Ten Hidden Perplexity AI Features Every European Healthcare and Research Professional Should Know

Focus Modes and Model Selection: Shaping Intelligence Before You Ask

The row of topic shortcuts beneath Perplexity's search bar is not decoration. Options including Academic, Writing, Math, and Health fundamentally alter how responses are generated. Each focus mode prioritises different source types and adjusts the underlying reasoning approach accordingly.

For a clinical researcher at, say, the Karolinska Institute or a policy team at the European Medicines Agency, this matters immediately. Selecting 'Academic' mode weights peer-reviewed sources more heavily and structures answers with greater analytical depth. 'Health' mode prioritises medical journals and verified health authorities rather than consumer wellness content. 'Writing' mode draws from style guides and editorial resources when drafting regulatory submissions or briefing papers.

The model picker adds another layer. Perplexity allows users to select which underlying AI powers a given response, and the performance differences are substantial. Faster, lighter models handle quick factual lookups well. More capable models handle nuanced synthesis and multi-step analysis. Critically, regardless of model choice, Perplexity maintains its citation-first approach throughout.

Conversational Context That Actually Remembers

Traditional search engines treat every query as an isolated event. Perplexity does not. It maintains conversational threads, allowing researchers to start with a broad question and progressively tighten focus without restating background context on every turn.

For European healthcare professionals, this is particularly useful when navigating interconnected regulatory questions. You might open a thread by asking about AI medical device classification under the EU AI Act, then drill down into how the Medical Device Regulation interacts with those provisions, then compare the UK's Medicines and Healthcare products Regulatory Agency approach post-Brexit, all within the same thread and without losing the original context.

According to Perplexity's published documentation, the platform maintains context across up to 20 follow-up queries within a single thread. That ceiling is high enough to support a substantial research session without interruption.

Citations That Enable Real Verification

Perplexity numbers each citation and links directly to source material. This is not academic courtesy for its own sake; it is a practical verification mechanism that distinguishes the platform sharply from tools that generate confident-sounding text without traceable provenance.

Yann LeCun, Chief AI Scientist at Meta and a prominent European-rooted voice in AI research discourse, has repeatedly argued that verifiability should be a baseline expectation of any AI system used in professional contexts. Perplexity's citation architecture goes some way toward meeting that bar. Each claim traces back to its origin in two clicks. For healthcare teams where a misattributed statistic in a regulatory submission carries real consequences, that transparency is not a nice-to-have feature.

The citation layer also surfaces source quality. Users can distinguish between a preliminary conference abstract and a peer-reviewed systematic review, between a government consultation document and a finalised directive. That distinction is routine in clinical research but is largely invisible in conventional AI chat tools.

Advanced Query Techniques That Surface Hidden Insights

The most underused capability in Perplexity is the ability to interrogate its own sources rather than simply summarise them. Instead of asking for a summary of the literature on, for example, AI diagnostic accuracy in radiology, try asking what the cited sources disagree on, or which perspective is most critical of the prevailing consensus.

Joanna Bryson, Professor of Ethics and Technology at the Hertie School of Governance in Berlin and one of Europe's most cited AI policy researchers, has consistently argued that surfacing disagreement and uncertainty in AI outputs is more valuable than presenting false consensus. Perplexity's query architecture supports that principle directly when users know how to invoke it.

A practical set of advanced query approaches for European healthcare and research professionals includes:

• Ask Perplexity what its cited sources disagree on to surface genuine expert debate rather than averaged opinion.
• Request a timeline view of a complex regulatory or clinical development to understand how positions have shifted.
• Use it as a comparative evidence engine by asking it to contrast recent trial results with citations attached.
• Ask it to identify where evidence is weak or contested to flag areas requiring independent verification.
• Request the same answer framed for different audiences, such as a clinician versus a patient versus a regulator, to stress-test clarity.

Research Collaboration Over Simple Answers

The most significant shift available to Perplexity users is conceptual rather than technical: treating the platform as a research collaborator rather than an answer dispenser. Prompts such as "What questions should I be asking about this topic?" or "What is missing from this discussion?" turn Perplexity into an active thinking partner rather than a passive retrieval mechanism.

This approach is particularly valuable during the early stages of a research project, a funding application, or a regulatory impact assessment, when the most important unknowns are the ones you have not yet thought to ask about. The platform is reasonably good at flagging knowledge gaps, provided users prompt it to do so explicitly rather than waiting for it to volunteer the information.

For EU healthcare teams working with datasets that span multiple languages and regulatory environments across member states, this gap-identification function has practical value. Understanding what remains uncertain or contested in the literature is frequently as important as confirming what is established.

Putting the Features Together in a European Context

The EU AI Act, which entered into force on 01/08/2024, classifies AI systems used in medical diagnosis and treatment as high-risk applications, triggering requirements around transparency, human oversight, and data governance. Separately, the UK's AI Safety Institute has been assessing frontier model capabilities with particular attention to outputs in sensitive domains including healthcare.

In that regulatory climate, the features that matter most in Perplexity are precisely the ones most users skip: citation verification, source quality differentiation, uncertainty flagging, and model selection. These are not power-user luxuries. They are the baseline behaviours that responsible AI use in clinical and policy contexts requires.

The gap between casual Perplexity use and informed use mirrors the gap between reading a Wikipedia article and conducting structured literature review. The interface looks the same from the outside. The outputs and their reliability do not.