Perplexity

What Perplexity is

Perplexity is an AI tool that answers a question with a synthesised response and cites the sources it drew on. It does not return a list of results for you to work through. It reads across sources, writes one answer, and shows where the answer came from. That is why it is often called an answer engine, and it is the sense in which the site uses the word.

On this site it counts as a chatbot and as one example of generative AI, the wider category of systems that produce new content instead of only sorting or finding existing content. It rests on a large language model like the others. What sets it apart is the emphasis on showing its sources.

Two places it appears on this site

Perplexity appears in two separate parts of the site, and the difference between them is the useful thing to know about it. In the courses it is one of the chatbots worth knowing. In the work on AI citation it is something else: a system whose behaviour is being measured.

That second role is unusual. The other assistants on the glossary are tools a team learns to use. Perplexity is also a tool whose output a business wants to appear in. Whether it names your business when someone asks a question in your field is a fact about your visibility, and the site treats it as one.

In the Courses

Perplexity has no course of its own. The catalogue has courses of their own for three tools, ChatGPT, Microsoft 365 Copilot and Mistral Le Chat, and Perplexity is not among them.

It appears as a comparison. The ChatGPT Enterprise course lists it beside Gemini, Claude, Copilot, Mistral, Grok and DeepSeek. The catalogue overview says the ChatGPT and Copilot sessions compare their tool against Gemini, Claude, Perplexity and DeepSeek. The aim is that a team leaves able to tell which job belongs to which tool, and does not memorise one interface.

Why Answers with Sources change the picture

A list of results asks the reader to do the reading. An answer with sources does it for them, and that changes what visibility means for a business. If people get their answer from the synthesis, the question is no longer where your page ranks. It is whether the answer names you.

The mechanism underneath is retrieval-augmented generation: pulling a live source into an answer instead of relying only on what the model learned in training. That is the technique behind a tool that cites, and it is close kin to grounding, which ties an answer to specific source material. A tool that shows its sources gives you something to check, and that is part of its appeal.

How it behaves differently from the others

The site’s article on ranking reports what Ahrefs measured in 2025, in data collected in July of that year. Google’s own AI Overview cited a page in the top 10 on Google for the same query 76% of the time. ChatGPT, Gemini and Copilot did so 12% of the time on average. Perplexity sat in between, at 28.6%.

So Perplexity leans on conventional ranking signals more than the other standalone assistants do, and behaves closer to Google’s summary than to them. Two points keep the figure honest. It is one dataset from one period, and the article notes that the AI Overview figure has already moved a great deal since. And the article’s central lesson is that ranking on Google and being cited are separate outcomes that move independently.

The Research behind getting cited

The site’s article on what gets content cited rests on one peer-reviewed study, presented at the ACM SIGKDD Conference in 2024 by a team from Princeton University, Georgia Tech, the Allen Institute for AI and IIT Delhi. It tested nine content techniques against a simulated generative engine and validated its strongest findings on Perplexity as a real-world check, not only as a simulation.

Three techniques gave the largest gains. Adding statistics raised AI visibility by 41%. Adding direct quotations raised it by 28%. Citing external sources raised it by 115% for lower-ranked content specifically. Keyword stuffing produced negligible or negative effects, and adding more words on its own produced no improvement. The article’s reading is that the signal is data density and source credibility, not length.

The same article reports that the overlap between top-ten Google results and what AI systems cite has fallen from around 75% in mid-2025 to between 17% and 38% in early 2026, and that brand mentions correlate about three times more strongly with AI visibility than backlinks do. It is careful to add that none of this replaces conventional SEO. The techniques sit on top of it.

Measuring whether it names you

None of that is worth much without a way to see whether it changes anything. Am I Cited, the tool this business partners with, monitors how AI chatbots such as ChatGPT, Perplexity, Gemini and Claude mention or cite a brand, product or store when answering a buyer’s question. It reports how that visibility compares against competitors.

The ranking article makes the underlying point plainly. Rank tracking answers a narrower question than it used to, and citation frequency needs its own tracking. It also recommends a habit that needs no tool: check by hand, now and then, what an assistant actually says for the few queries that matter most, and which sources it names while saying it. A page that used to rank first can find a competitor cited in its place, and a rank tracker will not show it. The partners page sets out where this work sits.

A Citation is not proof

Using Perplexity, as opposed to appearing in it, needs one habit above the rest. A cited answer looks more trustworthy than an uncited one, and that is the risk. The hallucination page explains that a model completes a pattern and does not check a record, so a properly formatted citation can sit on top of a claim the source does not make.

The courses treat this as a working habit. Ask for the source, open the source, and ask whether the claim is actually in it. The answer is a draft, and the person who sends it on is responsible for what it says. A tool that shows its sources makes that check easier to do. It does not do the check for you.

Mix-ups worth avoiding

Four confusions come up often enough to name. The first is treating a citation as verification, when it only shows where the answer says it came from. The second is treating ranking first on Google as a guarantee of being named by an assistant, when the data the site reports says otherwise for every assistant it measured.

The third is treating Perplexity as a course, when it is one of the assistants the courses compare. The fourth is treating an answer engine and a chatbot as opposites. Perplexity is a chatbot that leads with its sources, and the same checking applies to it as to any other.

Where it goes wrong

Perplexity can produce an answer that reads well and is wrong, like every generative AI tool. The presence of sources narrows the distance between an answer and its evidence, but it does not close it. Only someone who opens the source can see whether it says what the answer claims.

The larger the decision resting on the answer, the more the check matters. For anything a client, a regulator or a colleague will rely on, the answer is where the work starts and not where it ends.

Perplexity compared with other Assistants

The site does not rank the leading assistants, and the reason is stated in the catalogue. Which tool suits which job depends on the work and on what an organisation already has. Microsoft 365 Copilot answers from content inside your own applications. Gemini does the same inside Google Workspace, and is also the model family behind Google’s AI Overview. Claude is a hosted assistant reached on its own. Perplexity is the one built around showing where an answer came from.

The site publishes nothing on Perplexity’s own pricing, plans, models or data terms, so this page states none. A business considering it for regular use should read those terms directly for the account it would use, in the same way it would for any other assistant.

Where it is taught and where it appears in an Audit

The ChatGPT Enterprise course lists it among the leading chatbots, and the core courses page explains how the comparison runs across the catalogue. Copilot Essentials is one of the six core courses, and the rest sit in the wider training catalogue.

An audit meets Perplexity from the other side. An AI audit checks whether a site’s content is structured to be cited and not only ranked, run through the five-step audit methodology and starting with a discovery call about where a site stands now. Once you know what an assistant says about you, the work is to improve it, and to keep measuring.