Spaces and Deep Research
How to organize topic-based spaces and when to launch the long Deep Research mode.
Spaces
Spaces are "topic environments" in Perplexity. Inside a Space you fix a system prompt, sources, files, and preferences. Instead of a general search you ask inside the Space — and get answers strictly within its bounds.
Examples:
- Space "Competitors Q2-2026" — a list of domains, documents, a report template.
- Space "AI Regulation in the EU" — official documents, laws, guides.
- Space "Technical research on vector DBs".
Deep Research
Deep Research is a mode where the Perplexity agent spends tens of minutes, walks through hundreds of sources, and assembles a detailed report. Useful for:
- Competitive analysis in an unfamiliar segment.
- Preparing for a big meeting on a new topic.
- A regulatory scan.
Pitfall: sometimes Deep Research produces a large summary with duplications and a "safe" tone. Verify the sources and the final structure.
Create a Space for one of your recurring topics. Upload 3-5 sources. Run Deep Research and assess the result.
Copy and adapt to your context. Text in angle brackets should be replaced.
(In Perplexity, in a new Space) Space context: <…> Research topic: <…> Which report format: <…> Include opposing opinions and risks: yes / no Which sources to prefer: regulators / academic / industry
- They use Deep Research for every question — it's overkill.
- They don't narrow the Space — focus is lost.
- They accept Deep Research as a final report without verification.
- Use Spaces for recurring topics.
- Deep Research is best for preparing for serious meetings and for assessing markets.
- After Deep Research always — a separate human step: fact-check + condensing.
Recurring research tracks + occasional deep reports.
A simple quick search.