Perplexity: the cited answer engine

0:00 Perplexity

Traditional search gives you ten blue links and leaves the reading to you. A general chatbot gives you an answer, sometimes without saying where it came from. Perplexity sits in between. It searches, reads and writes an answer with numbered citations. In this lecture you will learn what Perplexity is best at, how to ask it better questions, and how to read its citations like a professional.

0:29 Why an answer engine

Why use an answer engine? Because it combines the speed of an AI answer with the traceability of search. Every important sentence points to where it came from, so checking takes seconds instead of minutes. Think of it like a research assistant who hands you a summary with sticky notes on every claim, showing the page they got it from. You still decide whether to trust it, but you can check instantly.

1:00 What Perplexity offers in 2026

In twenty twenty six, Perplexity includes search and Pro Search for cited answers, with its own Sonar models and a choice of third party frontier models on paid plans. Deep Research for multi source reports, and ways to compare answers across several models. Spaces to organise research with files and instructions. Pages to turn research into shareable articles, and Labs to build reports and dashboards. And Comet, its free AI browser with an agentic assistant. Plans run from Free to Pro, Max and Enterprise, and the details change, so check the pricing page.

1:40 Best uses

What is it best at? Current facts and news with links. Finding primary sources quickly. Quick competitor and market scans. And research before a meeting. For long form drafting, complex reasoning over your own files, and writing in your own voice, a general assistant is usually better. Many professionals pair them. Perplexity to find and verify, an assistant to synthesise and write.

2:07 Simple example

A simple example. Ask Perplexity, is the Louvre Abu Dhabi open on Mondays? The answer cites the museum's official website. Click the citation, find the opening hours line, and confirm. Ten seconds. That is the whole habit. Answer, then source, then confirmation. When the stakes are low, you might skip the click. When you are repeating the fact to a client, you never do.

2:35 Better questions

Now, better questions. Best CRM gets you a generic listicle. Try instead, CRMs suitable for a ten person UAE agency that need WhatsApp integration and Arabic interface support, comparing pricing models and data residency options, using vendor documentation and reputable reviews from the last twelve months, as a table. Every clause narrows the search and raises the quality. Add time, place, source type and output shape, and your answers improve immediately.

3:06 Reading citations

Citations make checking fast, but they do not guarantee accuracy. Open the citations behind the claims that matter, and confirm the page actually says that. Check the date and the authority of the source. Watch for claims supported by just one source, because one blog is not a consensus. And treat any sentence without a citation as unverified.

3:31 The pre-call workflow

Here is a workflow worth saving. Before a sales call, ask Perplexity, using sources from the last twelve months, what the company does and who its customers are, recent news, launches, hires or funding, likely priorities relevant to your service, and three smart questions to ask, marking anything uncertain or single sourced. It takes a minute to run and five minutes to check, and it changes the quality of your first conversation.

4:02 Business example: Karachi consultant, Saudi retailer

A realistic example. A consultant in Karachi is preparing for a call with a Saudi retail chain. The pre call workflow returns news of two new store openings and a new head of e commerce, citing a regional business publication and the company's newsroom. He opens both sources and confirms the openings. He finds the hire was announced only on LinkedIn, so he notes it as single sourced. He walks into the call with three sharp questions and no embarrassing errors. In the call, one of his three questions, about how the new stores would handle online returns, led to a twenty minute discussion and a request for a proposal. Afterwards he saved the pre call prompt as a template in a Space for all his prospects, with instructions to prefer official company and regulator sources, so every future brief starts from the same standard.

5:05 Common mistakes

Four common mistakes. Treating the number of citations as proof of accuracy. Asking one word questions. Forgetting that sources can be outdated or promotional. And typing confidential client details into search queries, which can leave your control. Keep queries general and do the private reasoning in your approved workspace.

5:26 Comparing models

On paid plans you can switch the model behind an answer, and Perplexity has introduced ways to compare several models' answers to the same question. Use this for important questions. When models agree and cite the same primary source, your confidence can rise. When they disagree, go straight to the primary source. Remember, agreement is not proof, because they may all be reading the same flawed page, but disagreement is a reliable signal to dig deeper.

5:59 Narrow the sources

One more technique. Narrow where Perplexity looks. Ask for official sources only, or academic sources, or name specific sites, such as the regulator's website, and set a time window like the last twelve months. A narrower source set means fewer promotional pages and outdated posts in the mix, and far less work when you verify. Precision in the question saves time in the checking.

6:27 Watch me do it, part 1

Let me prepare the Karachi consultant's call with a Saudi retail chain. In Perplexity I type the pre call prompt. I have a sales call tomorrow with this retail chain in Riyadh, using sources from the last twelve months, tell me what they do and who their customers are, recent news, likely priorities relevant to e commerce consulting, and three smart questions, and mark anything uncertain or single sourced. The answer arrives with six numbered citations. It mentions two new store openings, a new head of e commerce, and flags that the hire is supported by one source only.

7:10 Watch me do it, part 2

I open the three citations that matter. The regional business article confirms both store openings. The company newsroom confirms the dates. The hire is only on a LinkedIn post, so in my notes I write, single source, ask carefully. Then a follow up. Using official company sources only, what have they said about e commerce plans? It returns a line from their annual report about growing online sales, with a citation. I open it and it matches. My three questions for tomorrow are now grounded in things the company has actually said.

7:50 Recap and try this now

Recap. Perplexity gives answers with sources, which makes verification fast. Ask specific questions with scope, time, place and source preferences. Open the citations behind anything important, watch for single sources and missing citations, and keep sensitive details out of queries. Try this now. Take your next real meeting, run the pre call prompt, open the three most important citations, and note whether each one fully supported the claim.