Budget Direct keeps AI search lead as brand visibility and recommendation split

Somantra’s latest audit of 20 Australian insurance brands finds Budget Direct still leads combined AI search mentions, while NRMA has edged Allianz on ChatGPT and brands are increasingly diverging between being named and being recommended. The report points to a growing gap between raw visibility and AI-driven consideration across ChatGPT and Google AI Overviews.

Why it matters: - Somantra’s audit suggests AI search is no longer just about getting mentioned. Brands also need to show up as the likely recommendation. - That split matters for insurers because AI answers can influence which brand consumers see as the default option before they compare policies. - The report also shows much of insurance advice in AI search still does not name a brand at all, leaving room for insurers that can win unbranded queries.

What happened: - Somantra released its third AI Search Visibility audit on August 31, 2026. - The report tracks how 20 Australian general insurance brands appeared across ChatGPT and Google AI Overviews from May through August 2026. - The dataset is based on more than 34,000 consumer conversations. - Budget Direct held the top combined AI search position in August with 9,960 observed mentions. - Budget Direct also led Google AI Overviews with 5,771 mentions. - NRMA overtook Allianz to lead ChatGPT for a second straight audit with 5,649 mentions, 83 ahead of Allianz. - On the matched July-to-August query set, Budget Direct fell 4.4%, Allianz fell 5.4%, AAMI fell 6.0%, and ING rose 4.2%.

The details: - Somantra’s Brand Consideration metric measures whether an AI answer frames a brand as the most likely recommendation rather than only naming it. - Shannons ranked No. 10 in raw visibility but No. 1 in most-likely framing in the elite, premium and vintage motor vehicles segment. - ING ranked No. 11 in visibility but No. 3 in consideration. - AAMI ranked five places lower on framing than on raw mentions. - NRMA ranked four places lower on framing than on raw mentions. - The report added three diagnostic layers not included in prior audits. - A Modifier Test found that changing one word in a prompt can materially change an AI answer. - The prompt phrase “most trusted” produced the largest positive sentiment shift, at +0.498, across tested brands. - The word “cheapest” produced the largest semantic distance from baseline, at 0.626, meaning it changed the substance of the answer rather than only the tone. - A First Seen Audit found Google AI Overviews places a tracked brand before the scroll line 85.6% of the time. - ChatGPT does that 34.0% of the time, and its longer answers push the first brand mention to a median of word 186. - Citation-layer analysis covered 7,945 cited domains across the two engines. - Canstar remained the top-cited domain across all three audits. - Every one of July’s top 100 cited domains stayed present in August, even as thousands of long-tail domains entered and exited. - In detailed ChatGPT conversational queries, 78.9% of category-level answers named no tracked insurer. - That brandless rate was only slightly different from the roughly 82% rate recorded in May and July. - Life insurance was the least-branded category, with 96.9% of ChatGPT conversation answers naming no tracked brand. - Somantra said that creates whitespace for insurers that can build content and answers around unbranded, in-depth advice queries. - The full report is available at the AI Search Visibility Report. - Somantra is a Sydney-based AI search brand mindshare monitoring platform focused on Answer Engine Optimisation and Generative Engine Optimisation. - Somantra tracks brand presence across ChatGPT, Google AI Overviews, Claude and Perplexity.

Between the lines: - The leaderboard is only part of the story. A brand can win mentions while losing recommendation strength. - Shannons’ and ING’s stronger consideration scores suggest AI systems may favor clearer category fit in some niches than broader, higher-volume brands. - NRMA’s strong ChatGPT visibility but weaker consideration score shows that repeated naming does not automatically translate into being the preferred answer. - The brandless-query data suggests AI search is still early in category recommendation behavior, especially for insurance.

What's next: - Somantra said the visibility gap will matter more as brands try to influence AI answers, not just appear in them. - The company expects more insurers to target unbranded advice queries and prompt contexts that improve recommendation framing. - The audit’s category-level breakdowns and visibility model appendix are now available in the full report. - Somantra also invites brands to use the new modifier, first-seen and citation-layer diagnostics to understand where AI answers shift and why.