In my final article, I made the case that search quantity is screening out your finest satisfied opportunities. The abbreviated type is that the highest-intent questions your buyers ask are increasingly being stated to an AI assistant, and they rarely display up in a keyword tool. Near the end of that piece, I listed a few places to discover those questions instead: People Also Ask, forums, your own revenue calls and assistance tickets, and the assistants themselves. That catalog was lone four projectile points, and the most average answer I’ve had since is several type of “OK, but how do I really do that?”
So, this is the applicable follow-up. I’ll go through how I collect these questions, how I use Search Console to inspect which ones are already reaching a site, and how I rotate the entire lot into item a satisfied squad can activity from.
Start With Your Own Calls And Tickets
I continually commencement here, since it’s the origin most teams overlook, and it’s normally the finest one. Sales calls, finding calls, assistance tickets, live conversation logs, and the messages group dispatch before booking a demo are all group describing their issue in their own words. That’s extremely near to how they’d depict it to ChatGPT or Gemini.
Here’s a genuine example from a latest finding call of quarry (anonymized, alongside the particulars changed adequate that you couldn’t acknowledge the business). The potential was a furnishings retailer, and this was approximately how they opened the call:
“We moved to Shopify a brace of years ago and misplaced a big chunk of our rankings. Some of it came back, and afterward we dropped again this year. Is that motionless the migration, or are our own gathering pages competing alongside all other?”
A few minutes later, they asked a second question:
“We market exactly the identical products in France and Germany, but those sites barely get any traffic. Do we need entirely distinct satisfied for all country?”
Neither of those would always appear in a keyword export. The closest you’d get is item akin “shopify immigration SEO” or “international SEO,” which tells you nearly nothing concerning the genuine situation. But if you appearance at those two questions as a satisfied brief, they’re incredibly specific: a platform, a timeline, two imaginable causes, and a decision to make.
This is exactly the benevolent of inquiry a buyer takes to an assistant, and it’s the benevolent of inquiry your satisfied have to be capable to answer.
How I Pull These Out At Scale
If you document your calls (I use Fireflies, but most call recorders create a copy you can export), I’d obtain the final three months of revenue and finding calls and run them through an AI aide alongside a immediate akin this:
“From these call transcripts, catalog all inquiry the potential asked or implied. Rewrite all one the way they would expression it to an AI assistant, and keep their constraints, specified as business size, budget, timeline, platform, and what they’ve already tried. Group akin questions together and note how many calls all collection came up in.”
The final education current is critical. If the identical circumstance comes up in five distinct calls, that’s the closest item to petition data you’ll get for a inquiry alongside no hunt volume, and it’s a lot additional persuasive in a gathering than “I think group are asking this.”
You desire to observe out alongside a brace of things here. Strip out names, business details, and item identifying before transcripts go into a third-party tool, and create certain your call record consent covers this benevolent of use. Also, revenue calls skew toward group who are already near to buying, so I’d add assistance tickets and pre-sale inquiries into the mix to grasp earlier-stage questions too.
Use Forums To Fill In The Rest Of The Situation
Your own calls lone inform you concerning group who already established you. Forums assistance alongside everyone else.
Reddit is the apparent one, but Quora, industry Slack and Discord communities, Facebook groups, and assessment sites specified as G2 or Trustpilot can all be fair as helpful depending on your market. The value normally lies in the beginning post, anywhere person explains their constraints in full, and in the replies, anywhere another group ask the follow-up questions an aide would additionally need answered before it could provision a decent response.
Sticking alongside the immigration example, a hunt akin site:reddit.com shopify immigration rankings dropped should bring up plentifulness of threads anywhere shop owners depict the identical issue in their own words, frequently alongside particulars my potential didn’t mention, specified as redirect issues or how lengthy they waited before panicking.
I copy the most applicable threads into a document and run the identical extraction immediate complete them. That way, forum questions and revenue questions end up in the identical format, and you can collection them together later.
See also: The Review Gap: Finding Client Opportunities In Competitor Feedback
See How The Assistants Break A Question Down
Once you have a set of genuine questions, the next stage is operating out what an AI aide would really hunt for whenever answering them.
Google is open concerning the fact that AI Overviews and AI Mode may use a query fan-out technique, issuing multiple connected searches to build a sole response. You can’t see those searches for a genuine AI Mode session, but you can get a sensible approximation through the Gemini API.
When you use Grounding alongside Google Search, the API reply includes a site called webSearchQueries, which lists the searches the example ran to answer your prompt. Google’s own records describes it as helpful for understanding the model’s reasoning process.
If you nourish in the archetypal immigration inquiry from earlier, you’d anticipate to see sub-queries covering things akin how lengthy rankings obtain to regain following a Shopify migration, how to inspect redirects following a re-platform, and how to place keyword cannibalization between gathering pages. Your own output volition change from run to run, but all sub-query is efficiently a division your leaf would need to shield to be helpful as a source.
You don’t need to be specialized for this. Go Fish Digital published a walkthrough that runs it through Screaming Frog alongside a copy-and-paste tradition JavaScript snippet and an API key from Google AI Studio. Their type sends all page’s H1, but swapping in your own catalog of questions is straightforward.
I desire to be apparent concerning what this output is, though. It shows you how a Gemini example expands a question, which is a fair proxy for Google’s AI features, but it isn’t a document of what AI Mode ran for an genuine user. I treat it as a safety checklist, and I wouldn’t current it to a client as item additional conclusive than that.
For ChatGPT, Perplexity, and the another assistants, the manual path motionless plant fine. Ask all one your question, note which sub-topics it decides to shield and which sources it cites, and difference that against the leaf that’s expected to answer it.
See also: The 2003 Framework That Was Already Doing Query Fan-Out
Check Search Console For What’s Already Reaching You
Search Console is improved as a validation stage than a finding tool for this, since it shows you which of these questions are already finding their way to your site.
Long conversational queries do rotate up in the Performance report. Suganthan Mohanadasan has already covered four ways to exterior them in a lot of detail, so I’d propose checking that out if you haven’t already. For a quick look, I use the Custom (regex) choice in the Query filter. This form returns queries of eight words or more:
^([^ ]+ ){7,}[^ ]+$ And this one returns queries that commencement alongside a inquiry word:
^(how|what|why|which|should|can|do|does|is|are|when|where|who)\s
Then I difference what comes rear against the questions from calls and forums. If the immigration circumstance shows up as a lengthy query, that’s fine confirmation that genuine group are bringing it to Google. If it doesn’t display up, it could be a safety gap, or the queries may merely be too rare for Google to display you.
That final item is important since Google anonymizes rare queries, and conversational queries are rare nearly by definition. Suganthan established that 57.7% of his impressions sat in the anonymized pond complete a 59-day period, so item you discover current is an undercount.
If you have admission to the generative AI achievement report, the pages perspective is value a appearance too. It shows which of your URLs are already appearing in AI Overviews and AI Mode. There’s no query data in it (yet), and sites alongside few AI impressions may not see the study at all, but it does inform you which pages Google is already comfortable using as a source.
Turn It All Into A Content Plan
By now, you’ll have a messy catalog of questions from multiple sources. To get it into a form a satisfied squad can really use, I put everything into one sheet. Here’s how a sole row power look, using the immigration example:
| Column | Example |
|---|---|
| Question | “We moved to Shopify, misplaced rankings, restored some, afterward dropped again. Is it the immigration or are our gathering pages competing?” |
| Source | Discovery call, affirmative multiple Reddit threads |
| Frequency | For example, 3 of the final 20 calls |
| Situation | Ecommerce, Shopify, migrated about two years ago, second autumn this year |
| Sub-questions | Recovery timelines, redirect checks, gathering leaf cannibalization |
| Existing page | Shopify SEO assistance page |
| Coverage | Partly answered (covers migrations, not cannibalization) |
| Priority | High |
For the precedence column, I use the identical proxies from my former column: endeavor value, evidence that the inquiry is being asked, how big the safety gap is, and whether you can answer it alongside genuine specifics. The frequence pillar feeds direct into the second one.
Once the panel is filled in, you’ll normally notice that most of the questions collection about a fistful of situations. Those clusters are your satisfied plan. In my experience, most of them can be handled by strengthening a page you already have, which is quicker than publishing item new and tends to execute improved too.
What This Won’t Tell You
I’m conscious this is a lot small tidy than sorting a spreadsheet by volume, so I think it’s value being up forefront concerning the limitations.
You motionless won’t have a quantity fig next to any of these questions, which method you’re prioritizing on judgement and frequency. Your own calls are biased toward group who already cognize you exist, so questions from audiences you don’t attain yet volition be underrepresented. The Gemini fan-out is an approximation of Google’s behavior, and Search Console lone shows you part of the image since of anonymization.
However, I don’t think any of that is a logic not to do it. It fair method you should current the output as a well-evidenced shortlist, particularly to stakeholders who are used to seeing a figure next to all mark or goal.
Over To You
If you desire location to commencement correct now:
- Export three months of call transcripts or assistance tickets, anonymize them, and run the extraction immediate above. Even 20 calls volition likely provision you additional usable questions than your final keyword export.
- Take your five most common scenarios and run them through Gemini alongside grounding switched on to see which sub-queries arrive back.
- Check those five against the leaf that should answer all one, and fix the gaps on your strongest existing leaf first.
Keyword tools are motionless helpful for understanding what group category into a hunt box. But a expanding portion of the questions that decide whether a buyer picks you are being asked in a conversation alongside an assistant, and the fine news is that plentifulness of them are already sitting in your inbox and call recordings.
More Resources:
- AI Answers About Your Locations Are Often Wrong – Check Before Customers Do
- The AI Conversations Leaking Into Your Search Console
- AI Visibility Measurement: What To Track & What To Ignore
Featured Image: kmzway/Shutterstock