Perplexity for Professionals: Spaces, Pages & Deep Research

Part 1 covered what Perplexity is and how it compares to Google and ChatGPT. This article is for those who’ve moved past the first experiment and want to make Perplexity a genuinely productive part of their workflow. We cover the tools that make it different in a daily work environment: how to organize accumulated research, how to produce cited reports in minutes, and how to choose the right model for each task.

① Basic Search vs. Pro Search — The Practical Difference

Before getting into advanced features, there’s a fundamental distinction worth understanding between the two core search modes:

	Basic Search	Pro Search
Search method	Single direct search	Multiple sequential searches, each adjusted based on what the previous round found
Sources consulted	5 – 10 sources	20 – 50 sources
Asks clarifying questions?	No	Yes — may ask for clarification before starting
Best for	Fast, direct questions	Comparisons, analysis, multi-faceted topics
Available free	Unlimited	5 – 10 per day only

The practical rule: Basic Search for “what is” and “when” questions; Pro Search for “why,” “how do these differ,” and “what’s most accurate.”

② Deep Research — From One Question to a Cited Report

Deep Research is Perplexity’s most distinctive feature. Rather than returning an answer in seconds, it takes between two and four minutes to do the following:

1. Breaks your question into more precise sub-questions
2. Runs 20 to 50 sequential searches, each building on the previous round’s findings
3. Reads and weights sources against each other
4. Assembles a structured report with headings, paragraphs, and a clickable reference list

Perplexity Deep Research achieved a 21.1% accuracy score on Humanity’s Last Exam — a benchmark for complex multi-step reasoning tasks — outperforming models including o1 and DeepSeek-R1 on that specific benchmark. The number doesn’t mean it’s error-free; it indicates a meaningful edge on complex, layered research tasks.

How to Activate It

Go to perplexity.ai → from the mode selector below the text box, choose Deep Research → write your question → wait 2–4 minutes.

Deep Research Prompts for Translators and Writers

Necessary reminder: Deep Research carries the same error rate noted in Part 1. The report it produces is a documented starting point, not a finished product — always verify claims against the original sources before citing anything in published work.

③ Spaces — A Research Environment That Remembers Everything

The problem Spaces solves is simple but painful: you worked on a topic for two weeks, ran twenty searches in Perplexity, then lost the thread entirely. Spaces are persistent research projects — you open them any time and find every previous conversation, uploaded file, and custom instruction ready exactly where you left them.

Capability	What It Means in Practice	Plan Required
Persistent custom instructions	Write once: “Focus on academic sources after 2020” — applied to every search inside this Space	Pro
File uploads and analysis	Upload a translation draft or reference text and analyze it alongside live web research simultaneously	Pro
Cross-session context memory	Yesterday’s research builds today’s — no need to re-explain the project every session	Pro
Invite collaborators	Add a proofreader or project partner to see the same research and contribute to it	Pro

Ready-to-Use Space Instructions for Translators

To access: from the left sidebar in Perplexity → Spaces → New Space.

④ Pages — Turn Research Into a Shareable Report

Pages converts any research thread into a cleanly formatted published page shareable via a direct link. Practical uses for content professionals:

• Send a research summary to a client without manual formatting in Word
• Share a cited reference with a colleague or language editor
• Keep a readable version of a long research thread instead of copying the raw text

After any completed research thread → click Share → Create Page → add a title → publish. The page gets a public link you can share or keep private. See Perplexity’s Pages gallery for published examples.

⑤ Model Council — The Right Model for Each Task

Model Council launched in early 2026, with memory added later the same year. Rather than one model for everything, you choose the most appropriate model from inside Perplexity itself:

Model	When to Choose It	Plan
Default (Sonar)	Everyday fast research — the quickest and most efficient for most tasks	Free + Pro
GPT-5 / GPT-5.2	Text tasks requiring high linguistic precision and polished phrasing	Pro
Claude (Anthropic)	Long-form text, complex context analysis, document summarization	Pro
Best (automatic)	Let Perplexity decide — selects the model based on query type	Pro

The added memory feature means Perplexity remembers your model preferences across sessions — you don’t need to re-select every time if you have a consistent preference.

⑥ File Uploads — Combine Your Sources With the Internet

On Pro and higher plans, you upload PDFs, Excel spreadsheets, and images, then ask questions that draw on both the file content and live web research simultaneously:

File	Question You Ask	What Perplexity Does
Translation contract (PDF)	Which clauses differ from GALA standards for translation contracts?	Reads the contract, searches GALA standards, compares them
Industry report (PDF)	Summarize the key figures and compare them with the most recent available data	Combines what’s in the file with current web findings
Article draft	Are the statistics and sources cited in this draft accurate and current?	Fact-checks each claim against the live web

A Complete Workflow: From a Translation Brief to a Cited Report

Suppose you’re translating a report on the Gulf translation services market and need a documented background before you start:

1. Create a new Space titled “Gulf Translation Market 2025-2026” and set your custom instructions
2. Upload any reference files you have: the client brief, domain terminology, required writing style
3. Start with Pro Search for a broad question: “What is the size and main trends of the translation services market in GCC countries?”
4. Activate Deep Research for the deeper question: compare Saudi Arabia, UAE, and Qatar on demand for specialized translators in legal and technical fields
5. Create a Page from the results and share it with your editor or client for review
6. Go to the original sources and verify figures before including them in the final translation

The principle that holds at every step: Perplexity saves you time finding sources. You protect credibility by verifying them.

---

The Complete Series

• Part 1: What Is Perplexity AI? And When Does It Beat Google? — The three-way comparison, pricing, and the necessary warning
• Part 2: You’re here — Spaces, Pages, Deep Research, and Model Council

---

Sources: Perplexity — Introducing Deep Research | perplexity.ai/spaces | perplexity.ai/pages | Perplexity Changelog