What is the difference between conversational search and a chatbot?

A chatbot can answer from what its model already knows. Conversational search fetches live web results for each turn and links to them, so the answer can cite current pages. ChatGPT search, Google AI Mode and Perplexity all work this way.

Is voice search the same as conversational search?

No. Voice is an input method. A voice query can be a single one-off search, and a typed query can be part of a long conversation. What makes search conversational is that the engine keeps the context of earlier turns.

Do follow-up questions trigger new web searches?

Usually, yes. Engines that search the web treat each follow-up as a new retrieval step, rewritten with the context of the conversation. That is why a page can be cited on the third turn even if it never appeared on the first.

How do I make my pages work for conversational search?

Answer the follow-up questions buyers ask in plain text: price, fit, compatibility, limits and comparisons. Make sure the search crawlers of each engine can reach those pages and that Google can show them with a snippet.

AI search glossary

Conversational search: asking follow-up questions instead of new queries

Conversational search is a way of searching where the user asks in natural language and keeps refining with follow-up questions, while the engine remembers the earlier turns and answers each one in context, usually with links to its sources.

Also called: conversational AI search, chat search

Why it matters for AI search

In a conversation, the buyer rarely stops at the first answer. The second and third questions (which one is cheapest, does it work with Shopify, what do people complain about) are where a shortlist turns into a pick. A brand that is only visible for the opening question can drop out of the conversation by the follow-up, so coverage of the follow-up questions matters as much as the first one.

How conversational search works

In classic search, every query starts from zero. You type "project management tool", scan ten links, then type a new query with more words. In conversational search, the second question builds on the first. You can ask "which of those work for a team of five?" and the engine knows what "those" means.

Google describes AI Mode as offering "the ability to go deeper through follow-up questions and helpful links to the web". ChatGPT search and Perplexity work the same way: each turn can trigger fresh web retrieval, and the answer carries links to the pages it used. Behind the scenes, the engine usually rewrites the follow-up into one or more full searches, the same mechanism described in query fan-out.

An example conversation

"What is a good email tool for a small online store?" The engine names four or five options with short reasons.
"Which of those integrates with Shopify and costs under $30 a month?" The engine searches again, now for integrations and pricing, and narrows the list.
"What do people complain about with the cheapest one?" A third search, this time pulling reviews and forum threads.

Each turn draws on a different set of pages. A brand that has a clear pricing page and a documented Shopify integration stays in the conversation. A brand that only has a strong homepage can be named in turn one and gone by turn two.

What it changes for your site

Follow-ups decide the pick. Opening questions produce a shortlist. The follow-ups (price, fit, limits) decide which name the person acts on.
Context travels. The engine carries constraints forward, such as budget or platform. Pages that state those facts plainly are easier to match.
Fewer clicks, more qualified ones. Many conversations end without a click, which is the pattern behind zero-click search. The clicks that do happen tend to come late, from people close to deciding.

What to do about it

Write down the three or four questions a buyer asks after the first one, and check that each has a crawlable answer on your site. Confirm that search crawlers such as OAI-SearchBot and PerplexityBot are allowed, and that Google can show your pages with a snippet, which its AI features require. Then test whole conversations, not single prompts, on engines like ChatGPT and Perplexity, and note at which turn you drop out.

Questions

What is the difference between conversational search and a chatbot?: A chatbot can answer from what its model already knows. Conversational search fetches live web results for each turn and links to them, so the answer can cite current pages. ChatGPT search, Google AI Mode and Perplexity all work this way.
Is voice search the same as conversational search?: No. Voice is an input method. A voice query can be a single one-off search, and a typed query can be part of a long conversation. What makes search conversational is that the engine keeps the context of earlier turns.
Do follow-up questions trigger new web searches?: Usually, yes. Engines that search the web treat each follow-up as a new retrieval step, rewritten with the context of the conversation. That is why a page can be cited on the third turn even if it never appeared on the first.
How do I make my pages work for conversational search?: Answer the follow-up questions buyers ask in plain text: price, fit, compatibility, limits and comparisons. Make sure the search crawlers of each engine can reach those pages and that Google can show them with a snippet.