Skip to content

News

How To Find The Questions Your Customers Ask AI (And Your Keyword Tools Miss)

Customer calls, support messages and forum discussions can reveal specific questions that keyword tools may not show. Here’s how to collect those questions, check them against your site and turn them into a practical content plan.

19 November 2026 · 3 min read
A person typing a question into a laptop
"Workspace @ Nottingham Flat" by David Wellbeloved, CC BY, via Openverse. — David Wellbeloved · CC BY

Search volume can miss the detailed questions people ask when they need help choosing a service or solving a problem. A Search Engine Journal guide recommends looking at sales calls, support messages and forum discussions to find those questions, then checking which ones already reach your website and organising them into a content plan.

Start with questions people already ask you

According to Search Engine Journal, sales and discovery calls, support tickets, live chat logs and pre-sale messages capture people describing their situations in their own words. Those details can make a question more useful than a broad keyword: the guide gives an example involving a business’s platform, a drop in search rankings and uncertainty about the cause.

Gather recent conversations and identify questions people asked directly or implied. Keep constraints that affect the answer, such as the platform they use, what they have already tried, or their timeline. Group similar questions and note how often each group appears. Repeated scenarios can help a content team prioritise questions even when a keyword tool shows no search volume.

Use forums to broaden the picture

Your own conversations show what people who have found your business ask. Forums can add situations from people who have not. The guide names Reddit, Quora, industry Slack and Discord communities, Facebook groups, and review sites such as G2 or Trustpilot as possible places to look, depending on your market.

Read the opening post for the person’s circumstances and the replies for follow-up questions. Collect relevant threads, then extract and group questions in the same format as your customer conversations. That makes it easier to spot recurring needs and plan content around real situations.

Work out what a useful answer needs to cover

One customer question may contain several related issues. Search Engine Journal explains that Google’s AI Overviews and AI Mode may use “query fan-out” to run related searches while building a response. The guide describes using Gemini’s API with Grounding with Google Search to view a webSearchQueries field, which lists searches used to answer a prompt.

Treat those related searches as a coverage checklist, not as proof of what an AI feature searched for in a particular person’s session. For ChatGPT, Perplexity and other assistants, the guide suggests asking the question manually, noting the subtopics and cited sources, and comparing them with the page intended to answer it.

Check what is already reaching your website

Use Google Search Console as a validation step: it can show which conversational queries are already finding your site. In the Performance report, the guide recommends using the Query filter’s Custom (regex) option to look for queries that start with a question word. Its example pattern is ^(how|what|why|which|should|can|do|does|is|are|when|where|who)\s.

Bring the results together with the questions from calls, messages and forums. The resulting content plan can focus on the situations customers describe, the follow-up questions an answer should address, and the queries already bringing people to your site.

Source: searchenginejournal.com