AI search is no longer a future consideration for Australian businesses. It is a present-tense distribution channel. Gartner predicted in February 2024 that traditional search engine volume would drop 25 per cent by 2026 as users shifted to AI chatbots and virtual agents. Whether or not the full 25 per cent has materialised, the behavioural shift is real. AI Overviews now reach over 2 billion monthly users across 200+ countries. AI Mode crossed 1 billion monthly users by May 2026. And generative AI use for local business recommendations jumped from 6 per cent to 45 per cent in a single year (industry survey data, 2026).
We wanted to know what actually happens when Australian buyers type their questions into ChatGPT, Perplexity, Gemini and Google’s AI surfaces. Not what the theory says. What the answers look like. So we ran the queries ourselves, across the industries we serve at 21 Webs, and documented the patterns.
This article is the write-up. It covers the six findings that matter most, and what each one means for your SEO and AI visibility strategy.
Before we go further, some context on scale. Google’s own data shows AI Overviews now appear on roughly 50 per cent of US search queries (blog.google, I/O 2025 and 2026 announcements). In specific verticals like healthcare, that figure exceeds 80 per cent. AI Mode, Google’s conversational search layer, has crossed 1 billion monthly users. And ChatGPT itself processes approximately 1.6 billion daily queries. These are not niche platforms. They are mainstream discovery channels. And the rules for appearing in them are different from the rules for ranking on Google.
Here is what we found.
Finding 1: The Local Block Comes First, and It Runs on Reviews
When you ask ChatGPT a local business query (for example, “best plumber in Melbourne’s western suburbs” or “NDIS provider near Werribee”), the response almost always opens with a local business block before the text answer. This block looks like a ranked list: business name, location, and a brief reason for inclusion.
The pattern we observed is consistent with what independent research has confirmed:
Review count is the entry threshold. Research published in 2026 found that businesses below roughly 150 reviews per location rarely get named in AI recommendations at all. ChatGPT references reviews in approximately 58 per cent of its local responses, while Perplexity uses reviews in 100 per cent (industry analysis of 800,000 AI-generated answers, 2025-2026).
Review quality and specificity matter as much as volume. A review that says “great service, five stars” gives the AI nothing to work with. A review that describes what was done, how fast and what the experience felt like gives the AI the language it needs to recommend you for specific queries.
NAP consistency is the trust check. ChatGPT cross-references business information across dozens of sources before including a recommendation. Name, address and phone number need to match character for character across your website, Google Business Profile, Bing Places, Apple Maps and every directory you appear in. Inconsistency reads as unreliability and quietly drops you from consideration.
Review velocity outweighs total count over time. A steady flow of 2 to 4 reviews per week over 90 days outperforms a burst of 50 reviews followed by months of silence. The AI surfaces businesses that look consistently active, not ones that ran a one-off review campaign.
Across the four engines we test, the local block is the constant. ChatGPT leads with it. Perplexity leads with it. Gemini and Google AI Overviews lead with it. The ordering changes and the reasons given change, but the block always comes before the prose. That consistency is the finding. The AI answer to a local question is built on the same raw material everywhere.
The implication is clear. If your small business SEO strategy does not include a structured review generation programme and rigorous NAP management, you are invisible to the fastest-growing discovery channel in Australia. Reviews are no longer a reputation afterthought. They are the first filter an AI applies before it decides to name you at all.
Finding 2: Directories and Awards Dominate the Text Answer
After the local block, ChatGPT’s text answer for agency and service-provider queries consistently draws from two types of sources: curated directory listings (particularly Clutch) and industry awards lists (particularly the APAC Search Awards and similar regional programmes).
This is not a secret. ChatGPT itself, when prompted, will flag that these lists are influenced by sponsorships, self-submitted data and pay-to-play dynamics. But it cites them anyway, because they are the most structured, entity-rich sources available for the query. The AI is not endorsing the integrity of the list. It is using the list because it is machine-readable and consistent.
What this means for you:
Directory presence is a citation lever, not just a referral channel. Even if you have never received a lead from Clutch, your presence there (with complete profile data, verified reviews and accurate service categories) increases the probability that AI surfaces you in a recommendation.
Awards and recognition lists function as entity validators. When ChatGPT sees your business named in a structured awards list alongside other verified entities, it treats that as a corroborating signal. It does not make you the recommendation. But it puts you in the pool from which recommendations are drawn.
Completeness beats ranking on these platforms. A fully completed directory profile with 30 verified reviews and consistent data will outperform a top-ranked but incomplete profile in AI citation terms. The AI cannot extract much from a shell of a profile, no matter where that profile sits in the directory’s own ordering.
If you are a service-based business in Australia and you have not claimed, completed and maintained your profiles on the major directories (including industry-specific ones for your vertical), you are leaving AI visibility on the table. We wrote separately about what AI search actually means for SEO in 2026 and how these citation mechanics work in practice.
Finding 3: Reddit Is a Dominant Citation Source for AI Recommendations
This one surprised us initially, but the data is consistent across every major study published in 2025 and 2026.
Reddit is one of the most-cited source domains in LLM responses globally. An independent analysis of 30 million sources ranked Reddit as the most-cited source across ChatGPT, Google AI Mode, Gemini, Perplexity and AI Overviews combined (industry research, April 2026). A separate study found that domains with significant brand mentions on Reddit averaged 3.9 times more ChatGPT citations than domains with minimal Reddit presence (industry analysis of 129,000 domains, 2025-2026).
For Australian agency and service queries specifically, Reddit threads like “best SEO agency in Melbourne” or “who do you use for web design in Australia” carry real weight in LLM citation. In our own tracking, Reddit was the most frequently cited third-party source domain in AI responses to Australian agency queries, appearing in a substantial share of the mention data we collected. When we pulled the sources behind a set of AI answers for Australian agency terms, Reddit threads kept surfacing where brand-owned content did not.
What this means practically:
You need genuine Reddit presence. Not promotional posts (moderators and AI ranking signals filter those within hours). Real participation in relevant subreddits. Answering questions. Sharing specific experience. Contributing frameworks.
Third-party mentions on Reddit are a citation lever you can influence but cannot buy. The signal is authenticity and discussion depth, not link authority. A Reddit thread where someone recommends you because a job went well is worth more to an LLM than a directory listing you paid for.
Monitoring Reddit mentions of your brand and your competitors is now an AI visibility task. If your competitor is being discussed and recommended in relevant threads and you are not, that gap shows up directly in AI recommendations.
Our own approach is simple: our team answers questions where our clients’ customers actually ask them. That means contributing genuinely useful answers in the subreddits that match the industries we serve, with real specifics about how a build was structured or why a campaign moved. We do not treat Reddit as a link farm. We treat it as a place where a useful answer today becomes a citation tomorrow.
Finding 4: Traditional Ranking Position and AI Citation Are Weakly Coupled
This is perhaps the most important finding for any business that has been told “just rank on Google and the AI will follow.”
Research published in early 2026 found that only 17 per cent of AI Overview citations come from pages already ranking in the organic top 10 (industry tracking data, February-March 2026). That is a dramatic drop from 76 per cent in mid-2024. The correlation between Google ranking position and AI citation has weakened substantially as AI systems have developed their own citation logic.
We see this in our own tracking across client campaigns. A page ranking position 3 for a target keyword is not guaranteed to appear in the AI answer for that same query. And pages that do not rank in the top 10 can and do appear in AI citations, if they carry the right signals: entity consistency, third-party mentions, structured data and review corroboration.
This does not mean traditional SEO does not matter. It means SEO alone is no longer sufficient. The signals that drive AI citation overlap with, but are distinct from, the signals that drive Google ranking. An SEO consultant who is only optimising for traditional ranking is optimising for one of two parallel discovery systems, and the second one is growing faster.
To put this in context: Google AI Overviews now appear on approximately 48 to 50 per cent of all US search queries, an almost 8x expansion from January 2025 (industry tracking data, February-March 2026). Google itself has said AI Overviews reach “roughly 50 per cent of US queries.” In Australia, the expansion has followed a similar trajectory, particularly in high-intent verticals. That means roughly half the time a potential customer searches for a service you offer, they encounter an AI-generated answer before they see a single organic result. If your page is ranking but not being cited in that AI answer, you are visible in one system and invisible in the other.
The practical response is straightforward. Continue building organic ranking strength. But layer the llm seo signals on top: entity consistency, structured data, review depth and third-party mentions across platforms the AI trusts. This is the discipline of optimising for how large language models retrieve, cite and recommend, and it now runs alongside traditional ranking as a separate workstream. The businesses winning in both systems are the ones treating traditional SEO and generative engine optimisation as complementary, not interchangeable.
Finding 5: AI Visibility Is Not Evenly Distributed Across Industries
Not every Australian industry is experiencing AI search disruption at the same rate. The data shows significant variation:
E-commerce accounts for a large share of Australian SEO spend. Industry estimates suggest e-commerce represents roughly 35 per cent of total search marketing investment in Australia. AI Overviews and product-related AI answers have expanded rapidly into shopping queries, with AI Overviews appearing on 14 per cent of shopping queries by March 2026, a 5.6x increase from November 2024 (industry tracking data, 2026). The shopping answer is being rebuilt around comparison, and comparison needs structured product data.
Healthcare AI visibility is growing fast. Healthcare-related queries now trigger AI Overviews at the highest rate of any industry tracked, at approximately 88 per cent (industry data, February 2026). Healthcare marketing budgets in Australia have grown approximately 18 per cent year on year, partly in response to the need for AI-optimised content in a sector where trust signals and clinical accuracy are critical.
Trades and local services face the “local block” challenge directly. For plumbing, electrical, HVAC and similar queries, the AI local block is the entire answer. If you are not in that block, the user never sees a text recommendation. This makes review count, review quality and NAP consistency even more critical for trades than for other sectors. We built our industry-specific web design and SEO practice around these vertical differences.
The takeaway: your AI visibility strategy needs to be calibrated to your industry, not copied from a generic playbook. A national retailer and a suburban electrician are not playing the same game, even though both are being recommended by the same engines. The source mix, the entry threshold and the winning signals differ by vertical.
In our monthly audits, we calibrate the query set to the vertical before we run anything. A dental clinic gets a different query list and a different signal audit from a roofing supply business. The engines do not treat them the same way, so we do not either. If you copy a generic AI visibility playbook across two different industries, you are optimising for a market that does not exist.
Finding 6: What You Need to Build vs What You Have Been Told Matters
Based on our analysis, here is what actually drives AI citation for Australian businesses, compared with what many agencies have historically prioritised:
What drives AI citation:
Review velocity and review specificity (detailed, recent, steady)
Entity consistency across every data source (NAP, schema, directory profiles)
Third-party mentions on platforms AI trusts (Reddit, YouTube, industry forums)
Directory presence with complete, verified profiles
Structured data (LocalBusiness schema, FAQPage schema, service-area markup)
Content that answers specific questions with verifiable, experience-backed detail
What has been historically prioritised but is insufficient on its own:
Keyword ranking position (weakly coupled to AI citation as of 2026)
Domain authority score (LLMs use a different trust hierarchy)
Backlink volume without accompanying entity signals
Vanity keyword reports that do not track AI surfaces
This does not mean you should stop doing SEO. It means you should demand that your SEO strategy includes the AI citation layer. If your agency cannot tell you where your business appears (or does not appear) across ChatGPT, Perplexity, Gemini and Google AI Overviews, they are optimising for half the discovery landscape. For a full breakdown of what AI search means and what it does not, read our analysis on what AI search actually means for SEO in 2026.
How We Measure This: The Monthly AI Query Audit
At 21 Webs, we run a structured monthly query audit across four AI surfaces: ChatGPT, Perplexity, Gemini and Google AI Overviews. Here is the method:
We identify the 20 to 30 queries most likely to drive commercial intent in each client’s industry. These are not vanity keywords. They are the questions a buyer types before making a purchase or booking decision.
We run each query across all four AI surfaces and document who gets cited, how they get cited and what source signals the AI references. This includes the local block, the text answer, the linked sources and any disclaimers.
We compare the results against the client’s current citation profile. Where are they appearing? Where are they absent? Which competitors are being cited, and what signals are those competitors carrying that the client is not?
We map the gap to a prioritised action plan. Review velocity. Directory completions. Reddit and forum presence. Schema deployment. Content updates.
The output lands in the client’s monthly report as a citation scorecard: how many target queries now surface the business, where it is still absent, and which three actions close the widest gaps first. The report does not stop at “your citations are up.” It names the specific queries, the specific missing signals and the specific next move.
This is not a one-off exercise. AI answers change. Citation patterns shift. Reddit’s share of LLM citations grew 73 per cent in commercial categories between October 2025 and January 2026 alone (industry report, Q1 2026). YouTube overtook Reddit as the most-cited social platform in AI responses by early 2026, with 16 per cent of LLM answers containing YouTube-sourced information compared to 10 per cent for Reddit (industry data, January 2026). A visibility snapshot from three months ago is already stale.
The measurement framework we use tracks four dimensions monthly:
Citation presence. For each target query, does your business appear in the AI answer? In the local block? In the text response? In the source links?
Competitor citation. Who else appears for your target queries, and what signals are they carrying that you are not? This competitive layer is critical. AI answers are zero-sum in a way Google results are not: ChatGPT recommends only about 1.2 per cent of local business locations (industry analysis of 350,000+ locations, 2026). If your competitor is in that 1.2 per cent and you are not, you have lost the interaction entirely.
Signal gap analysis. Which specific citation signals (reviews, directories, schema, mentions, content depth) are you missing relative to the businesses that are being cited? This is where the action plan comes from.
Trend tracking. Are your citation appearances increasing, stable or declining over time? AI visibility is not set-and-forget. It requires the same ongoing attention as traditional ranking, because the underlying models and citation patterns evolve continuously.
If you want to understand how AI SEO works in practice and how it differs from traditional optimisation, we have published a detailed explainer on what AI SEO is and how it differs from traditional SEO.
What This Means for Your Business Right Now
If you have read this far, here is the practical summary:
Check who AI recommends in your industry today. Open ChatGPT, Perplexity and Google. Type the query your best customer would type. Look at who appears. If it is not you, find out why. That five-minute test tells you more about your AI visibility than any ranking report.
Build the signals AI actually uses. Review velocity. NAP consistency. Directory completions. Schema markup. Third-party mentions on platforms like Reddit and YouTube. This is the new baseline.
Stop treating Google ranking as the whole picture. It is one of two parallel discovery systems. The second one is growing faster, and it uses different signals.
Demand AI visibility reporting from your agency. If your monthly report does not track who AI recommends for your target queries, it is missing the channel where buyer discovery is shifting fastest.
We have built our AI SEO services around exactly this approach. We track AI citations. We build the signals that drive them. And we report on the results monthly, tied to the queries that actually matter for your business.
For more context on how AI traffic converts differently and what that means for measurement, read our analysis on how AI traffic converts differently in 2026.
Frequently Asked Questions
Why does ChatGPT recommend my competitor?
ChatGPT cites businesses with the strongest combination of review volume, review specificity, NAP consistency, directory presence and third-party mentions. If your competitor carries stronger signals across these five areas, the AI will surface them first, regardless of Google ranking position.
Does ranking on Google get me cited by AI?
Not necessarily. Research shows that only around 17 per cent of AI Overview citations come from pages in the organic top 10. Google ranking helps, but it is not sufficient on its own. AI citation requires entity consistency, structured data and third-party corroboration that go beyond traditional ranking factors.
What makes ChatGPT recommend a business?
ChatGPT retrieves live web pages at the moment you ask, then synthesises an answer from what it finds. The businesses that get recommended are the ones whose information is easy to retrieve, consistent across every source and corroborated by independent third parties, including reviews, directory listings and community mentions.
How do I check who AI recommends in my industry?
Run your 10 most important buyer queries across ChatGPT, Perplexity, Gemini and Google. Document who appears in each answer. Compare their citation signals (reviews, directories, schema, mentions) against yours. The gap is your AI visibility action plan.
The Bottom Line
Google ranking is one of two parallel discovery systems. The second one is growing faster, and it uses different signals.
The AI answer to a local question is built on the same raw material everywhere: review count, review quality, NAP consistency, directory presence and third-party mentions. Reviews are no longer a reputation afterthought. They are the first filter an AI applies before it decides to name you at all.
This does not mean traditional SEO does not matter. It means SEO alone is no longer sufficient. The businesses winning in both systems are the ones treating traditional SEO and generative engine optimisation as complementary, not interchangeable.
Want to see who AI recommends for your industry?
We track AI citations across ChatGPT, Perplexity, Gemini and Google AI Overviews, tied to the queries that actually matter for your business. Talk to our team or explore our AI SEO services.
Sources
[1] Gartner, “Gartner Predicts Search Engine Volume Will Drop 25% by 2026 Due to AI Chatbots and Other Virtual Agents” (February 2024). https://www.gartner.com/en/newsroom/press-releases/2024-02-19-gartner-predicts-search-engine-volume-will-drop-25-percent-by-2026-due-to-ai-chatbots-and-other-virtual-agents
[2] blog.google (Google I/O 2025 and 2026 announcements). https://blog.google/
