Insights

How we use SEO, Google Ads and AI rankings together (and why looking at them separately wastes money)

Most businesses we talk to are running three dashboards.

One for Google rankings, one for ad spend, one for "we're showing up in ChatGPT now." Three tabs, three stories, three vendors telling them three different things to do next.

The problem isn't any single dashboard. It's that the answers only show up when you look at all of them on one screen.

A roofing business we work with shows up in AI answers for every prompt that matters. Hundreds of five-star reviews. The AI visibility chart climbs every week. The traffic chart stayed flat. Until you put those two lines next to each other you'd think you were winning.

That's the gap this guide is about. Here's how we look at SEO, Google Ads and AI rankings as one picture instead of three.

The four-quadrant frame

We put every account we run into the same four boxes.

1. AI Rankings - ChatGPT, Gemini, Perplexity, Copilot, Claude, Google AI Overviews

2. Google Rankings - keyword positions, Search Console impressions, AIO appearances

3. Traffic + Behaviour - GA4 sessions, conversions, source/medium, page performance

4. Paid - Google Ads spend, search terms report, cost per conversion

Quadrants 1 and 2 are visibility. Are you findable.

Quadrants 3 and 4 are what happens next. Does that visibility convert into anyone clicking, and what does it cost when you pay for the click directly.

The mistake we see weekly is treating any one of these as the whole answer. SEO agencies show you Quadrant 2 and stop. AI visibility tools show you Quadrant 1 and stop. Your ads guy shows you Quadrant 4. Nobody is stacking them on one screen and asking "what's the next move."

Quadrant 1: AI rankings

The first thing people want from an AI visibility tool is to know whether their business shows up when someone types "best [thing] in [city]" into ChatGPT.

Fair question. The portal answers it directly. We track prompts the business actually cares about, run them across six engines (ChatGPT, Gemini, Perplexity, Copilot, Claude, Google AI Overviews), and tell you who got mentioned and what the engine cited as its source.

Here's the part most tools skip: the citation list.

When an AI engine mentions your business, it's almost always pulling from somewhere. A review site, a directory, a blog post, your own homepage, a third-party comparison page. The engine doesn't know you. It knows the sources that talk about you.

Which means the work to win AI rankings is not new work. It's mostly the same SEO foundations - structured pages, schema markup, presence in the directories and review sites the engines trust, clean content that says clearly what you do. The engines crawl the same web everyone else does.

So here's what we look at in this quadrant:

• Mention rate - across all the prompts we track, how often does this business get named at all? A new client usually starts at 10-30%. We aim to get them to 60%+ over three months.
• Engine coverage - is the business getting picked up by all six engines or just one or two? If only one, we look at what's different about that engine's source mix and try to replicate.
• Citation sources - which third-party pages are the engines actually reading. Often the same five or six URLs come up repeatedly. Those URLs become the priority for SEO and outreach work.
• Sentiment / framing - when mentioned, is the business framed as a leader, an option, or an alternative-to? This is the one most tools don't surface and the one that matters most for trust.

And here's the trap.

Rankings on AI engines do not directly translate into traffic the way Google rankings do. When ChatGPT mentions your business in a response, the user doesn't always click through. They might just read the answer and call you. Or read the answer and forget. Or read the answer and Google your name to verify - in which case the click shows up as branded organic, not as an AI referral.

So a chart showing "you're mentioned in 80% of prompts" is half the story. The other half is in Quadrant 3 - did anyone actually arrive on the site, did the phone ring, did the form get filled.

Quadrant 2: Google rankings

Google rankings are the oldest quadrant. Everyone's looked at one at some point. We still look at it daily.

Three things matter here, and most "SEO reports" only cover one.

Position by keyword. The obvious one. We track 20-50 keywords per client that map to actual buying intent, not vanity. "Plumber Sydney" is a buying-intent keyword. "What is plumbing" is not, and chasing it is a waste. We pull positions daily so the chart shows movement, not a snapshot.

Impressions in Search Console. Position is what your page ranks. Impressions is how often Google actually showed it. The two are not the same chart. A page can rank position 4 for fifty keywords and barely get impressions because nobody searches them. We pair the rank tracker with Search Console impressions to know which rankings are worth defending and which to drop.

AI Overview appearances. Google AIO is now showing on a big chunk of search results. When it appears, the blue-link click rate underneath collapses. The portal flags every keyword where AIO is present and tells you whether your page is cited inside that AI Overview. If you rank position 2 for a keyword but Google's AI Overview answers the question above you with citations to three competitors, you've lost that page even though your rank looks healthy.

That last point is where Quadrant 1 and Quadrant 2 stop being separate things. Google AI Overviews pull from the same citation graph as ChatGPT. The pages that get cited in AIO are the same pages that get cited by ChatGPT and Perplexity. Structured content, clean schema, third-party validation, on-page clarity. The work is the same.

We tell every client the same thing in their first month: the cheapest way to win AI rankings is to do basic SEO properly. Schema markup, FAQ blocks, clean H1/H2 structure, internal linking, fast page load, fixed canonicals. The businesses that already did this work five years ago for Google are the ones showing up in ChatGPT now. No magic AI tooling required.

What we look for when a client's Google rankings are stagnant:

• Pages targeting keywords with no buyer intent (drop them)
• Pages competing with themselves on the same keyword (merge or pick one)
• Pages missing schema markup that competitors have (add it)
• A flat homepage with no supporting topic pages (build the topic cluster)

None of that is exotic. All of it shows up in the portal as one view per client.

Quadrant 3: Traffic + behaviour

Visibility is half the story. Whether anyone arrives, and what they do when they get there, is the other half.

This is where GA4 sits.

We hook GA4 into every client's portal so the same dashboard that shows AI mentions and Google rankings also shows whether traffic is moving. Three views matter:

Sessions by source/medium. Where the traffic is coming from. Organic search, direct, paid, referral, AI referrals where we can identify them. When AI rankings move and direct traffic moves with them, you can connect the dots. When AI rankings move and nothing else moves, you've got a different problem - the engine is recommending you but the user isn't converting curiosity into a click or a call.

Conversions and conversion rate. Form fills, phone clicks, checkout completes, whatever the business actually measures success by. The number we care about is not visits. It's whether the visits did anything useful.

Page-level performance. Which pages bring traffic and which pages convert it. Often not the same page. A blog post might bring 60% of organic traffic and 5% of conversions. The contact page might bring 2% of traffic and 40% of conversions. That tells us what to feed, what to fix, and what to leave alone.

Here's where the cross-quadrant view earns its keep.

A professional services client we work with had three things running at once: an SEO programme, a Google Ads account, and a sudden uptick in AI engine mentions. Each conversation with each vendor felt like a small win. But when we put all three onto one screen and stacked GA4 on top, the answer was clear: ad spend was driving 80% of the contact form fills, organic was driving brand searches but not converting them on the website, and AI mentions were generating curiosity but no traceable traffic at all.

That single chart shifted the conversation. The question stopped being "how do we improve SEO" and became "where's the next dollar best spent." Different question, different answer.

The trap in Quadrant 3 is treating GA4 as a standalone report. It's not. It's the truth check on every other quadrant. Without it you're guessing whether visibility is converting into business.

Quadrant 4: Paid

Paid is the only quadrant you can change tomorrow.

Organic SEO takes weeks. AI rankings move month over month. GA4 tells you what already happened. Google Ads is the lever you can pull on Monday morning and see a result by Friday.

We use the paid quadrant in two ways.

As a lead generation channel in its own right. Obvious use case. Set up campaigns, structure them properly, watch cost per conversion, scale what works, kill what doesn't. We do this for every client running Ads.

As a test bed for organic. This is the use case most agencies miss.

The Google Ads Search Terms report shows you the exact phrases real customers typed into Google to land on your ad. Not the keywords you targeted - the keywords that triggered your ad. The list is gold. It tells you:

• What language your buyers actually use (vs what you think they use)
• Which long-tail variations have intent (and could become organic content)
• Which adjacent searches are wasting your budget (and need to be negative-keyworded)

We pull this report monthly for every client. The high-converting search terms from paid become the targeting list for new SEO pages. The low-converting search terms get added to the negative keyword list so the budget stops bleeding.

The waste finder is the part most clients have never seen. Open the Search Terms report, sort by spend, look for the rows where money went in and nothing came out. An e-commerce client we run had a Shopping campaign quietly burning a chunk of monthly budget on search terms that were sending shoppers to product pages with conflicting shipping rules. Three negative keywords and a Merchant Centre fix recovered the spend the following week. No ad copy change. No bidding strategy change. Just a look at what Google was actually showing the ad against.

We also use Quadrant 4 to test new positioning before it ships organically. Want to know if a new tagline converts? Run it in an ad for two weeks. The click-through rate tells you whether the angle lands long before you commit to writing a homepage rewrite around it.

The deeper point: Quadrant 4 is the only quadrant with a real-time feedback loop. The other three teach you slowly. Paid teaches you fast. We use it as the front-of-house experiment lab for everything we eventually plan to do in organic.

One real client view

Here's what this looks like on a real account. Anonymised, but the shape is real.

E-commerce business. Sells physical products. Mid five-figure monthly Ads spend during peak season. Organic SEO running in parallel. Started showing up in AI engine recommendations around month three of working together.

Before we put it all on one screen, the monthly conversations went like this. SEO chart: rankings up, two new featured snippets. Ads chart: cost per conversion holding steady, return on ad spend healthy. AI visibility chart: mentions climbing across ChatGPT and Perplexity. Three good news stories. Three separate conversations.

Once it was stacked into one view, the picture changed.

The AI mentions were almost all citing a third-party review site, not the client's own pages. The Google rankings were strong on category pages but weak on individual product pages - exactly the pages the Ads campaigns were paying to land traffic on. GA4 showed that the highest converting traffic was coming through branded search, not category search, suggesting customers were finding the brand somewhere else first and Googling the name to verify.

That somewhere-else was the AI mentions and the third-party review site they kept citing.

What we did with that picture:

• Doubled down on the review site (asked customers for reviews, replied to all existing ones, made sure listing data was current)
• Shifted ad budget from category keywords to product-name keywords where intent was higher
• Wrote individual product page content with the language pulled from the Ads Search Terms report
• Built schema markup on product pages so AI engines could cite the brand directly instead of only the third-party site

Six weeks later: AI mentions cited the brand's own pages 40% of the time (up from 5%), cost per conversion dropped, organic traffic to product pages tripled.

None of that happens if you only look at one quadrant.

What we've stopped doing

Three habits we dropped when we started looking at all four together.

Reporting position rankings as the headline. Position 1 for a keyword nobody searches is worth less than position 8 for a keyword that converts. We stopped leading reports with rank movement and started leading with conversions and revenue.

Building content for keyword volume alone. Keyword research tools love volume. Volume without buyer intent produces traffic with no business value. We now cross-check every target keyword against either the Ads Search Terms report (does it convert when paid?) or a manual SERP review (is the existing top result a transactional page or an informational one?). If the intent doesn't match what the client sells, we don't write the page.

Treating AI rankings as a separate workstream. We've stopped selling "AI visibility" as a standalone product. The work that wins AI rankings is mostly SEO done properly plus presence on the right third-party sites. We do it inside the same monthly programme as everything else. Different chart, same work.

The agency layer

The portal exists so any business can run this stack themselves. Track AI rankings, watch Google positions, see GA4 traffic, monitor Ads performance. One screen. AU$50 a month. No contract.

For the businesses who don't want to run it themselves, our marketing team takes the same view and runs it for them. Same dashboard, same four quadrants, plus the hands on the levers. That arm of the business handles SEO programmes, Google Ads management, AI visibility work, and the connective tissue between them. Same logic as everything above - we just do the work.

Either way the picture is the same. The portal is the engine. The agency layer is optional.

Final word

If you're running three dashboards and getting three different stories, the cheapest move is to put them on one screen and look at what they say together. The portal does that for AU$50 a month. If you'd rather hand the work over, the marketing arm is the other lane.