Agents · Lesson 02 Pro ~10 min read Persistent research

Spaces: research that lives somewhere.

Perplexity Spaces solve the problem with one-off AI searches: your research disappears when you close the tab. Spaces are persistent workspaces — they hold sources, custom instructions, and conversation history. They can be shared with a team. For any ongoing research project, Spaces is the right home.

Workflow 01 Create a Space for an ongoing project

1

When to use a Space vs. a regular search

Spaces are right for research that continues over days or weeks, or that you want to share. One-off searches stay in regular Perplexity.

The prompt that works

Space setupUse a Space when: • Research will continue over multiple sessions • Multiple people need to contribute or view • You want to add custom instructions or focus the AI's behavior • You need to upload specific documents as context • The topic benefits from accumulated context Don't use a Space when: • It's a single question you'd ask once • You're just looking up a fact • The topic is sensitive and shouldn't have a persistent record Space setup: 1. Name it clearly: 'Competitor analysis: Acme Inc' not 'Research' 2. Add a description so future-you (or teammates) understand the focus 3. Upload key reference documents at the start 4. Write a custom instruction that tells the Space how you want answers formatted

Best use cases

  • Ongoing competitive analysis
  • Multi-week research projects
  • Team research with shared findings
  • Reusable research templates (e.g., 'New market entry research')
Don't pile every research session into one Space — they get cluttered. One Space per distinct topic.
Time savings: Reusing accumulated context: 10-15 min saved per follow-up search.

Workflow 02 Custom instructions per Space

2

Focus the AI on what you actually need

Custom instructions in a Space change how Perplexity answers within that space. Format preferences, focus areas, what to avoid.

The prompt that works

Custom instructionsUseful custom instructions: For a sales-research Space: • 'Always include the company's revenue if findable' • 'Cite primary sources (company filings, official press releases) over secondary' • 'End every answer with one specific question I should ask in my sales call' • 'Don't include generic company descriptions — assume I know what they do' For a technical-research Space: • 'Prioritize official documentation and primary sources over blog posts' • 'Include code examples in answers when relevant' • 'Flag if information might be outdated (>12 months old)' For an executive-prep Space: • 'Keep answers under 200 words' • 'Lead with the bottom line; supporting detail after' • 'Cite sources but don't include URLs inline'

Best use cases

  • Role-specific research patterns
  • Format preferences (bullet vs. prose, length, tone)
  • Source quality requirements
  • Removing noise from generic answers
Too many instructions confuse the AI. Pick the 3-5 most important; revise from there.
Time savings: Each query returns more usable output, less editing needed.

Workflow 03 Sharing Spaces with your team

3

Collaborative research patterns

Shared Spaces enable team research where everyone contributes. Especially useful for sales teams, research teams, and any context where multiple people benefit from one knowledge pool.

The prompt that works

Team sharingSharing patterns that work: • **Team competitor Space** — one per major competitor, all sales reps contribute. New questions and findings accumulate. • **Account-specific Spaces** — one per top-tier customer or prospect, shared with the account team. • **Project research Spaces** — one per active project; PM, eng, design, and PMs all contribute. • **Industry intelligence Spaces** — one per market or vertical; product marketing maintains it. Governance: - Designate an owner per Space (responsible for hygiene) - Review quarterly for relevance — archive stale ones - Use clear naming so Spaces are findable - For sensitive research, use private Spaces, not shared

Best use cases

  • Sales-team competitive intelligence
  • Cross-functional project research
  • Industry watch (multiple analysts contribute)
  • Customer success knowledge bases
Shared Spaces are public to whoever you share with. Don't put confidential customer data in a Space without considering who has access.
Time savings: Team knowledge that compounds vs. individual searches that vanish.

Build a Space for your most active project

Pick the work topic you research most often — a customer, a competitor, an ongoing initiative. Create a Space for it. Upload key docs. Write custom instructions. Use it for a week and notice how much faster follow-up questions are.

Back to Perplexity Mastery → Try a prompt in the playground →

What you can do now

  • Create Spaces only for research that extends beyond a single session
  • Write clear custom instructions to focus the AI per Space
  • Designate owners for shared Spaces with quarterly review
  • Archive stale Spaces — clutter degrades search experience
  • Use private Spaces for anything sensitive
Pro
Up next in Perplexity Mastery

Lesson 03 · Deep Research — Perplexity's analyst mode

Perplexity Deep Research is the long-running research agent — now built on Claude Opus 4.5. Multi-source synthesis, deliverables include presentations and spreadsheets. See the track →