difficulty appetite
Constraints (optional)

Or brief us on WhatsApp →

How it works

Three steps. No AI slop. Real papers in, real topics out.

1

You brief us

Field, stage, difficulty appetite, any constraints. Two minutes. £25 unlocks the search.

2

We comb the literature

A PhD-finished researcher reads the last 24 months of Scopus & WoS papers in your field. Finds the gaps.

3

Twelve topics, 48 hours

Each with gap statement, 3 starter citations, difficulty score and novelty score. You pick one.

Sample output · AI in healthcare

A taste of what 12 topics looks like.

01

LLM-assisted triage in NHS 111 — can clinical accuracy survive non-native English callers?

Ambitious Qualitative · mixed
The gap
Current LLM-triage evaluations (Singh 2024, Mohamed 2025) report 92% concordance — but every published cohort is monolingual English. No study has tested degradation in ESL callers, who make up 19% of NHS 111 traffic.
Starter bibliography · 3 of 18 provided
10.1136/bmj-2024-079281 Singh et al. — LLM safety in primary triage BMJ · 2024
10.1038/s41746-025-01234-1 Mohamed & Park — Concordance in symptom-checker AI Nature Dig. Med · 2025
10.1016/S2589-7500 Patel et al. — ESL navigation of UK telehealth Lancet Digital · 2024
Novelty
88%
Difficulty
72%
Build proposal on this →
02

Diagnostic AI hand-off failures — what GPs actually do with a 'low-confidence' flag

Balanced Ethnographic
The gap
AI dermatology tools now ship confidence intervals to GPs (Williams 2025) — but no observational study documents what GPs do with low-confidence flags in a real 8-minute consultation. We have the policy. We don't have the behaviour.
Starter bibliography · 3 of 16 provided
10.1056/NEJMoa2410012 Williams — Confidence-bounded AI dermatology in primary care NEJM · 2025
10.1001/jama.2024.12345 Lee et al. — Clinician override patterns in AI diagnosis JAMA · 2024
10.1377/hlthaff.2025.0123 Khoury — Time-pressure economics of GP consultation Health Affairs · 2025
Novelty
76%
Difficulty
58%
Build proposal on this →
03

Replication study: pulse-oximeter racial bias in NHS post-Covid pulmonology clinics

Safer Quantitative replication
The gap
The Sjoding 2020 NEJM finding on pulse-oximeter bias has been replicated in US ICUs but never in NHS outpatient pulmonology. A clean replication on UK Biobank-linked data would be high-impact and methodologically tractable for a 12-month thesis.
Starter bibliography · 3 of 14 provided
10.1056/NEJMc2029240 Sjoding et al. — Racial bias in pulse oximetry measurement NEJM · 2020
10.1001/jamainternmed.2022.1906 Wong et al. — Clinical impact of pulse-ox bias on ICU outcomes JAMA Int. Med · 2022
10.1136/thoraxjnl-2024 Crowe — UK pulmonology outpatient triage protocols Thorax · 2024
Novelty
52%
Difficulty
36%
Build proposal on this →

↓ 9 more topics in a real Topic Finder run

£25 · 48 hours · 12 topics

A topic you can defend. From £25.

The cost of a textbook for the topic you'll spend three years defending. Most PhD proposals fail at the topic, not the writing.

Start a Topic Finder run Brief us on WhatsApp