Trust-by-design marketing in Australia & New Zealand
My takeaways from Ernst & Young’s 2026 AI sentiment data (with APAC comparisons)
Last updated: 14th October, 2026
Privacy Awareness Week (4–10 May) has a habit of surfacing the same uncomfortable truth: trust isn’t something we “say” in a campaign, it’s something people feel in the experience we deliver.
In my view, that’s becoming even more important as AI becomes embedded in everyday products and marketing operations.
EY’s Global AI Sentiment Survey 2026 frames this shift as AI moving “from advice to authority”, from helping people decide to making decisions (and taking actions) on their behalf.
For marketers, that’s not a theoretical debate. It changes what audiences will tolerate, what they’ll question, and what they’ll report as “creepy” or unfair.
In this post, my focus was on the intersection of three recent concepts and events:
- The practical trust framework I’ve written about before (trust, transparency, accountability and security)
- The OAIC/Privacy Awareness Week, which reminded us that privacy expectations are lived in real moments (opt-in, preference, complaint, unsubscribe)
- EY’s recently published AI sentiment study, which found that trust is now a market-by-market deployment problem, not a universal brand message
Australia ranks equal lowest on global AI sentiment as use rises, EY study finds
“Australia ranks equal lowest for AI sentiment, trailing 21 other countries, according to EY’s latest EY Global AI Sentiment Study. The result highlights a trust gap that could slow confidence in adoption and value creation.
AI use and exposure in Australia continue to rise, with 77 per cent of Australians reporting they use AI. However, Australia’s EY AI Sentiment Index is 52 out of 100 (well below the global average of 66), reinforcing the gap between use and trust.”*
My take on Autonomous AI vs. an AI Assistant
At its core, autonomous AI refers to artificial intelligence systems that can think, make decisions, and execute tasks independently without constant human intervention or step-by-step programming.
Instead of just responding to a single prompt (like a standard chatbot) or following a rigid script (like traditional software), an autonomous AI agent is given a high-level goal and left to figure out the “how” on its own.
What EY’s 2026 AI sentiment study suggests
EY surveyed 18,152 people across 23 markets, including Australia and New Zealand.
The report’s big idea is that a meaningful minority is already delegating decisions to AI, and that minority is large enough to quickly reshape expectations.
One table that jumped out at me shows where people would prefer autonomous AI versus AI as an assistant in everyday scenarios. Even among people who don’t use AI today, EY notes that over a third would trust autonomous AI to resolve service issues, a sign that “minority behaviour” can tip into mainstream expectation once autonomy proves reliable.
From a marketing and CX perspective, my key takeaway is this: trust will be won (or lost) in the operational details, not in lofty statements about “responsible AI”.
Automating broken processes built on poor data foundations is not going to deliver great experiences that engender trust.
To balance out the steady flow of “Kool-Aid” from the tech sector, there are current examples of organisations that have gone “AI headfirst” and are now taking a step or two back.
Gartner’s structural prediction: Half of AI workforce cuts abandoned
10th September, 2025
Further context from market analysts confirms that this isn’t just a few isolated incidents. Research from Gartner (highlighted by CX Network) predicts that by 2027, half of the companies currently planning to slash their call centre staff in favour of AI will abandon those plans.
Firms are realising that completely agentless organisations are unrealistic. Instead, the industry is forcing a shift into a hybrid model: letting AI act as a basic filter, but actively routing anything remotely complex, sensitive, or high-risk back to a human call centre agent.
Swedish fintech Klarna begins rehiring human agents
15th September, 2025
After heavily publicising its massive success in replacing human customer service tasks with an AI assistant, the fintech giant Klarna quietly began walking back its fully automated ambitions.
According to reporting from CX Network, Klarna began rehiring human support staff after relying too heavily on AI agents.
They discovered that while AI is great for quick, basic inquiries, the sheer complexity and emotional nuance of financial customer interactions required a human touch.
The great AI rollback: 74% of Chatbots pulled offline
8th June, 2026
A major, widespread study reported by The Media Copilot revealed just how common this buyer’s remorse has become. A global survey of over 2,500 enterprise decision-makers found that 74% of companies that deployed AI agents in customer communications have been forced to shut them down or roll them back. The press coverage details a frustrating loop: customers repeat themselves multiple times, receive “confidently wrong” AI answers, and completely give up. To save their brand reputation and avoid losing customer trust, nearly three-quarters of these firms had to yank their bots offline and pivot back to human-led or heavily human-assisted frameworks.
Why APAC comparisons matter
(even if you’re focused on AU/NZ)
Here’s something I find genuinely interesting: Australia is included in the 23 markets, but it doesn’t appear on EY’s bubble chart (below) showing “Pioneer”, “Transitional” and “Lagging” markets. New Zealand does appear on that chart.
So rather than pretending EY has neatly categorised Australia, my view is we should use the report in a more practical way:
- Treat New Zealand’s placement as a useful reference point (it appears in the mid-pack)
- Use nearby APAC markets that are shown (Singapore, South Korea, Japan, Hong Kong SAR, China, India) as comparators for how different trust environments can be
- Use the Australian case study EY includes (Lendi Group) as a concrete example of “advice → authority” happening locally
On the bubble chart, EY’s message is that the shift to autonomous AI is not happening at the same speed everywhere. In “Pioneer Markets”, AI use is broader, more frequent, and more embedded in everyday life, creating conditions for delegation to happen sooner and at greater scale.
EY says that within Pioneer Markets, 94% report using AI, and nearly a quarter have used autonomous AI.
The chart also shows a handful of APAC markets sitting closer to the Pioneer cluster (notably Singapore and South Korea, with India, China and Hong Kong SAR also in the higher-use area), while Japan appears at the more cautious end of the sentiment spectrum. New Zealand appears around the middle.
My practical interpretation for AU/NZ marketers: APAC is not one trust environment. If your campaigns, preference models, and personalisation logic are designed for a high-adoption market, they can land badly in a more cautious one and vice versa.
A local Australian example of “advice → authority”: Lendi Guardian
EY includes a case study that’s worth calling out because it’s not “AI in the abstract”. It’s a consumer-facing service that moves from recommending to acting.
EY notes that Australia-based Lendi Group launched Lendi Guardian in 2025: an always-on, AI-powered home and loan digital companion that monitors mortgage rates and equity changes, and automatically pre-fills refinancing applications when rates drop. EY reports that moving from recommendation to execution increased platform engagement by 40%.
Whether you love that idea or hate it, it’s a clean illustration of the trust challenge marketers now face:
- If a system is going to act for me, I need to understand the limits
- I need to know what data it’s using
- I need to be able to stop it, override it, or correct it
Those are marketing questions as much as product questions, because the promises you make in acquisition and onboarding are the promises you’ll be held to when something goes wrong.
My trust framework for marketers
(updated for AI-shaped campaigns)
Last year, I wrote that “beyond slick campaigns that might convey trust, the best way to build trust is in the experience (CX) you deliver.” I still believe that. If anything, AI makes it more true.
Here’s how I’m thinking about trust-by-design marketing for Australia and New Zealand right now.
How do I build trust in email and lifecycle campaigns? Start with clarity.
EY’s report links higher comfort in Pioneer Markets to experience and training, and notes that people in those markets feel more comfortable sharing data to enable personalisation.
My view: in Australia and New Zealand, you often need to earn that comfort through control mechanisms:
- Offer topic-level preferences (not just “subscribe/unsubscribe”)
- Offer channel choice (email, SMS, phone) and cadence choice
- Provide an easy “reset” path: update details, correct preferences, or pause comms
Control is not just a compliance feature. It’s a conversion feature.
REPORT: Automated decision-making and public reporting under the Freedom of Information Act
The OAIC report reveals a critical lack of transparency, with only 17% of surveyed government agencies (keep reading) clearly disclosing their use of automated decision-making (ADM).
Whilst focused on the public sector, this systemic shift signals tightening regulations for email marketers, who must ensure data-driven audience segmentation and personalised offers remain transparent and explainable under emerging standards.
To-Do List for Email Marketers:
- Audit Data Sets: Review segmentation rules to eliminate hidden algorithmic bias.
- Update Privacy Policies: Clearly communicate the logic behind automated audience profiling and targeting.
- Ensure Explainability: Avoid “black-box” models; ensure you can explain why an individual received a specific offer.
How do I reduce low-grade anxiety about AI content? Be explicit about what’s automated.
One of the strongest trust signals in the EY report is the desire for transparency about AI influence. EY reports that an overwhelming majority agree AI outputs should clearly show when something is created or influenced by AI, and that nearly three-quarters worry they won’t be able to tell what’s real or fake as GenAI becomes more widespread.
In my experience, that worry shows up as disengagement: fewer replies, more silent unsubscribes, and more “mark as spam” when messages feel generic or inauthentic.
Practical ways I’d handle this in AU/NZ campaigns:
- If AI is generating customer-facing content, don’t pretend it isn’t; disclose it in a human way
- Use named ownership and a real reply path (not “no-reply”)
- Link to source pages and policies so people can verify claims quickly
How do I demonstrate accountability? Show your limits.
A concept I like (and that maps to EY’s “visible constraints”) is to state what you won’t do.
Examples:
- “We won’t sell your data.”
- “We won’t use sensitive attributes for targeting.”
- “We won’t auto-enrol you into new topics without asking.”
Then back it up in the experience:
- Make escalation paths obvious (“If this looks wrong, here’s how to fix it”)
- Put humans in the loop for high-stakes journeys (complaints, billing disputes, hardship)
In my view, this is where marketers can lead. We’re often the ones designing the moments where trust is either reinforced or broken.
A Quick Note on Ownership: Marketing Cube has no commercial relationship with Ernst & Young, and we haven’t received any compensation for this review. We simply found the insights in their 2026 AI Sentiment Report incredibly valuable and wanted to break them down specifically for our primary audience: email marketers. On a personal note, our author, Derek Bell, was employed by Ernst & Young from 2007 to 2009, but this review is entirely independent.
Turn privacy complaints into a trust-building CX
What’s inside the playbook?
- A 6-step complaint handling flow you can adopt internally
- A complaint-prevention checklist for opt-in, preference centres, and why you’re receiving this messaging
- Suggested complaint categories for marketing teams (so issues don’t get lost in inboxes)
- Sample acknowledgement copy you can adapt when someone complains
- A quick guide to building a learning loop (so you reduce repeat issues)
Marketing Automation Agency Services
Since 2007, Marketing Cube has been a trusted partner for marketing automation clients, specialising in Eloqua. In 2018, we expanded our expertise to include services for Salesforce Marketing Cloud clients. Our team excels at implementing your chosen platform and providing comprehensive agency services.
With a track record of assisting nearly 180 customers across diverse industries and geographies, we offer traditional application support and efficient help desk services.
If you’re seeking agency services, such as strategic guidance, end-user coaching, training, campaign support, and more for Oracle CX Marketing, Salesforce Marketing Cloud or HubSpot, reach out to us today to discuss how we can best assist you.
Frequently Asked Questions
Q. What is ‘trust-by-design’ marketing?
A. It’s about earning trust throughout the customer experience, not just with campaign slogans. With AI now making decisions on a customer’s behalf, trust is won or lost in the operational details, such as clear consent and visible safeguards.
Q. Why are APAC comparisons important for Australian and New Zealand marketers?
A. The pace of autonomous AI adoption varies across the region. Comparing your audience to nearby APAC markets helps you avoid applying campaign logic designed for highly adopting areas to more cautious audiences, and vice versa.
A. As AI moves from advice to authority, customers will tolerate less. If a process is unfair or opaque, they are far quicker to question it or report the experience as ‘creepy’. You must be explicit about the system’s limits.
Q. What is the fastest way to improve trust in email campaigns?
A. The fastest path to trust is clarity. Make the ‘why’ obvious when you ask for something. For example, explain why you need certain data under a form field or clearly state the trigger for an automated message.
Q. How can I make personalisation feel helpful instead of intrusive?
A. You must earn that comfort by giving control. Offer topic-level preferences, channel choice, and cadence options, not just a simple on/off switch. Control is a vital compliance feature, but it’s even better for conversions.
Q. What should I do if AI is generating my customer-facing content?
A. Be explicit about what’s automated. Disclose the AI influence in a human way and use a real reply path, not ‘no-reply’. This transparency counters people’s worries about what is real or fake in GenAI content.
Q. What does ‘demonstrating accountability’ mean in trust-by-design marketing?
A. This maps to ‘visible constraints’. State clearly what you won’t do with their data. Crucially, ensure escalation paths are clear and involve humans for high-stakes customer journeys, such as complaints or billing disputes.
