A Search Console query such as "yes go on" looks useless if you treat keyword research as a list of standalone phrases. It becomes much more interesting if you treat search as a conversation.
Search Engine Roundtable reported that marketers are seeing query strings including "yes," "yes go on," and "yes, pricing" in Google Search Console. Google's John Mueller pointed to AI Mode documentation explaining that follow up questions are effectively counted as new queries. The main performance report can therefore contain pieces of a conversation whose earlier context is invisible to the marketer.
This does not make keyword research obsolete. It changes the unit of research. A modern keyword plan needs to understand the first question and the likely next questions. That is a natural extension of long tail keyword research, but AI Mode makes the sequence itself more important.
The keyword is becoming one turn in a longer journey
For years, SEO teams grouped keywords by topic and intent. That remains useful. What AI Mode adds is continuity. A person can move from discovery to comparison to pricing inside one interface without restating the topic every time.
Consider a buyer researching CRM software. The first prompt might ask which systems suit a ten person consulting firm. The next prompt could ask which options integrate with Outlook. The next could be "compare the first two." The next could simply be "pricing." Search Console may expose one of those later fragments without the chain that came before it.
Classic keyword research would try to estimate volume for each self contained phrase. Conversational research asks a different question: what information does someone need next after receiving an answer?
That distinction matters because a page can rank for the opening query and still fail the rest of the conversation. A buyer may need proof, comparison, implementation details, price ranges, limitations, alternatives, or a clear next step. If those answers are absent, an AI system can retrieve them from other sources.
Build question chains around commercial intent
The most practical response is to turn keyword clusters into question chains. Start with the main discovery query, then map the logical follow ups that would move a buyer toward a decision.
A page about an accounting platform, for example, may need to answer what it does, who it is for, what it costs, how implementation works, which alternatives exist, what integrations are supported, and where the limitations are. Those topics are not filler. They are the likely branches of a conversation.
This approach is closely related to answer engine optimization, where clarity and direct answers matter because systems may retrieve individual passages rather than evaluate the article only as a whole.
The content architecture also matters. A site can distribute the answers across product pages, comparison pages, FAQs, case studies, guides, and pricing pages. The goal is not to cram every possible follow up into one enormous article. The goal is to make the next useful answer easy to find and easy to understand.
Search Console should become a prompt research input
Search Console remains one of the best sources of first party search demand because it reflects queries that actually generated impressions for your site. AI Mode makes the dataset noisier, but that noise can reveal emerging behavior.
Instead of deleting short conversational fragments from reports, classify them. Look for confirmation language, follow up modifiers, comparison terms, pricing language, implementation questions, and requests for more detail. Then connect those patterns to the pages that received impressions.
The SEO reporting model also needs to evolve. A report that only lists top keywords can hide the change in how people search. Add a section for conversational query patterns. Track whether those patterns are increasing. Identify which pages appear in those journeys and whether they support the next likely question.
This will not give you a complete transcript of AI Mode. Google does not expose that. It can, however, help content teams build a more realistic map of the buyer's information journey.
Stop treating every query as a page title
AI Mode makes an old SEO mistake even more obvious: a query is evidence of demand, not an instruction to create a page with exactly that phrase in the title.
Nobody needs a page optimized for "yes, pricing." The marketer needs to understand that pricing is a likely follow up in a conversation where the product has already been introduced. The answer may belong on a pricing page, a comparison table, a product overview, or a concise FAQ.
This is where content briefs should become more useful. A strong brief can include the main query, related intents, expected follow up questions, required evidence, internal links, and the role each section plays in the buyer journey.
The question for writers is no longer only "What keyword are we targeting?" It is also "What will the reader probably ask after this section answers the first question?"
Keyword research is becoming conversation architecture
The search industry has spent years improving keyword clustering. AI Mode suggests the next step is conversation architecture.
The process starts with familiar data: Search Console, customer questions, sales calls, support tickets, competitor pages, and search results. The difference is how the data is organized. Instead of stopping at groups of semantically related terms, marketers can arrange questions in a sequence that reflects how understanding develops.
That sequence can then shape site structure. Discovery questions lead to educational pages. Comparison questions lead to alternatives and versus pages. Trust questions lead to evidence and case studies. Commercial questions lead to pricing and conversion pages.
The "yes go on" query is therefore more than an oddity in Search Console. It is a reminder that the user may already be several turns into a decision process when your content is retrieved. Keyword research that ignores those turns will increasingly miss how search actually happens.
Originally published on the Mustard Seed blog.
Top comments (0)