Cameron Bryant.
Cameron Bryant, partner, Brainlabs in Australia, co-founder, Sparro.
If you don’t know how your brand appears in Generative AI results, you have a blind spot.
According to Salesforce in February 2025, 49 percent of Australians are frequent AI users. But the real story isn’t just usage; it’s trust and intent. From our own data at Sparro by Brainlabs, we’re seeing conversion traffic from Generative AI sit at 2–7x that of traditional search. While total volumes are still emerging, traffic from AI surfaces has grown more than 1,200 percent in the past year.
Why the outsized performance? It’s a shift in user psychology. Traditional search forces people to carry the cognitive load: sift through ads and blue links, compare options, cross-check reviews, and make sense of conflicting information. Generative AI flips that dynamic. It behaves less like a directory and more like a trusted assistant—curating, summarising, and recommending. When a user asks an LLM for advice, they get a definitive answer rather than a list of options. So by the time someone clicks through to your site from a ChatGPT or Gemini response, they aren’t browsing; they’re validating a recommendation they already feel confident in.
Trust in AI is contested. A recent University of Melbourne survey suggested only 46 percent of users are willing to trust AI. But behaviour tells a more nuanced story: people may be skeptical yet the data shows they still rely on AI for shopping and service recommendations where the payoff is clear and the friction is low. People say one thing and do another: they claim low trust but still follow AI recommendations when it’s useful. That behaviour—revealed preference—drives the higher conversion we’re seeing.
Despite this, most brands are flying blind. They aren’t measuring their visibility, share of voice, or the quality of citations within these platforms. The common objection is that LLMs generate unique, context-aware outputs for each user, making measurement “impossible.” But isn’t this true of most media? We’ve long used personas, panels, and survey methods to approximate and track performance. Why should AI be any different?
In practice, we’re seeing consistent patterns. For example, queries like “best restaurants in Sydney” tend to produce stable recommendations over time, even across different users. The models draw from similar sources, apply similar heuristics, and reflect dominant signals—quality content, strong reputation, consistent customer feedback, and clear product information.
So how do you start measuring? Build personas to stress-test your visibility. Ask: How does our brand appear to a price-conscious shopper? To a local who values convenience? To a loyalist of our biggest competitor? To someone researching a high stakes purchase versus a quick buy?
Then train models or structured prompts to assess your presence and performance against those personas. At a minimum, track:
- Mentions and citations: Are you being named? Where and how often?
- Share of voice: Relative presence versus competitors in relevant queries.
- Sentiment and visibility quality: Are references positive, neutral, or cautious? Do they reflect your actual value proposition?
- Referral traffic: Click-throughs from AI results to owned properties—site, app, store locator.
- Visibility into owned traffic: How do AI-referred users behave once on-site? Do they convert faster? Which pages and products do they favour?
The tools are new and imperfect. But if you wait for perfection, you’ll miss the baseline. Establish it now and track change over time. That’s how you’ll understand correlations—between content investments and visibility, between review quality and recommendations, between product availability and conversion.
Practically, three moves will put you ahead:
- Audit your AI footprint: Run core queries across major LLMs and AI-powered search experiences. Catalogue how, where, and why you appear—or don’t.
- Strengthen your signals: Ensure product data, FAQs, policies, reviews, and expert content are structured, current, and consistent. AI systems prize clarity and consensus.
- Close the loop: Tag AI-origin traffic, build reporting on the metrics above, and set an operating rhythm to test, learn, and iterate.
Generative AI is already shaping discovery and decision-making. Brands that embrace measurement and optimise for this environment will capture the intent-rich clicks and compound advantage. Those that ignore it risk becoming invisible in the next great shift in customer behaviour.