When Anthropic's Claude topped a new US consumer satisfaction ranking last month, the finding barely touched Australia's insurance conversation directly. The survey covered American users, not Australian ones. But the timing lines up with a run of local developments: Commonwealth Bank has been steadily deepening a strategic partnership with Anthropic, the company recently opened its first Australian office in Sydney, and Australia's biggest general insurers have spent the past year building their own AI capability largely without publicly naming which foundation model sits underneath it. The US satisfaction data offers a useful lens on that local picture, even if it can't be read as a verdict on Australian preferences.
Anthropic's Claude posted a net satisfaction score of 59.3 among current and former US AI users surveyed between February and July 2026, in newly published YouGov BrandIndex research, ahead of Apple Intelligence (57.7), OpenAI's ChatGPT (53.5), Google's Gemini (51.3) and Perplexity AI (49.0). The score is a net figure: the share of people who say they're satisfied minus the share who say they're dissatisfied. DeepSeek, Amazon's Alexa, Microsoft's Copilot, Apple's Siri and xAI's Grok rounded out the top ten, with Grok recording the lowest score at 39.0.
This is a notably different order from raw popularity. ChatGPT remains the runaway leader on preference in the same US research, chosen by roughly a third of American AI users as their tool of choice. Satisfaction is a narrower test: it only counts people who've actually used a product, not everyone who recognises the name. On that test, Apple Intelligence and Perplexity AI, both mid-table on preference, jump into the top five.
The more telling split sits between current and former users of each tool, which works as a rough proxy for how a product holds up once the novelty wears off. Current users of every major AI brand report high satisfaction: DeepSeek edges out Claude at the top (76.4 versus 76.0), with Apple Intelligence (75.1), Perplexity (73.4) and ChatGPT (72.5) close behind. But once people stop using a tool, the numbers diverge sharply. ChatGPT's satisfaction score among former users collapses to -2.8, a 75.3-point swing from its current-user figure and the widest of any brand tracked, while Gemini falls from 69.7 to 10.7. Claude and Apple Intelligence hold up best after people move on, with former-user scores of 30.4 and 33.5 respectively.
It's tempting to read a US satisfaction ranking straight across into Australian procurement advice. That would be a mistake, for a couple of reasons specific to this market.
First, Australian general insurance is unusually concentrated: Insurance Australia Group, Suncorp, QBE and Allianz control around 74% of the market between them, a far higher share than comparable markets in the UK, US or Europe. That means the AI decisions of four boardrooms carry outsized weight for the whole sector, and none of the big four has publicly hitched its brand to a single named foundation model the way some US carriers have with Claude or ChatGPT. Instead, the pattern locally looks more like buying specialist capability built on top of a model, rather than naming the model itself.
Second, when US quoting app Insurify launched inside ChatGPT in February, sparking a wave of broker-stock selling in the US, the same news moved Australian insurance stocks too: Insurance Australia Group fell 3.3%, Suncorp dropped 2.4% and QBE slid 3.7% in a single session, according to contemporaneous reporting by Finance News Network, even though none of the three had any direct exposure to the US launch.
That's one sign the local market is watching US AI-distribution news as a proxy for where its own industry is headed, though it also shows how much of the reaction is sentiment rather than a reflection of what these insurers are actually doing with AI day to day.
Publicly, the picture at Australia's major insurers looks less like a single-vendor story and more like dozens of narrower deployments stitched together. Suncorp told Insurance Business it is banking on AI investment and a multi-year core systems overhaul to make cover more accessible, alongside a broader technology spend the outlet has separately put at roughly AUD 560 million, with about 90% of workloads now migrated to public cloud. Group-wide, Suncorp has told industry publication iStart it has more than 100 AI and machine-learning models in production and runs 15 separate AI chatbots that handled 1.6 million customer interactions in a recent half-year period, up 28% on the prior period.
That total includes AAMI's voice-and-camera assistant Ava and a Shannons Virtual Assistant that helps classic-car customers buy cover conversationally.
QBE's approach has leaned toward specialist partnerships rather than one flagship model. QBE Ventures, its investment arm, has backed insurance-specific AI vendors including Cytora for risk selection and pricing, Hyperscience for document digitisation and Tensorflight for property inspections, according to a summary published by data centre operator NextDC, layering large-language-model capability into narrow, already-built workflows rather than deploying a general-purpose assistant across the business.
IAG, for its part, has continued upgrading the digital systems its brokers and CGU and WFI-branded commercial lines rely on. The same NextDC summary reports that IAG also uses AI to personalise its quoting process and support fraud detection, alongside stated commitments to bias testing and impact assessment. Allianz, Suncorp, IAG and QBE have all deployed AI-driven catastrophe technology to give earlier warning of extreme weather, a use case with particular weight in a market this exposed to bushfire, flood and cyclone risk.
On the broking side, market commentary published on Livewire and Market Index has reported that AUB Group's BizCover platform has rolled out more than 35 AI-driven tools across customer engagement, compliance and claims workflows, and that Suncorp now conducts 78% of sales and 59% of servicing digitally, with around 65% of natural hazard claims lodged online.
None of this activity is framed publicly around a single foundation-model brand the way HUB International's Claude rollout or Liberty Mutual's ChatGPT quoting app have been in the US. It's mostly delivered through named insurtech vendors, whose own underlying model choices are rarely disclosed to the insurer's customers, or even to the insurer itself in every case.
The clearest exception sits just outside insurance proper: Commonwealth Bank has built and repeatedly expanded a strategic partnership and investment in Anthropic, first announced in March 2025 and extended since, and Anthropic has since opened a Sydney office specifically to grow its presence with Australian and New Zealand enterprise customers, citing financial services as one of the sectors it's prioritising locally.
That relationship is the most concrete, named foundation-model commitment in Australian financial services to date, and it gives Australian insurers a closer, adjacent-sector case study to watch than anything happening in US insurance specifically.
Insurance Business's own model-by-model guide for brokerages is a useful check on how uneven the economics of these platforms already are, regardless of which one tops a satisfaction chart. Microsoft's advertised US$30-a-seat Copilot Business tier works out closer to US$42.50 per user per month once the mandatory Microsoft 365 base subscription is included, two to four times the cost of a standalone Claude or ChatGPT subscription, while Claude's own enterprise pricing isn't published at all and requires a direct conversation with Anthropic's sales team.
The guide's broader advice to brokers is to match tools to tasks rather than standardise on one platform: Claude or GPT-5.5 for document-heavy analytical work where accuracy carries compliance risk, Copilot as the operational layer for firms already committed to Microsoft 365, and narrower tools like Gemini or Grok for specific jobs rather than as a primary platform.
That advice matters more once contracts scale to the size Suncorp's AU$560 million technology spend implies. A brokerage or insurer that builds workflows, staff training and client-facing tools around one vendor's document handling and context window isn't signing a monthly subscription. It's making a multi-year bet that's expensive to unwind if the chosen model turns out to be the kind that scores well on day one and sours by month six. That's roughly the pattern the US current-versus-former-user data flags for ChatGPT and Gemini, and roughly the pattern Claude and Apple Intelligence appear to avoid.
None of this means Australian insurers should read a US consumer survey as a procurement recommendation. But it is a useful early warning system. The vendors Australian insurers increasingly buy from, from specialist underwriting tools like Cytora to broker platforms like BizCover, are themselves built on top of a small number of foundation models, most of them the same ones YouGov is tracking in the US. If a model's consumer reputation sours enough to show up in swings like ChatGPT's post-use collapse, that reputational risk eventually flows upstream into the vendor tools built on it, whether or not anyone at an Australian insurer ever sees the underlying model's name on an invoice.
The satisfaction and preference figures above are drawn from YouGov BrandIndex US data collected between February 1 and July 31, 2026, covering 18 tracked AI brands, with satisfaction tables based on samples of more than 530 current and former US AI tool users (more than 220 current users and more than 280 former users for the current-versus-former comparison). Notion AI and Mistral were excluded from the satisfaction results due to low sample size. No directly comparable Australia-specific YouGov AI satisfaction study has been published at the time of writing; the Australian context in this piece draws on separate, publicly reported material about local insurers' AI deployments, market structure and stock market reaction.
