Perplexity
Cited-source AI research assistant for desk research you need to be able to defend
Perplexity answers questions by searching the live web and citing the sources it used, so every claim can be followed back before it reaches a briefing note. For comms work that changes the calculation: the value is less the speed of the answer than the fact that the working is shown. It suits desk research, competitor scanning, first-pass fact-finding and quick sense-checks of a claim you have been handed. It is also one of the answer engines the GEO tools in this directory exist to monitor, which is a reason to understand it from the inside.
For comms teams: desk research where you need to check the source, not just the answer.
Best for
- • Desk research where you need to check the source, not just the answer
- • Competitor and issue scanning
- • Comms teams testing what AI answer engines say about their sector
Pricing
Free; Pro $20/month or $16.67/month billed annually; Enterprise Pro around $40/seat/month; Enterprise Max considerably higher.
Figures gathered from secondary sources in September 2026. Confirm current tiers at perplexity.ai before quoting. Enterprise tiers add SOC 2 Type II, MDM deployment, domain-level agent controls and per-session audit logs.
Pricing last verified Sep 2026 – check vendor site for current rates.
What works well
- + Citations make claims checkable rather than asking you to trust the answer
- + Fast at the scanning stage of a piece of research
- + Enterprise tier has the controls a regulated team will be asked about
Watch out for
- - Citations can be confidently wrong, or point to a source that does not support the claim
- - Free-tier data handling is not appropriate for confidential client material
- - Summarises what is published, so it inherits the biases of whatever ranks
When to choose this
Choose Perplexity when you are researching the open web and need to see and check the sources behind an answer before you use it.
Who should look elsewhere
If your research is against documents you already hold (annual reports, consultation responses, transcripts, policy papers), NotebookLM keeps the model grounded in your own sources and will drift less. If you need drafting and reasoning rather than search, Claude or ChatGPT are the better seat. And if the material is confidential, check which tier you are on before you paste anything into it.
Key features
- • Cited answers with links to every source used
- • Follow-up questions that hold the thread of a search
- • File and document upload on paid tiers
- • Focus modes for academic, social and general web sources
- • Enterprise controls including SSO, audit logs and domain-level restrictions
Governance & data
Governance information is based on publicly available vendor documentation. Verify with your vendor before procurement decisions.
Integrations
Available on
More in Research & Intelligence
Deciding whether to adopt it?
Three free templates take you from a review to a decision you can defend, and a record of it.
Making the call across a whole team or tool stack? Deploy Comms With AI builds the policy and the register with you.