It's not overhyped, but the way most people are measuring it makes it look that way.

On a pure traffic basis, AI tools currently send about 1.24% of traffic to the average Australian website. Ahrefs data shows Google sends 190 times more traffic than ChatGPT and all other LLMs combined. If traffic is your only lens, you'd be forgiven for thinking this is a sideshow.

But traffic is the tip of the iceberg. The more meaningful metric is influence — and we estimate AI is influencing somewhere between 20-30% of purchasing decisions, most of which never show up as a direct click in GA4.

Here's why: 68% of users start a research query in ChatGPT but then go to Google to verify and action it. Branded search in Google (people Googling a brand name after encountering it somewhere else) is up 15% across our client portfolio — but direct traffic is down. That gap points to people finding brands in AI environments and then Googling them. The AI shaped the consideration; Google got the credit.

Add to that: AI-referred traffic converts at 3x the rate of Google traffic, because users arrive already researched. One in three Australian consumers is now using AI tools on the path to a purchase. In B2B, that figure is 92%. By 2028, revenue flowing through AI search is projected to reach $750 billion globally.

So: not overhyped. Just poorly measured.

No, and we'd push back on any recommendation to shift significant budget away from Google right now.

Google reported 19% growth in search and other revenue in Q1 2026. Across our pool of 100+ Australian businesses, impressions (how many times brands appeared in Google search results) were up 43% year-over-year. Google is not shrinking.

What has changed is the relationship between ranking and traffic. Traffic is down despite impressions being up, primarily because AI Overviews now appear in almost 30% of searches (up from 10% a year ago), and when an AI Overview is present, click-through rate to position one drops by 58%.

The more accurate framing isn't Google vs. AI, it's that the same buyer journey now moves across both. 77% of consumers use AI tools and Google together. They start their research in one, verify in the other. Your presence in both environments is what matters.

SEO (Search Engine Optimisation) is optimising to rank in traditional search engines, primarily Google.

GEO (Generative Engine Optimisation) is optimising to be cited or referenced by AI tools like ChatGPT, Claude, Gemini, and Perplexity.

AEO (Answer Engine Optimisation) is sometimes used interchangeably with GEO, sometimes used more specifically to describe optimising for featured snippets and direct answer formats, including Google's AI Overviews.

The important thing to understand is that they're not the same discipline, but they're not opposites either. A strong SEO foundation is a prerequisite for effective GEO. If you have no SEO strategy, we'd recommend building that first and treating GEO as a growth layer on top.

Where they diverge: if you rank in Google's top 10, you have a 43% chance of appearing in Google AI Overviews, the overlap is significant. But for ChatGPT, that overlap shrinks to around 15%, and 80% of the brands ChatGPT cites don't rank in Google's top 10 at all. That gap is where GEO-specific strategy operates.

For most Australian businesses, ChatGPT is the priority. It holds around 68% of LLM market share (down from 85%, but still dominant) and drives 87% more referral traffic to websites than all other AI tools combined.

Claude grew 51x year-over-year and sits at 1.76% market share. Gemini is growing. Both matter, but the good news is that the tactics that improve your visibility in ChatGPT have significant overlap with what works in other tools. Optimising for ChatGPT gets you most of the way there across the ecosystem.

The main exception is Schema markup: ChatGPT doesn't currently use it, but almost every other AI tool does. Implementing Schema is still worth doing because of its value for Google AI Overviews, traditional SEO, and future-proofing for when ChatGPT likely adopts it.

The data points to B2B as the highest-adoption context: 93% of B2B buyers are using AI somewhere on the path to purchase, and 92% of purchasing influence in B2B now involves AI. Professional services, technology, legal, financial services, and recruitment are all sectors where decision-makers are actively researching in AI environments before they ever visit a website.

B2C adoption is lower but growing fast, 49% of all Australians and 78% of knowledge workers are using AI monthly.

For industries like emergency services, government, or categories where people are responding to urgent needs (rather than researching considered purchases), AI search is less of an immediate priority. The higher the consideration involved in a purchase or decision, the more AI is likely influencing it.

When someone types a query into ChatGPT, here's what happens under the hood:

First, ChatGPT checks whether its internal training data can answer the question without going to the web, about 69% of the time, it will use that internal knowledge. The other 31% of the time, it triggers a web search.

When it does search, it doesn't search for your one query. It performs what's called a query fan-out, breaking your single query into 8-10 related searches. "Accounting software for small businesses" becomes: "best accounting software for Australian small businesses," "Xero vs MYOB for small businesses," "Xero reviews 2026," and so on. You're effectively optimising for a cluster of searches, not a single keyword.

It then reads the sources it finds, often visits your website directly to verify that you do what you claim (and that you're available in Australia), and synthesises a response — typically citing 4–5 brands or sources.

The other key characteristic: ChatGPT is lazy. Whatever makes its job easier — clear structure, direct answers up front, fast-loading pages, significantly increases your chance of being cited. Whatever creates friction, like slow sites, JavaScript-heavy rendering, vague or hedging language, gets skipped.

It's a combination of two things: training data (which has a cutoff date and is updated periodically, not in real time) and live web search (which runs in real time when triggered).

For queries where ChatGPT uses its training data, the information reflects whatever was on the web at the time of its last training update. For queries where it triggers a web search (about 31% of the time), it's accessing live content.

This is why publishing date visibility matters, content that clearly shows a recent publication or update date is significantly more likely to be cited than undated or outdated content. Studies suggest recently updated content is up to 3x more likely to be cited.

It also means that for brands with negative information or outdated details circulating on the web (old directories, old press coverage), that information may persist in AI responses even after you've updated your own website.

Hallucinations are real and ongoing. ChatGPT will sometimes return incorrect information, especially when it can't access current or authoritative sources and present it with the same confidence as accurate information.

The best mitigation available to you is controlling the quality of what AI finds when it looks for your brand. That means:

• Your website content should be the most accurate, detailed, and up-to-date source of information about your business anywhere on the web
• Third-party profiles (Wikipedia, LinkedIn, directories, review platforms) should be claimed, accurate and current, AI doesn't just go to your website; it checks everywhere your brand is mentioned
• If your site is blocking AI crawlers, it will still find information about your brand, but from sources you don't control, which is where hallucinations tend to originate

If you discover that an AI tool is consistently returning incorrect information about your brand, the approach is to ensure that accurate sources are more authoritative and accessible than the incorrect ones. There is no direct "correct this" mechanism with most AI platforms currently.

In order of priority:

First: Fix your technical blockers. If AI crawlers can't access your site, nothing else matters. Check whether you're blocking AI bots via Cloudflare settings or your robots.txt file. Check your server logs to confirm that AI crawlers are actually reaching your pages. Fix slow page load speeds. ChatGPT will wait 3-5 seconds and then bounce to a competitor.

Second: Get your measurement in place. Build a keyword tracking list structured for how people actually search in AI (more on this in the measurement section). Set up a GA4 channel grouping for AI referral traffic. Know your baseline before you start optimising.

Third: Run a brand sentiment audit. Go to ChatGPT, Claude, and Gemini. Search for your brand, your category, and your competitors. Read what comes back. Understand the narrative that exists in these environments before you start trying to change it.

Fourth: Restructure your highest-value pages. Add TL;DR summaries above the fold. Front-load answers under H2s. This is fast to implement and has a measurable impact.

Fifth: Build out your content clusters. Identify the 8–10 related queries ChatGPT fans out to when someone searches your category. Make sure you have content that addresses each one.

The authority work. PR, reviews, citations, awards is the long game. Start it in parallel, but don't wait for it before doing the foundational work.

The core principle: make it as easy as possible for AI to extract and use your content.

Structurally, this means:

• H1 that directly answers the user's query: not a clever headline, a direct answer
• TL;DR summary at the top: a brief, clear answer before you get into the detail. 44% of LLM citations come from the first 30% of a page's text
• H2s written as questions: 78% of question-linked citations came directly from headings
• Answers placed immediately under the relevant H2: not buried in the paragraph below
• Clear publication/update date: recently updated content is far more likely to be cited

Content format matters too. Listicles, FAQ sections, pros-and-cons comparisons, and side-by-side competitor comparisons all perform strongly because they pre-structure the information for AI. Original data, statistics, and expert quotes increase citation likelihood because AI tools specifically look for authoritative sources to back up their responses.

Your meta description and title tag matter as well. AI tools read these first to assess whether your page is what it says it is. Including a direct answer in the meta description has been shown to increase citation likelihood.

The ideal page structure for AI search:

1. 1. URL and title tag: include the core topic/query directly
   2. H1: directly addresses what the user is searching for (not a creative headline)
   3. Meta description: includes a brief, direct answer to the query
   4. TL;DR / summary section: 2-4 sentences that answer the question before the full article begins
   5. H2s as questions: structured around the sub-questions someone would naturally ask
   6. Direct answers under each H2: don't make AI hunt through a paragraph to find the answer
   7. Supporting evidence: tables, stats, original data, expert quotes
   8. FAQ section: naturally captures the adjacent questions from the query fan-out

The general principle: AI is reading your page the way a very efficient researcher would, looking for clear signals that this page answers the query, extracting the clearest direct answer, then moving on. Structure your content accordingly.

More than most people currently appreciate.

Page speed: If your site takes more than 3-5 seconds to load, ChatGPT bounces. It's not a ranking signal in the way Google uses it, it's a hard threshold. Aim for an LCP under 2.5 seconds.

JavaScript: Sites that rely heavily on JavaScript to render content are difficult for AI crawlers to read. What the crawler sees may be a near-empty page rather than your actual content. If your site is JavaScript-heavy, this is a priority issue.

Schema markup: ChatGPT doesn't currently use Schema, but almost every other AI tool does, Gemini, Claude, Perplexity. It also remains critical for Google AI Overviews and for traditional SEO. Implement it now.

Crawlability: Set up Google Search Console and Bing Webmaster Tools as baseline diagnostics. When ChatGPT performs web searches, it uses search API infrastructure, if Google and Bing are having trouble indexing your site, ChatGPT is almost certainly having the same trouble.

Blocking: Many businesses unknowingly block AI crawlers, either via Cloudflare's default settings or through security rules designed to prevent scraping. Check your log files to see what's actually accessing your site.

Yes, with important adjustments.

Traditional keyword research is still valuable, but it needs to be supplemented with an understanding of how people search in AI environments. In Google, people use 3-4 word searches. In ChatGPT, searches run 6-12 words, are conversational, and are often comparative or use-case specific. "Running shoes" in Google becomes "best stability running shoes for flat feet under $200 in Australia" in ChatGPT.

This means your content needs to address more specific, conversational, and comparison-oriented queries, not just the short-tail terms you'd traditionally optimise for.

The foundations of SEO; technical health, quality content, domain authority, backlinks are still the bedrock of GEO performance. AI Overviews have a 43% overlap with Google's top 10. If you rank in Google, you have a meaningful head start for AI visibility. GEO builds on SEO; it doesn't replace it.

This is one of the most important and most overlooked aspects of AI search strategy.

When someone asks ChatGPT "X vs Y" or "what's the best [category]," that's one of the most common research queries going into AI tools. Your brand can influence these comparisons through:

Owning the comparison content: Create pages and content that directly address your brand versus competitors. If you've published a well-structured, direct, and accurate comparison, ChatGPT is more likely to draw from it than construct its own.

Managing your review presence: Studies show brands with 100+ reviews on platforms like G2, Capterra, or Trustpilot are 3.2x more likely to be cited than brands with fewer than 20 reviews. Review platforms are key sources for comparison queries.

Brand sentiment: Run a sentiment audit to understand what AI currently says when it compares your brand to competitors. If the response is consistently negative or if AI is recommending competitors over you, you need to understand why and what content or narrative changes could shift that over time.

Tracking comparison terms: Your AI search keyword list should include "[your brand] vs [competitor]" terms. This is where people are doing their research.

The ChatGPT-Shopify integration is significant for e-commerce, but it's not the whole picture.

ChatGPT has launched a product discovery feature that allows you to submit a merchant feed directly, similar to how Google Shopping works, so your products can be surfaced in AI search results. If you're not on Shopify, the merchant feed route is still available and worth pursuing regardless of platform.

More broadly, e-commerce is one of the highest-impact areas for AI search because of how consumers are using it: "What's the best laptop for video editing under $2,000?" is exactly the kind of query ChatGPT is being used for, and the answer drives purchase intent before a user ever opens a product page.

For e-commerce businesses, the priority areas are: submitting your merchant feed, ensuring product pages have clear structured data, generating genuine product reviews, and creating comparison and buyer-guide content that covers the queries people are researching in AI tools.

The core tension is real: regulated industries often can't make the definitive, concrete claims that AI tools favour, and legal review slows down the content creation and refresh cycles that AI search rewards.

The practical approach:

Focus on what you can own definitively. Even in regulated industries, there is usually factual, non-advisory content you can publish with authority, like process explanations, industry data, regulatory context, case study structures, "how to evaluate" guides. This is the content most likely to be cited.

Accuracy as a competitive advantage. In industries where AI hallucinations are highest-risk (financial advice, medical information), being the most accurate, sourced, and current voice in your category is a genuine differentiator. AI tools will gravitate toward sources they can verify against other authoritative references.

Control the brand narrative through owned and earned channels. Ensure your website, newsroom, and any published thought leadership is the most comprehensive and up-to-date information about your firm anywhere on the web. Then invest in being cited in authoritative trade publications, industry bodies, and recognised award programs — these third-party citations shape how AI perceives your brand.

Monitor what AI is actually saying about you. For regulated industries especially, a brand sentiment audit is not optional. If AI is returning incorrect or misleading information about your products or advice, you need to know and need a plan to address it through the sources AI is drawing from.

This is genuinely the hardest part of AI search right now and the most under-resourced area we see in Australian businesses.

The challenges are real: no native analytics from any AI platform, low click-through rates that make GA4 data thin, personalised results that vary across searches, and no keyword volume equivalent for AI queries.

What you should be doing:

• Manual search tracking: Build a keyword list structured for AI search (see Q18) and search each term in ChatGPT a minimum of 10 times per month. You need multiple searches per term because results are personalised and variable. If you use a tool, run 50–100 searches per term to get a statistically reliable average.
• GA4 channel grouping: Set up a dedicated channel group for AI referral traffic so you can isolate it from organic, direct, and other referral sources. Without this, AI traffic gets swallowed into referral or direct and is invisible. (Rocket has a blog post walking through the exact setup — search "Rocket Agency how to track AI search.")
• Google Search Console: Still the best proxy for overall search health and useful for mining longer-tail queries that can inform your AI keyword list.
• Brand sentiment auditing: Regularly query ChatGPT, Gemini, and Claude for your brand name, your category, and competitor comparisons. This is qualitative but critical; it's the only way to understand the narrative AI is building about your brand.
• Tools: Several enterprise platforms now offer AI visibility tracking, but many run $2,000-$5,000 per month just for the measurement layer. For smaller and mid-market businesses, a combination of manual tracking, GA4 configuration, and periodic brand audits is a credible and cost-effective alternative.

The mistake to avoid: using your Google keyword list in ChatGPT. People search very differently in AI tools (6-12 words, conversational, comparative) versus Google (3-4 words, short-tail). Tracking the same terms in both environments gives you misleading data.

Build your AI search tracking list across these categories:

• Core category terms (5-10 terms): "best [product/service] for [use case] in Australia"
• Branded terms (3-5): "[your brand] reviews," "is [your brand] good," "would you recommend [your brand]"
• Competitor comparisons (1-2 per competitor): "[your brand] vs [competitor]," "[competitor] alternatives"
• Category/vertical splits: If you serve multiple segments or product lines, track them separately
• Location-specific terms (if relevant): "best [service] in [suburb/city]"

Run each term at least 10 times per month manually. Track: are you being cited? In what position? How often? How does that compare to your closest competitors?

Because AI traffic in GA4 is thin and variable, the most meaningful metrics are:

• Citation rate: Out of your tracked searches, what percentage return your brand? Even appearing in 30-50% of searches for a core category term is meaningful visibility.
• Share of voice vs. competitors: For generic category terms, how often are you cited relative to your main competitors?
• GA4 AI referral sessions and conversions: Low volume but high quality — AI-referred traffic converts at 3x the rate of Google traffic
• Branded search growth in Google: A rising trend in branded Google searches (people Googling your name) can indicate growing AI-driven awareness, especially if it's not correlated with increased paid media
• "How did you find us?" surveys: Qualitative but high-signal; if customers are reporting finding you via ChatGPT or Gemini, that's the gold standard confirmation

For businesses making the case internally for GEO investment, the framing that tends to land is: AI is changing the research phase of the buyer journey, and citation rate is our visibility metric for that phase in the same way impressions are our visibility metric for Google.

Because AI traffic in GA4 is thin and variable, the most meaningful metrics are:

• Citation rate: Out of your tracked searches, what percentage return your brand? Even appearing in 30-50% of searches for a core category term is meaningful visibility.
• Share of voice vs. competitors: For generic category terms, how often are you cited relative to your main competitors?
• GA4 AI referral sessions and conversions: Low volume but high quality — AI-referred traffic converts at 3x the rate of Google traffic
• Branded search growth in Google: A rising trend in branded Google searches (people Googling your name) can indicate growing AI-driven awareness, especially if it's not correlated with increased paid media
• "How did you find us?" surveys: Qualitative but high-signal; if customers are reporting finding you via ChatGPT or Gemini, that's the gold standard confirmation

For businesses making the case internally for GEO investment, the framing that tends to land is: AI is changing the research phase of the buyer journey, and citation rate is our visibility metric for that phase in the same way impressions are our visibility metric for Google.

More important than most SEO practitioners expected and structurally different from how brand shows up in traditional SEO.

In Google, brand authority flows primarily through links. In AI search, it flows through sentiment; what AI tools actually say about your brand when asked. And AI forms that sentiment by synthesising everything written about your brand across the open web: your website, Wikipedia, LinkedIn, industry publications, review platforms, directories, social media, press coverage, forums.

We did a brand sentiment audit for Optus as a demonstration example. 72% of the time, ChatGPT returned a negative result for Optus, citing network reliability, poor customer service, and actively recommending competitors. If one in three Australians is now using AI to help choose a mobile provider, that sentiment profile has direct commercial consequences that are invisible if you're only monitoring Google rankings and click traffic.

The implication for marketers: brand monitoring in AI environments is not optional. It's a new category of always-on listening, understanding not just whether you're being cited, but what is being said.

Significant and probably the most underappreciated lever in GEO strategy right now.

AI tools weight authoritative publications heavily. In our Australian research, The Australian newspaper ranked as the third most cited brand in ChatGPT results, after Wikipedia and Reddit. If your brand is being quoted, referenced, or featured in major publications, that coverage is actively shaping your AI brand profile.

This means traditional PR, getting your brand and your experts into authoritative media, has a direct GEO benefit beyond the traditional brand awareness value. Digital PR specifically (the kind that generates authoritative links and citations in major outlets) is the GEO equivalent of link-building.

The practical implication: if you're doing PR, brief your agency or team on the GEO angle. Target publications that are likely to be in AI training data and frequently cited. Getting listed in a "best [category]" article in a major publication is far more valuable in an AI search world than it was in traditional SEO.

Reviews are a meaningful citation and authority signal for AI tools, particularly on platforms like Google, Trustpilot, G2, and Capterra.

The data point: brands with more than 100 reviews on G2, Capterra, or Trustpilot are 3.2 times more likely to be cited than brands with fewer than 20 reviews. These platforms appear as explicit citation sources in AI responses.

The practical work here is straightforward but requires consistency: ask for reviews at the natural close of a purchase or project. For B2B businesses, a direct request to satisfied clients after delivery is usually the most effective approach. For B2C, post-purchase email sequences or in-app prompts work well.

This is also worth monitoring. If ChatGPT is citing your review profile in a negative context,  drawing on a period when reviews were lower, or citing a competitor's review count favourably. That's a signal to prioritise review generation.

This is a genuine constraint for manufacturers and intermediary businesses, but it's not a blocker.

Your goal in AI search isn't necessarily to capture the transactional query, it may be to shape the consideration phase. If someone asks ChatGPT "what's the best [product] for [use case]," you want your brand name to be in the response as the recommended manufacturer, even if the click-to-purchase goes to a distributor.

The tactics that help here:

• Category and product education content on your own site: definitive guides, comparison content, specification pages, use-case guides. This content positions you as the authoritative source for your product category.
• Industry and trade publication presence: getting your brand and products covered in specialist media that AI draws on for category recommendations
• Review and award presence: product reviews in specialist publications, industry award wins, certification bodies. These become the third-party endorsements AI cites when recommending products.
• Distributor content alignment: where you can influence distributor content (product descriptions, FAQ pages, etc.), ensure it's structured in an AI-friendly way and accurately represents your product

The pressure on traditional paid search is real but gradual. AI Overviews currently appear in around 30% of searches, and only 5-15% of searches in most industries are commercial or transactional, the type most likely to trigger ad competition. The immediate impact is greater on informational organic traffic than on paid.

The longer-term dynamic to watch: as AI Overviews expand and zero-click behaviour increases for more query types, the pool of searches that result in a click (and therefore an ad impression) will continue to shrink. That will likely increase CPCs on commercial queries as competition concentrates.

Google has also signalled they are experimenting with ads within AI Overviews and AI Mode. The shape of paid search in an AI-dominant environment is still forming.

The near-term implication for paid search practitioners: track whether AI Overviews are appearing for your highest-value commercial terms, monitor click-through rate trends carefully, and be ready to adapt bidding strategies as the volume/CPC dynamics shift.

A few things we're confident about based on where the data and the platforms are moving:

The research phase of the buyer journey is increasingly happening in AI. The shift isn't from Google to AI, it's from Google-as-research to AI-as-research, then Google or direct for the decision. Brands that are visible and well-regarded in AI environments will win the consideration phase before buyers even get to search.

Brand sentiment in AI will become as important as SEO rankings. The brands that invest now in understanding and actively managing their AI narrative — through content, PR, reviews, and direct channel ownership, will have a structural advantage that compounds over time.

Measurement will improve. The current lack of native AI analytics is a temporary constraint. As platforms mature and add attribution capabilities, the measurement story will get clearer. Building the habits and infrastructure now means you'll have historical data to work with when those tools arrive.

Agentic AI will raise the stakes for technical readiness. The shift toward AI agents that autonomously complete tasks (booking flights, purchasing products, filling forms) means your website's technical accessibility to AI crawlers and agents will become increasingly business-critical. Google published a white paper on agent-friendly website design earlier in 2026. This is the direction of travel.

The practical advice: don't wait for certainty. The businesses that are building visibility in AI environments now, through the tactics in this piece, are building the same kind of compounding advantage that early SEO adopters built in the 2010s. The window to establish that advantage ahead of competitors is right now.