By Ron Arnold, founder 11eight

The insurance sector is experiencing an AI revolution. Insurers are investing heavily to automate back office functions, accelerate claims processing, detect fraud and personalise offerings. AI promises to handle complex workflows, collaborate across underwriting, compliance and customer engagement, and even learn from real-world scenarios to optimise outcomes. The focus has been internal: how can AI make insurers faster, leaner and more profitable?

However, there are signs another, external, revolution is taking place – one that could fundamentally redefine insurance distribution and the way insurers interface with customers: AI shopping.

AI shopping could transform the way people buy items online. Imagine a world where, instead of doing everything themselves, consumers can use AI agents – digital assistants – that shop, compare and even make purchases on their behalf. 

Big payment companies such as Visa and Mastercard are investing in this technology, seeing it as the next big step in online commerce. Visa, for example, is launching Visa Intelligent Commerce, which will let users set spending limits and then have AI agents handle the rest: searching, selecting, buying and managing orders. Mastercard is introducing Agent Pay, which enables AI agents to shop and pay on behalf of users securely. Both companies say they are collaborating with major AI firms including Anthropic, Microsoft, IBM, Mistral AI and OpenAI.

This new approach is expected to transform shopping, payments and the level of control consumers have, making the process easier, more secure and more personalised.

David Sexton, VP and head of insurance, global growth markets at tech consultant Cognizant, puts it this way: “By the end of the decade, most insurance transactions will involve AI assistants acting on behalf of human policyholders ... These digital advocates will negotiate terms, complete purchases and manage ongoing relationships with insurers. But rather than eliminating the human element, they’ll elevate it to where it actually matters.”

Lessons from Australian insurers’ response to comparison sites

Agentic AI in the hands of consumers might be described as “comparison sites on steroids”. These services have been highly disruptive in the UK but less so in Australian personal lines insurance. Examples of comparison sites here include Compare the Market and Canstar. 

Australia’s major personal lines brands have resisted comparison sites, declining to share data and threatening legal action in some cases. This reflects incumbents’ concerns such as “race to the bottom” pricing and commoditisation, and it has limited the role and functionality of comparison websites.

If insurers apply the same tactics to AI shopping agents – restricting online access, blocking or stymying them – the tech may face similar hurdles.

Buying insurance isn’t like buying shoes: legal and regulatory hurdles

AI shopping agents, particularly in the general insurance sector, face significant regulatory challenges, including the regulation of AI-provided insurance advice, liability uncertainties when AI systems err, evolving frameworks to ensure transparency and data privacy, and risks associated with the absence of human judgment.

These challenges necessitate robust legal, regulatory and technological solutions to protect consumers and maintain trust. Caution will be critical.

Somewhat paradoxically, laws that govern agentic commerce, empowering consumers, are probably the same ones that insurers and distributors will have to follow when using AI agents to service customers. 

Where to from here?

Until now, much of the industry’s enthusiasm for AI has focused on internal transformation. However, the potential – and the challenge – posed by AI shopping agents is very real. If agentic commerce gains traction, insurers will encounter a new landscape: consumers empowered by “their” AI agents. 

Insurers need to start thinking about their approach, not just in how they utilise AI internally but also how they would set up, compete and thrive in a market where the balance of power could shift towards the consumer. 

In the near term, the focus when building these tools is unlikely to be insurance; it is more likely to be on far simpler transactions.

In the longer term, the ability of big tech, payment platforms and credit providers to advance a truly comprehensive agentic insurance shopping experience may depend on how willing insurers are to make their systems compatible with them. Some insurers will probably resist these changes, while others will embrace them.  

Affordability issues and difficulty comparing coverage across insurers provide strong ammunition for those in favour of supporting properly regulated AI shopping agents.

While this future isn’t imminent, the growing ubiquity of AI and increasing consumer awareness of its capabilities suggest greater openness may eventually be driven by both regulatory pressure and consumer demand. 

Ultimately, insurers may not have much of a say in how it plays out.

After a successful executive career in government and insurance, Ron Arnold founded 11eight to help organisations solve complex challenges and drive long-term success by delivering transformative strategies.

For more in-depth analysis, features and opinion, read the latest Insurance News magazine